The Latest

0:00 so met observation I'm starting to get chitchat a little more um it is a it's like a group chat like a WhatsApp chat but uh it's a little bit more like hanging out with friends async when um it's hard to get everybody in the same place or at the same time or some combination thereof the interesting thing about it is why is it different than just using a WhatsApp with voice memo um because in theory people could do that and people sometimes use voice memos on WhatsApp is it just the ux is it just the fact that you can kind of uh sort of do an audio draft and relisten to it um you know is it is it the concept of okay this is just psychologically a hangout that I have on the desktop and um you know I'm kind of hanging out with my friends in this audio room over here like you know like the way Discord works um but async audio rather than sync audio uh you know I guess maybe that's what it is and uh if so I think I can see it I maybe it's just because I process information visually much more quickly than I can hear it that um you know for just speed of information intake I prefer usually chats but I can start to see where async audio is interesting asynchronous uh group chat I see it also as a Pinterest of ideas where you can you have an idea you can start to put them together it's also a Production Tool so you can make a podcast you can refine them you can make uh something new out of it easily it also combined some of the paint points away from the traditional models where you doing a podcast uh there's usually sometime of editing post production going on and um it's a barrier to start doing something because

2:04 you know that that's the work which is waiting for you to do before you can publish anything so this is like a lure to get into to do something just quickly and then you realize that you have actually some substantial body of stuff to you have put together uh with these bites it's easy to see that everybody can basically have their own um customized um stream or or I don't know how do you call chapters or theme based uh composition of today's topics or something you missed or discussions going on so um like I'll go feed of Twitter but uh in this uh form chat put together the latest developments in the Multiverse discussions happening in the last week I would love to get something like that that going on and uh give me the arguments uh for and against the the the most prominent uh topics or the ideas uh coming out of the last month another way of describing uh what's going on here is like a fragmentation of ideas but in a positive way so there's like a thousand tiny pieces which you can stitch or clue together and make something out of them all new ways of seeing things um well in a way that's like what the llm models are at the moment uh it's it's not a new body of knowledge it's just different angles to the same existing uh stuff we already have throwing the latest tools where you can transfer between text image uh convert to video audio whatnot uh different language translations all put together and that's the the latest in the rich media experience combined with uh ideas thoughts

4:08 Concepts so is this the new Twitter in a sense that uh Twitter used to be to place for people who have uh ideas uh to express them and Instagram was for the visuals but uh this is the version two or three obviously I don't want to make too much of a comparison to the Twitter but in a way a place for the ideas to evolve and develop themselves The Genius of this platform is the same as uh with the original Twitter 140 characters maybe with the 280 characters as well that you have to be uh concise and have a Clarity in your thinking in order to in short uh time or amount of characters uh express your idea and this is usually the curse of the podcasts uh there's a low information density as I tend to call it um so basically a lot of words a lot of time goes by and not too many ideas not too many uh things are coming out of it obviously if you're just making entertaining content or stuff like that doesn't apply building your content uh out of these legal blocks of high information density uh bits I think is really the way to go into the future there are great tools like descript uh which actually was a a little bit of an inspiration for me to start a podcasting again 2020 uh but it starts from the processing side from the not from the content it starts from the tools and um I'm not to excited about that approach in general these are some very good insights about what's going on here so thank you for noticing it's pretty awesome that you can now uh create a

6:11 podcast episode which basically doesn't end or you can uh upend to it uh just whenever you feel like it uh add Co commentary replies and uh take it further uh still using the same link um in a way I think this reminds me of the Bachi open Kindle book uh which is rewriting well not itself but uh the user the end user who bought it once with 10 bucks uh gets always the latest version uh so it's like a placeholder for the episode um so it's the air chat episode a new season in addition to that you you don't want set speakers and you also want to leave long enough time between the different pieces of the conversation that people can genuinely think about what you somebody said and come back with a thoughtful response as opposed to having to give a knee-jerk reaction in the moment and then of course the listeners should be able to join in and you can also retake uh if you don't feel like that your answer was the exactly there which is impossible on the live unusual podcast situations yes as a consequence I think our exported podcasts make the speaker sound much better than Live recorded podcasts and not only the speakers but also the guests may feel more comfortable because they know that it's not a live situation is not uh one take uh or nothing situation and and they have their time to reflect things as well uh so um I think it's a win-win for everyone also for the audience because if you get more concise more thoughtful answers and that's good obviously doesn't work for all the all podcast which are shoting [ __ ] and stuff but uh for those where actually there are some uh things to talk about and Concepts and that type of more factual based

8:15 things this may also allow the par to actually become a guest on on a podcast because what I well I have on booed quite a many people who who are prominent uh done a lot of stuff doing even public speaking but they haven't done any podcast uh episodes before as a guest and and there's certain um barriers for people mental barriers to to speak publicly do this type of thing and and this sounds more like a casual chat and more familiar situation from other apps so I think that's a plus as well yes but people get stuck in their habits I brought some podcasters on here and they sort of Overlook all the advantages and and start just thinking about their familiar situation so somebody will have to break the ice it absolutely creates this middle Place between other social apps and sharing and content creation and there is an idea maze that happens in here that is just incredibly fabulous especially when you see it as in real time like during the banking crisis during the FD FX crisis it was amazing um and then and the last thing I would say is that it's doing all this Middle Ground stuff really well and that's incredible but it brings this human warm intimate element to it that maybe it's simple but no one else is doing it as well fly air chat I think this sets is this is like the Bitcoin and Fiat thing it is a perfect counterweight to tw Twitter Twitter is and I believe always will the fastest place to go get real time news and real time use and and consume a bunch of information albeit in a little bit of a cold fashion by reading and then come here and talk about it if you want to talk about it with people here not saying the Comm asso only have to happen here but the other thing is the way this audio is uh

10:18 and the way the warmth is if it became like everyone's trying to report up what's going on in the world in the way that Twitter does um I think that would be overwhelming in the audio format and even though the book mode is terrific and almost works like Twitter in terms of consumption um I think it would become overwhelming and yeah I think you know I don't it's up to them but I don't want them to be Twitter Twitter's Twitter it's fine this is such a perfect counterweight and again this reminds me of the Bitcoin Fiat discussion like if Twitter were to disappear the world in air chat would not be as great because we wouldn't have this like great source to go and just see the various opinions in real time and then talk about them um similar to if the US dollar disappeared uh Bitcoin wouldn't be able to solve so many problems that you know the the the traditional banking system actually has solved and I don't know um there's a big Chesters F fence thing that I think Bitcoin Advocates they get so obsessed with the price and like the um the dogmatism of like uh it almost becomes religious and they lose the fact that uh yeah it actually works pretty well works pretty well uh but it doesn't work well it's terrible but it works pretty well long shits for life just kidding uh hey man this is cool I like your channel long shits for life just kidding uh hey man this is cool I like your channel this ties into an old saying from their old Twitter days it's like the bigger the meeting the more it feels like you're on stage and the more feels like you're on stage the less free flowing philosophy occurs and that free flowing philosophy the Casual chitchat that's how you get innovative ideas to emerge where people don't have their

12:21 guards up so when you're on stage you have your guard up you know it's all about your Public Image how people perceiving you but when you're in a casual chitchat when you're in a conversation you're just immersed in the subject you're in in the interest in the Curiosity that's how the ideas bubble up to the surface from the subconscious you know when your guards are down that's how they break through it's funny we spent so much time crafting a project or you know whatever you're creating and at the end of it when it's done and does all the things that you hoped for originally you look back and you think to yourself man those days were the good old days where I was building that thing you know and now it's out there but like when you're in it and you're building it it's not like the good old days it's like gosh I got to get this working I got to get this working you know I got to finish the project I got to get it functional it's so funny how that is and then afterwards you look back and you're like man those were the good old days so it's just good to have that conscientiousness in mind and like you know afterwards you're going to be nostalgic about this period of your life you know so you might as well be nostalgic about it now and just realize like this is it like this is the fun part it's not once everything's done once everything's done then what we're just chilling you know it's like this is my fun uh you know so it's not the end product it's the process and the end product is also the whole point but you know without it there is no fun in the process but um all the fun isn't just end of the end it's it's now even though we can distribute

14:26 information and the speed of life so basically when somebody gets an idea we can distribute it immediately but yet it still takes years decades generations to get the new ideas to become mainstream enough that they will be act upon and that's also something hofman mentions but it's quite clear we've been losing more or less 100 years in in many science Fields uh if if we look now back and think what what what was happening at the early 20th century in itself that's nothing new love L and puppets invented the computer 100 years earlier than we basically implemented and did it uh so we could have a computer Revolution starting way way back if only the same thing with a lot of a lot of stuff it's you don't need a fancy technology you just need to sit down and think about things and new thoughts around and new ideas can be built to the world Ed at kantrum it's so easy to do this that you tend to jump the gun and start talking without thinking yet exactly because this is asynchronous and you can take your time to respond so that you can uh speak clearly you can be concise and you can have a Clarity in your thinking and in your speak which way to go is that really true I mean sometimes you can come with a design or an idea for something but the Practical technology to implement it and make it Mass scale doesn't quite exist yet and so because of that you can create these little Oddities here and there but you can't really make progress in the field because all the other pieces are not in place Robert prown 1827 looking through microscope Poland uh took um almost 80 years for Einstein and others to then Define prown and motion and there was basically nothing

16:31 between uh that the technology uh prevented to come up with that idea fair enough but I don't think eight years is actually a lot of time for a scientific breakthrough especially in a time period when you didn't have telecommunications technology and digital libraries and easy publishing and all that stuff so the pace of innovation now is getting faster I think the moment from now when some new fundamental breakthrough happens to when people start taking advantage of it is a much much shorter window in general I think breakthroughs in one area speed up breakthroughs in other areas uh we're operating on such an incredible infrastructure stack now of realtime news and which is built on top of electronics and internet which is built on top of just microprocessors and transistors which is built on top of electricity which is built on top of just bearing cables and lines you know which is built on top of of course fast communication and coding Shannon's work and information Theory which is built on top of just language so all these Innovations build on top of each other and then make it even easier and faster to innovate so AI for example I think is going to enable faster Innovation curve because it'll be easier to draw inferences across far-flung pieces of data yes in 8 years is not um a long time in the the course of human history but if you think about somebody say to me that you can can only have like uh seven dogs or something like that I don't remember exact figure the idea that if if the dog lives like 10 years uh aox you can have uh you can Pace your life uh to seven dogs or you know small number of animals pets you have and and that sort of puts some kind of limit to the things you can do so yes 80 years is not a long time in that sense but consider that human life is really short at this point uh it is a really long time considered that if if things could have been improved early and we have the means to do those things so basically 80

18:34 years is uh a few Generations waiting to die out and new ones picking up the stuff absolutely so the key question is what prevents us from moving faster obviously I think regulations are a big deal especially when it comes to human health um I think access to information is so good now that it would be hard to point to that as the reason um bad memes anti-rational memes which are everywhere um but yeah what is it that keeps us from moving faster as a society I'm used to Silicon Valley tech industry barrier which moves quite fast blisteringly fast bewilderingly fast why does the rest of the world not move that fast why don't all Industries move that fast my go-to answer is always regulation but maybe that's not the only one or maybe that's wrong I don't think it's wrong um I think that was part of the maybe it was the the Yan nber's book open the the story of human progress I hope I get it right um his point was that basically Europe was so bad at uh regulating uh blocking Innovation that we were the ones who succeeded you know many many centuries ago and and this is basically where the uh Western civilized world is based upon now so um that that's part of it then earlier on there was always times when everything was reseted there was some new open field to go to uh well us opened up and every was called Russ and well uh California uh and and then there was always internet came so there's always been some non-regulated some place where you can go you can explore innovate and do new stuff sort of a valve uh where this uh puddled up Innovation and and freedom uh has been able to directs uh it it's it's it's Focus but uh now we sort of stuck there hasn't been resets so to speak and everything is sort of putting

20:37 down to uh to Halt uh because there's no carpet colle collection in regulation and and these type of things but I have also other idea but that's uh for the next I think the biggest issue still is the human nature um the they all saying that you have to wait for the old generation to die for the new new ideas to pick up I I think that's basically the essence of it uh so it's bundled with um ego issues in a sense that you've been investing so much uh your reputation what do you know and maybe your sense of self uh to your career and then something uh comes and and puts you totally in a different position saying that hey this is not true anymore or you know we have better ideas or this doesn't work and it's really hard actually to reset everything and uh keep on going uh so basically I think that partly that's the one then then there's also the um uh the fact that uh a lot of the well publicly funded ideas are not exactly favoring Innovation they are more um favoring incumbent and the known knowns not the unknown unknowns uh risk taking is not part of it and um I I think that's a pretty important part of it as well in a sense that um Public Funding has been taking over uh a lot of fields which used to be private and it's really hard for some of the fields to get funding um and um well a lot of people gravitate towards where the meal is coming from in instead of Where the Heart Is did you know that you can uh make any PDF any text basically um a spoken audio and you can choose the and this is in iPhone choose the uh speaker AI voice at the speed and other parameters and and turn it into a

22:42 um well audio file um I I think this is pretty cool if you have a long PDF file and uh you want to go outside and be in the Sun or walk or run or do something and and hey you can listen to that one uh without uh having somebody else uh narrative The Voice did you know that you can uh make any PDF any text basically um a spoken audio and you can choose the and this is in iPhone choose the uh speaker AI voice the speed and other parameters and and turn it into a um well audio file um I I think this is pretty cool if you have a long PDF file and uh you want to go outside and be in the Sun or walk or run or do something and and hey you can listen to that one uh without uh having somebody else narate the voice that's a good phrase there's been no garbage collection and regulation you're right these laws stick around forever and they become more and more invasive so now we've fled to the internet now we' fled to crypto but there's regulation thanks to AI coming at math and crypto and code as well so pretty annoying I also don't think it's a coincidence that the big breakthroughs in Tech have really been an unregulated Frontiers recently almost in the domain of pure mathematics and code and that's crypto and that's AI absolutely true the ego gets stronger in most people as they get older and the lucky few it gets weaker and the stronger ego stands in the way of progress Douglas Adams had a funny saying along the lines where like everything that's invented before you were like 20 is like cool and fun and you know something that you can uh build a career upon everything between like invented between 30 and 50 is kind of interesting and everything invented after your 50 is an obstacle and a bad thing and should not be allowed his actual quote is much better I suggest you look it up but it was something along those lines one of the Prime uses of GPT 4 should be programming Apple shortcuts and apple script cuz this they're so hidden kudos to Apple for trying to

24:45 build a programming language and interface that ordinary humans can use but ordinary humans don't want to program at all the only programming language they want to use is English yes s Cuts is amazing uh if you just know how to do it because sometimes it's a bit weird how they execute things and you just have to know it and it's not obvious and intuitive but if GPD for would do the trick uh that's a way better way to do it than looking forward to that I already hacked the uh Series so uh I can buy and Via shortcuts uh use and ask anything and and Trigger it from uh quick actions and different places and uh yeah I get the power GPT uh from the city there's also this um history Rhymes but doesn't repeat itself uh type of thing going uh unfortunately with all these uh bad stuff in the world I mean really bad stuff genocide War all these sort of basic principles which take usually the well if your grandparents were part of it you still remember if if you were lucky that they were talking about those things and and you know you have a common shared memory but after that uh the the pond is broken in a way that uh the history doesn't pass from the history books uh the same way so uh people tend to repeat the same mistakes funly enough I think that's the same thing with the inheritance is the third generation who um spend it all uh tends to take a actual physical U memory of of sharing information and experiences with people not to repeat the same mistakes so this is an example of a shortcut that already does this pretty impressive that it was built within shortcuts itself so there's loads of these that just send things off to the chat GPT and then come back with the answer whereas this will detect certain keywords like playlists and certain other shortcut stroke um Siri Integrations and then turn what chat GPT

26:50 does into into a native iPhone action so a simple ones like a playlist give me top top 10 songs by this artist or whatever it might be and then it uh use chat GP to get that back and then uses shortcuts to turn it into an apple music playlist um there's also a few apps that I think PT on the IOS app store um includes a similar functionality but this shortcut is completely free uses your own API code and just shows um how where things can go so this is an example of a shortcut that already does this pretty impressive that it built within shortcuts itself so there's loads of these that just send things off to the chat GPT and then come back with the answer where is this or detect certain keywords like playlists and certain other shortcut stroke um Siri Integrations and then turn what chat GPT does into into a native iPhone action so a simple ones like a playlist give me the top top 10 songs by this artist or whatever it might be in there it uh use chap GP to get that back and then use his shortcuts to turn it into an apple music playlist um there's also a few apps that I think PT on the IOS app store um includes a similar functionality but this shortcut is completely free uses your own API code and just shows um how where things can go co times wear what they wear but uh I have a fun memories of Clubhouse the early days of Clubhouse I wasn't exactly the earliest there but still rather rather early um Malcolm mdov the actor in The Clockwork Orange He talks about the Stanley Krick stuff I don't think he ever spoke about that before I think was something some controversy about the post uh his name was supposed to be my if if I recall correctly and and then it was taken away in the last minute but he was really opening up it was the first

28:52 time he was in Clubhouse and and he spoke like two or three hours uh a lot of intimate discussions or that's at at least how it felt uh then I met a lot of um musicians who were really talented and couldn't play anywhere else Antonia already introduced in in here air chat Yi um I interviewed her in the in my podcast talks with Petra you can check her out uh and and many many more I wonder what comes uh out of Clubhouse next well hang on let's not forget about Ludwig uh boltzman boltzman did some incredible work that did rely on um Brown's work but he's the one who determined statistic he invented statistical mechanics and that rolls up into a lot of progress and then it also relates interestingly through where that that entropy which is initially uh a concept that's discussed in the context of uh gases really which is boltzman's kind of bread and butter uh that concept of entropy very much relates into information with information being the inverse and to this day I think that relationship between information and entropy is actually a pretty meaningful area of science that still has a lot lot of work to be done in it um and then obviously you know information gets into Shannon's like I think you know Von nman jokingly said to Shannon that uh he should use entropy because then no one could argue with him but there was actually a really deep insight there and so I think Von Newman knew it was a joke when he was saying it yeah shortcuts are I I don't understand why they don't emphasize it more there's been so many things I'm like oh how do I possibly do this or I Google it like is there an app to do this and what comes up instead is like go into your shortcuts do this it's pretty amazing it it it reminds me of the early days of uh like Visual Basic in Excel um we just kind of Cobble together stuff and then maybe it's not pretty but it works every time you hit the button so I see shortcuts very much as a p will

30:57 use the feature and there is that risk of overwhelming the standard user I think this is some of the discussions in the air chat feedback room is that you want Power users all want different features and I guess shortcuts as Apple's attempt at addressing that which just allows them to build it the sou versus versus Apple building it and then rolling out to everyone and overwhelming menus and confusing things it's that actually a really clever solution I hope they will actually import also the automator uh more advanced features like um I think I was doing Crown chops so you could do SK scheduled um well scheduled scripts and uh that was really helpful uh for a lot of automations but um at the moment shortcuts is quite clunky in that way that uh well you can do automations but only on the iOS uh uh and iPads but you cannot do it on on the computer yet and uh that limits a lot but you know if you could uh basically have some kind of triggers for automations that would be fantastic currently there's a lot of um new users in in Blue Sky two months ago they had a 200 people in the network and uh then basically they had a a lot of um invites which were not used and suddenly just you know in the last few weeks a lot of people been actually using them and obviously they also been giving new ones out and now there's 50,000 people in the network uh and you you can see that they starting to have a little bit of issues with the moderation the tools are not in place there are some health threats uh and um yeah it's uh getting out of hand in in a way but I guess it's positive that there's a uh well be people are interested of um decentralized networks and uh

33:00 competition is good it seems to be that there's a lot of uh creatives coming in uh to the network at this point so you if you want to C some Hollywood writers or uh directors uh now is a good time to join uh in a way you can say that this is a reset um you have a chance for well it's a tiny window but you have a chance to um get your presence uh to be known for some of these luminaries who have a lot of followers in the other networks and uh uh suddenly they are in a position where they might only have like a 50 and if you reply or you you you do something you make it uh their attention and and this in a way this is a fresh start for some people but um this doesn't obviously last uh too long but U I think it's a great that um young people or those who missed the the previous waves uh now have a chance as well to recreate the networks uh build them again or or reshape them in new format this one is uh from forast uh 16 of September 2021 when they had a first time 100 daily active users and I think now they have between 10 and 20K all together users so networks are taking different times to do things but obviously um as blue sky foraster is also invite only so they have a regulating valve how quickly they want to turn the knop this is pretty amazing what you can already now do um I don't know what are the YouTube uh API limits uh can you pull the whole Lex fredman uh transcripts uh out of it or maybe pred Hall YouTube channel is available soon in a way that you can ask chat GPT uh about topics and the whole content is uh accessible just by asking maybe somebody here Ed that can do that even I

35:06 think this quote is magnificent something I've experienced over the years many times sometimes the problem is really difficult can stump someone for years even decades and then a solution emerges that's so simple that it def your intuition you're like it can't be this simple and then you realize it is and then you think oh I could have done this myself or this this is so simple that it's trivial like this it's almost like impostor syndrome where it's not even an achievement but it is an achievement because the problem was so hard you know it's kind of like the idea of the bicycle oh yeah just has two wheels it's so easy you know anybody could have came up with the bicycle but nobody did for a very long time same with the car same thing with tons of things that we think are simple but are only simple because we're so familiar with them and because they're so easy to understand after the fact before they're immensely complex and it's like that quote I can't remember who it's from but he said uh if they would have gave me more time I would have made it simpler or I would have wrote a shorter letter condensing something down to that level of Simplicity that's like Masterpiece level that's like finished product before when it's all complicated usually it's not completely figured out yet so cherish the Simplicity but don't get me wrong Simplicity doesn't always equal ultimate truth to have the ultimate truth and the ultimate functionality merged with something so simple that almost anyone can understand it that

37:10 is the ultimate State and it takes a very very long time to get there so funny that it's underappreciated sometimes when you do arrive but if you have had the endurance to last that long then you won't mind the under appreciation CU you're never going to get to that finished Point unless you can keep on going despite people not appreciating uh you know you trying to find this masterpiece answer at the end of the rainbow so by the time you're there and you actually do find the answer everything will be unlocked anyways and I figure you'll be pretty excited whatever it is you're working on applies to most if not all Fields I think what's the value of the views the headset icon for non-authors yes it shows the agregate number but um air chat is not about the mass views advertising model so I'm I'm really questioning why it's there I understand the value for the author to see who's been viewing and listening to the chit but for the rest why uh there's been a lot of discussion and actually action as well in the norer network in the last few well I would say this year about the views the function of the views and because a lot of the stuff is coming from the advertising model and it's not really needed if you don't have an advertising model or even if you cannot count the views properly because the well there's no Central State um it has gone so far that uh it became almost like the chap Wars uh chapping meaning the tipping in the in the nosta world and uh the leading iOS client damos uh removed the like button and wanted to indicate that hey um if you want to acknowledge the order of the of the notes uh send a little bit of a fraction of a uh sent uh over and it's

39:14 way better signal than just liking something so there was an only zaps mode you could take it off if you didn't uh like uh but the thing was that you couldn't see the likes either for the other people's notes and there was a lot of people who were like oh no I don't want to that want to have that at all uh I'm I'm stop using damus um then actually Iris um implemented the option that you can take them all way all the reactions chaps sles retweets and even replies uh and now it's optional you can pick and choose whatever you like and um this is basically I think is a really good question to ask what are the ways to acknowledge what our ways to show your appreciation show you your presence uh when we are not really present but not taking the advertising or revenue or business models uh imposing things on us which probably lead to the wrong signals and uh well what do you measure that's what you were going to get Bazi settled his pets early and I don't know why but for some reason I'm puzzled with that um not the decision itself it's by by all means it's money but whether whether that actually undermines the signal he's sending why do it early why not wait um and I think that's probably the underlying issue underneath uh if there's a 10% probability or whatever it is and you settled it early why it's just standard on social networks these days people want to show off how many Impressions they got you saw on Twitter they put it back front and center so people could kind of get bragging right same reason people want you don't know how many followers they have we don't feel strongly about it whenever that uh right rail gets crowded like it may get crowded again because we have some new stuff coming up the first thing to go is the headset icon it gets

41:16 crammed underneath likes and you may see it disappear as early as 2 or 3 days from now by the way I wouldn't assume that this is not going to be an Advertising based business model after all in audio and podcast that is currently how people monetize and even if we don't want to monetize heavily we're still going to be forced to just to split revenues with the creators every big social network now is doing that and I think large creators are going to start expecting it more and more that said it wouldn't be our our first choice subscriptions or tips or some kind of a micro payments or purchasing model would be way better absolutely ADM model is nothing new people love free stuff and they want somebody else to pick up the tab and that's just how the world is uh but when there's a always a chance to do a little bit of a reset um it's always good to think about a bit maybe there's a ped away but I don't know how much the creators of of the tech apps can influence people to do things differently uh nevertheless that was just my whole point I don't know how true this is but I heard that somewhere around 10% of YouTube's Revenue was coming from subscriptions and 90% was coming from advertising so clearly they started with advertising and advertising on TV is big but still that's a little disheartening for someone who's trying to consider a subscription based model that said we want to keep a small lean mean product team we don't want to build out a big advertising and sales team and this is much further down the road obviously this thing will either have to massively succeed or fail by then so um but much further down the road we would rather much much rather try to find premium features where we can charge for subscriptions as opposed to have to do advertising but I think the pressure to distribute money to creators if this ends up being a hierarchical Creator network will force us towards advertising even then there's different kinds of advertising there's brand advertising where we have few large relationships which I would hate and

43:19 then there's kind of Bottoms Up grounds up advertising like the type that Google does where anyone can buy an adword and it goes into an algor algorithmic slot and you don't necess need a lot of sales people to sell it although Google still use a lot of sales people to sell it they would probably tell me I'm crazy for thinking you can do advertising without salespeople but I'm going to try anyway I think nosta is a really good example of now trying to I wouldn't say impose but nuts uh people to um send tips kind of the Netflix subscription model but you know more basis that for each and every um note you you can tip you can just uh click the button it's not much but it will agregate because it's so easy um it may become like a habit and if you start it early on I I think there's a chance that people get used to it it's like when you go to into restaurants at least in the US and some part of the world uh you know that you need to tip in some part of the world well where I am mainly that's not what we do uh but you pick up the customer and the Habit wherever you are so in that similar way I don't know whether it will actually be enough that if you could U siphon a bit of the tips uh from the content providers but that would be also incentive for the uh creators that if you can see that um whatever you create you're getting money uh so nothing new once again o um substack uh all those whove been testing those models for many many years already but the fact is that people are paying for services Netflix uh newspaper subscriptions and uh the rest uh the bigger issue I see uh is the alignment if you're going for the advertising model it tends to well corrupt in a sense your business model

45:25 because you need to listen to them because that's where your money is coming from so that will direct the the product as well and it conflicts uh in its uh core with the users because you have to serve for the advertisers and even though you're struggling and you're fighting against the notion but the fact is that your your Revenue model is in Conflict uh with the user's best in interest in some sense and uh well we see it in apple as well it was not too many years ago they were saying that they all for the privacy of the user but yeah subscriptions is more likely to work here than microtransactions if microtransactions work anywhere it's going to be on Noster and blue sky and so on where you already have cryptographic keys underneath your accounts so in a sense you already have a wallet um let's see if it works I think the tax Reg regulations United States and the tracking around that makes it difficult um and there may be other obstacles besides just user Behavior but um a tipping based economy would be nice there's also an argument that we don't want the kinds of content creators who want to get paid because eventually then the whole app degenerates into listicles and um selling vitamins and selling branded water and selling fashion and selling food and it just content quality goes way down maybe it is really about authentic conversations about people who generally want to share ideas and learn we're keeping a small team we don't need to make a lot of money I think ideologically we're built to serve users not advertisers and we don't care that much about the money that said you're right in that institutions take on a life of their own and then eventually get corrupted by incentives regardless of whatever ideology the founders might have so you're right so advertising is kind of a last resort model yes and I I think that's uh probably in the essence of of this uh if

47:29 you're building this for yourself and you love this type of discussions and uh well high quality content in some sense ideas thoughts uh argumenting and uh well to appeal for these type of things is pretty small considered to all the Tik toks and and whatnot where you just U lips and and and do dancing and and all all the silliness um but maybe that's the knce You could argue that is it good content if you're not getting paid for it uh yes uh there are public services yes if you're accomplished you don't need the money you don't care uh and in some cases you just do it anyways uh but uh money tends to be also a motivation for things but maybe it's for the secondhand dealers of ideas and uh maybe that's where the line is drawn yeah I think the views are they're good feedback because you can get a sense of whether the conversation you're having has any interest to anyone um so there's good feedback mechanism there and then if you go into them you can sort of see who's more interested as well and that gives you a good sense now I think that's what it's doing whether the fact that it's doing that is Meaningful to the platform at all I don't know since people liking chits and whatnot isn't really standard uh does seem like some sort of user feedback would be good but I don't know if it should be the headphones or not I love where you've mentioned and I think it would be a really cool thing to see happen if in particular AI platforms that have a cool tool but don't have distribution yet come to air chat and it's possible for an air chat user to subscribe to that AI tool being available to them in the app and revenue sharing there makes a ton of sense you know listening to your talk

49:32 about this cuz you're right the large advertisers eventually have asks that are anti-user so you know I but I also think money is a good way to like get real information and real feedback so it's tough uh you could maybe consider a hybrid version where you are having some large advertisers but the large advertisers are essentially paying into the pool that gets distributed somehow to the creators on the platform so you're not really in the middle of it and you I always find I always find it I'm terrible at advocating for myself but I always find it really easy to advocate for other people so then if I'm in that conversation I say look it's you know it's going to be a million dollars to sponsor the platform and it all goes to the users and I'm not going to you know sure maybe I would give you a break or take this pressure or whatnot but like really we serve the users that's the point and if you don't like it then that's fine don't be here um but it sends the signal that like yeah you've got to pay to be here cuz you're business but also uh you can't rug pull me because at the end of the day I don't really care I'm just going to distribute to the users and if it gets cut off then every us is just going to see you stop sponsoring you know or every Creator is going to see you stop uh sponsoring as far as the Tipping model uh as far as the Tipping model like I have a lot of experience in that in the hospitality industry and I actually did some work where um I helped the Divide between it that was cool I I I came up with a system that really helped address the wage gap between the front house and back house but what I was going to say is that so tipping is standard in the restaurant industry I don't like Auto gratuities ever unless it's like an enormous party because I want the incentive the incentive does work really well and that's why the service is exceptional um in the United States I think um but online you know I would almost say that using a tip I think doesn't get used as often as having a paid subscription where one of the options is don't pay right now it's a small little hack but you may as well take that side of it and you're right eventually you are going to have to monetize or rather you don't have to monetize but you'll want your creators to monetize in a way that doesn't turn them into like total clown shoes

51:35 attention Hogs that are you know blasting your platform with this garbage we see everywhere else it's a good observation that if we don't monetize for our users somebody else will or they will and they will use horrible tools to do it and they'll start just like blasting ads in the middle of their content so it's better that we come up with a single clean weighted do it that's good what do you think caused the Blue Sky invite inflection I haven't followed closely enough to to have anything definite uh but I have one story for you um I don't remember anymore when LinkedIn was uh found it even and but but I I do remember when it actually hit in Finland uh that was one weekend uh when basila a local entrepreneur was sick he was having a flow nothing serious but he was bored he was bored out of his mind oh well that's what I think at least and he started to send out invit to LinkedIn to his friends and he was well connected and uh lo and behold after that weekend more or less everyone who was everyone or anyone in the tech scene uh was already in LinkedIn so that's how it happened it happened because of a flu so maybe that's um the case with blue sky as well somebody was just bought out of their mind or they wanted to escape a a bird site or combination of all these things once upon a time there was a rich man who owned a media platform that requested access to a public decentralized media platform but there was a problem there was 50,000 people already in but they didn't have have any invite here's an app I don't have any compelling reason to come back to well probably I'm not in the target market but it's creating the all traditional news houses uh feed and learning by AI

53:41 or whatever algorithm your taste and it's by the Instagram Founders so they clearly think that there's something into this thing the fun of tech first you're like wow this is going to change the world it can do so many things amazing fantastic uh then you dive in you play around with it uh you figured out the limitations pretty quickly it's only the fancy demo cases that work the the rest is like a hack put together and and it it basically falls apart when you try to do some serious stuff and uh then you like disappoint it and it's like okay what happens now I have to wait uh oh so many possibilities but they're not here they come here but I have to wait or I have to build it myself duh the future of books and libraries previously you have to read take time hours days to get the information get the stories get the thrill well nowadays you can go to Netflix and check the fiction and you can even fast forward and it's way more faster well you're losing a little bit of imagination but the interface is different the content may be the same now you can ask questions and learn it that way can you summarize this chapter can you give me the main five ideas can you give me the the storyline of this thing what happens with that character maybe this works better with the non-fiction but yet it's a completely different way of observing the same knowledge faster I bet what's the future of Salesforce saps uh hotspots and whatnot when you can build tiny micro apps with the AI tools and you can put them together they have nice interfaces they can do one thing really well uh they are modifiable they are dynamic they are composable they can do wonders

55:46 if you couple them together and they can also change according to your business needs what becomes more Val aable when most of the content is done by AIS there's no intention there's no thought behind it it's just the product most of the social media phone calls Outreach campaigns are just done by BTS for BTS what's left the human thoughts the genuine connection between persons how do you cross over that barrier when it's practically a pot war and firewalls are up how about a delayed access by that I mean that um well you can invite your friends for free but uh if somebody who you don't know a third party wants to join the discussion um they have to pay so you're just creating your content that but the idea is that um the rest of the people will never get it if you keep it private but you know maybe it becomes publicly available in 15 30 days or some type of DeLay So as a Creator you don't basically need to do anything you just be yourself and you create content and those who are curious enough and want to get access early well they pay for that privilege I have something similar but even better in mind so stand by it's going to be fun I don't mean to be mysterious but we now have full-on competitors cloning us so unfortunately that's just the way it goes why the corporates are going to Blue Sky well first of all I think they need a place to go um the apis are really closing down now in Twitter so they can use it for either posting their stories anymore automatically or with the publishing

57:51 systems another thing is the same happens also with the with the DMS so they cannot use the customer service anymore so they clearly need us some place to reach out to their customers and also uh tell Their audience that there's some new stuff to read um then the other thing is that if you ask where to go well you could go to masteron uh but it's a bit awkward you could set up your own server but um yeah plus guy has a Twitter feel and look so um you can also use your own domain as a verification and obviously you could also put up your own server and in in that way have your own instance and maybe extend your offering for your readers as well and put a pay wall or whatnot in the future so it has lot of these appeals and if you ask where else to go not too many obvious answers nost maybe but um I think it feels a bit too radical uh for the corporate folks this could be seen as the next um predatory model or someone exploiting other people's content and repackaging it well if it's in public domain then that's what it is um on the other hand I I think Fred had also good points that uh well what is left it's not all in the papers it's not not all in the in the tweets is not all in the in the blog post it's not all about in the public discussions it's not all about in the YouTube That's the K stuff that's the old stuff that's the Frozen um ideas if you want to get the most dynamic version if you want to get the customized version you still have to get to the verion in real life and that comes back to the uh my previous chit about uh human touchs the connection

59:54 between people the undivided attention where you can only get when you have a presence in the moment with someone uh this is an example of um great technology what's available at the moment anything you do in your computer you can rewind anything you have done you can double check so would you like to have a perfect memory whatever you been doing um I was considering it and uh my answer was nope what happens when search engines start to give uh J GPT type of summaries so they're not just uh giving uh queries back with based on the keyword search but they are actually giving the concept so they are summarizing the content of the website so if there's like a tons of Articles and uh it's just giving actually an to the questions well if you have books like Google Books uh how about that one how is the fair use and the copyright laws going with this one so you never basically go back to the real sources anymore you just use the search engines to fetch things or is this basically the chat GPT the AI which becomes a search engine interesting times agents agents agents um I believe that it's you that was talking about Donald Hoffman's Theory Of Consciousness now it's been a while since I heard it and I'm certainly no Hoffman expert but from what I remember he put it together on it's a pretty decent podcast called Theory of Everything um toe the toe podcast but Hoffman had described his series of Consciousness as basically you know forget about what he thinks of what we're looking at or not looking at uh but Consciousness as you know an amalgamation of a lot of different agents and that are sort of different taking controls and being aggregated up

1:01:58 and down broken apart etc etc and I actually think it's pretty useful and I think that agent based approach um is going to be really powerful going forward uh cuz now we have ai agents so uh I think you're just going to be managing an elite team or a large team rather of um of intelligent agents that are go going out and all aware of a common goal and working on it in their own individual ways and that gets pretty close to you know uh how individual organisms survive and compete in an ecosystem this is an interesting twist in a decentralized uh protocol um especially with the corporate employees and uh those who are getting a lot of followers due to their job here's an idea for somebody using shortcuts and having Apple devices so whenever you open air chat uh your music will mute and whenever you go away from the app Up the Music resumes how many you can stack before it starts to shake I rest my case I asked [ __ ] chat which is now using the chat gbd4 uh to give a summary of that um archive or article named Laura low rank adoption of large language models and this is what U Bing is saying the article you provided is a scientific paper titled a deep learning approach for a detection of covid-19 from chest x-ray images the article does a really good job of giving basically the biggest talking points but all together I would say that is it better to have a large huge models which are slow to do in big companies and also you have a large body of um data sets but um they may not be as high quality as these smaller ones U also if you're individual and you're working in open source world you are not bound by the um licenses uh you can do stuff uh and and basically taking the uh Japan Japanese um automobile industry Innovation tiny tweaks which are not so

1:04:03 interesting uh as themselves but a lot of people doing those things individually All the Time Around the Clock uh just pipe your interest and being really good at those things combin and uh then becomes the issue also with the data sets if you have a black box can you really use it if if if there's no uh Clarity where the content uh the data was coming from and is the bottom line that the big is not better it's better to uh actually be in the C of California um sell the tools and keep away with the capital keep away from the capital expenses I've had this very same concern I don't think it's fair use to just ingest other people's data mash it up and then give you the answer without linking back or attribution or driving traffic search engines in the early days there was quite a bit of War around what the concept of fair use was and I think one of the reasons why search engines got away with fair use is because they were driving traffics and they were showing a very limited um transcript or the first few lines they weren't actually giving you the full answer and it's one of the reasons why I think Google's one box which is it summarized search results at the top have been held back to some extent I think Google crossed the line with that summarizing thing they were practically doing the chat GPD thing they they gave the full answer they they were ripping the whole content and not driving traffic anymore and um well evidently the advertising model is not working as as good as it used to do so so they take they need to do something but that's that's the Black Ops this raises the the larger issue of of using somebody else's content I remember there was a case where an artist latest piece of art was refused uh because the it it was too similar to the AI

1:06:07 generated uh pieces well the twist is that uh her previous art was used part of the data set without her consent and and that was the the result so in this literal sense the artist was actually replaced by her own work but on the other hand it's like in the Fred Wilson's article it's only the old stuff uh but I understand that uh when you're doing creative art uh it it you you have a fingerprint usually you have a style and you want to stick to that one or that's who you are so it's not like that you can do different stuff in that sense so it it is distin distinguishable uh and it is a huge problem in in that case that is a great point when you're valued for a fingerprint that you invented uh trademarks patents intellectual property law all of this exists to give you some protection around that fingerprint but if that fingerprint is swallowed and regurgitated is it uh plagiarism is it copyright theft or is it only if literally that fingerprint is produced how many times you have to remix that fingerprint how many ways across how many other fingerprints before it's considered original and I think most people have a poor definition of creativity so they think creativity is just remixing things and so for many people you can argue them into a corner that this is uh original but then on the other hand if they understand how it's made if they see this case of an artist being replaced by an AI version of that artist then they can also see that it's just a copy so this is going to be very difficult situation and I'm very interested in seeing where Society comes out on this Andy warhall is a great Point uh he he was creating art he was having a a literal Factory of of producing copies of his work he loved to to make money uh but he was also popularizing well pop art so I think what happens now is that everybody has their 50 minutes in

1:08:10 literal sense it may take that 1550 minutes to actually make it part of popular art but you know you don't have the time to actually commercialize it like Andy had Andy had years decades uh to to to build a on on that and if if you look like the from the '90s 2000s course and the other ones who are doing similar stuff I I think it's the same thing it's so morakami it's so distinct it's so identifiable but if you don't have time to actually commercialize it as a young artist I think that's a huge issue this documentary which is on Netflix um gives an idea about how he was produced using stuff and uh his creative Factory is the message the medium uh this is what I was thinking just today uh should I actually start to export some of the discussions to other other places social media but I'm not actually sure the reason is that um well they can listen to the things this is the Lan back mode uh but um the conversation is different it's not the same it doesn't have the same feel if it's only text uh if whether it's in Twitter whether it's in Noster whether it's in some other other places it's the place sets the tone and and the the context for the discussions as well and if you repeat out of the original content is not the same discussion anymore uh and if you just point towards this uh chit for example uh then it's not the same because the other people cannot at this point contribute either so it's I don't know that's right it's a conversation and to really get everything out of a conversation you have to be part of the conversation that said I still there will still be occasional chits or threads or pieces of chits that would benefit from export and being Standalone but um try too hard and you'll have people creating py aism on here and building Brands out of that and that would be a

1:10:14 shame the other thing I'm discovering at the moment is that um well I'm I've been having this uh steps in a stream uh room where basically I'm putting stuff uh during the day uh but should I Stitch things where I'm into another discussions like here so that somebody who wants to see what I've been up to can can follow also for me to actually refer back later on because sometimes could be really hard to find the discussions uh with s just replies to somebody else's rooms threats but nevertheless if you want to reuse them and also for later if you want to actually Stitch something Creer together first it was just a reply to something but then it became a topic and then it became a room uh how to do that um two weeks back two months back and actually a question for you as well uh do you uh is it is it possible to export all the all the chits lat on where I'm being part of those which are mine the rooms are okay but then if I've been just giving some uh random comments this and There and Everywhere are they also included in my backup if you will this interesting thread at the topic uh what it's here is not exactly the content of the thread but is when Apple was in serious trouble 1997 W asked um well they stuff and other people uh what what should be done and there were 101 answers to the question and obviously we know already now the answer but U it's uh humbling to see what what was the tested at the time and we are in that situation every day we don't know about the future but we have a lot of ideas yes this is a pretty exciting uh New Era Avenue Universe whatever word you want to um use combining not the decentralized uh public uh not service with payments uh and then you have ai World which is uh open sourced if you

1:12:17 combine these two together it opens up a lot of possibilities where you don't have to ask permission from anyone and you can explore and it just takes ideas Innovation brilliant people and a weekend and you can probably do something that the big guys haven't uh managed to do in a months or years before 100 billion here 100 billion there and soon we will start talk about real money if open AI is uh trying to achieve the AI uh and they're going to invest 100 billion for that Venture good luck uh and Google is investing in the next five years uh similar amount uh to development of the llms uh what about meta and Apple so uh throwing in other few players as well and and we are in a range of at least half trillion so in a federal terms it's like a half chip then Jeff uh pesos was a huge advocate of pizza box teams uh was it around two pizzas and and that's the maximum size of the team um the same with the micro apps so the tiny Services uh which then basically have these interfaces and talk to each other but apparently now it seems that U there's a bundling unbundling going on and now it's time to bundle things to the pickup packages and uh it's uh not always uh best to have these tiny micro apps uh in the Amazon Prime article they say that they reached a peak maximum capacity already at 5% in some of the services so it may not be always the most efficient way of running things this drives with my experience anytime that I've thought about doing

1:14:21 microservices it never seemed like it was really worth it uh that said I think there are some cases where it does make sense so I feel the same way about microservices as I do about optimization which is build it first as a monolith measure it really figure out that you actually do need one and only in that case then go for it so uh tall Spike short tail here some not so obvious uh advantages of a type of discussion is that you cannot interrupt people think about it you just cannot and that's a amazing thing uh but it also is a preventative thing because those people who actually want to dominate those who actually just want to roll over you and push their point of view be rude they cannot do it so they may even avoid uh that type of discussion because they cannot take it over uh this also helps uh people who are not so um comfortable speaking uh instant dly uh on in real life chat where you have to respond immediately so you can take your time uh but that's something I probably already mentioned and and obviously it also gives the voice for those people who are not so powerful in the realize situations dominating so you can have more equal voices and message discussions uh because many times if you have teams uh the team Dynamic is usually not representing the quality of the conversation and this puts it in a completely different level and maybe you also think more what you're going to say because it needs to stand on its own legs without the other um qualities you can input to your output uh while speaking in uh in in in

1:16:30 synchronous situation just this morning I was just thinking because I was installing Bing on my phone and uh well it's also in some other places now because the J cbd4 is there and it's becoming quite nice and never in the world I was thought about doing that because it's Microsoft and and all the packages comes with that but it it starts to become like an OP because it can create images you can ask questions uh it can do the searches uh from the internet as well and it soon having the plugins as well and you start to ask questions like okay why would I go to Google can can I do those things with Google and apple is sort of uh fallen from the three already uh it it's not in the space even so the options really are like open AI or or being and being is free in some sense take the example of learning a new habit and un learning an old one I think from the brain's perspective you create synapses and then milein based in repetition and there's a concept or at least it was um research it was a accepted research data that you need say 66 days to change a new habit and I think because we understand so little about how the brain works that we still believe that is true and I don't get why we couldn't do that faster if our technology is presenting information to us faster and we're able to summarize it and understand it faster why can we not implement it into our brain faster I have an interesting theory about this I don't understand why we are opposed to learning faster or anyone could be opposed to learning faster if you think that in the last 100 years our technology has exponentially sped up and

1:18:36 our nervous system is tuned into our environment I think it's just a matter of understanding our potential and we could learn much faster now than we did 50 years ago even 10 years ago I think we're heading into Neo's learning mode in The Matrix and I'm all in learning new that's okay giving up the old stuff that's harder and I think that's the resistance why things are changing so slowly I spoke to 3 and a half thousand students uh at the Australian student leadership conference the other day and my opening slide was and and the kids literally started screaming now was so excited it was the first person to set foot on Mars is currently sitting in a classroom and the response to that statement in itself that was made by an administrator of NASA that I shared on the screen was so large these young people are so excited about technology they're so excited about gamification they're so excited about computer science and AI um it's the teachers that aren't it's the curriculum developers that aren't and we need to start shaking things up so the students can advocate for themselves I to get SU is smaller still based on the um 19th century uh paradigms educating people to the factories and having a kindergarten or daycare for for the kids so it's not about learning it's it's just uh keeping them busy uh it's not about uh teaching them how to think it's not about uh um teaching them how to uh well giving them the tools for the future if because if if we would do that uh the way things are done it's not exactly how you would plan it to do it now if you start from scratch uh given all the tools the ways you can learn and you can do it

1:20:39 asynchronously like we doing here in air chat those tools would make it so efficient that maybe you need to educate yourself or learn something only a few hours a day and the rest of the day you can do something fun or create a project sex or whatnot but uh I don't think it takes 8 hours a day 6 hours a day to sit in a class that's not a normal way of doing things adults are not doing work anymore that way either well some of them are not hi thanks nbal for putting this room together so I'm a high school student at Alpha which is a school where we have no teachers we use apps so we can learn two times as fast within the span of three hours each morning so all of our course subjects are in those 3 hours and this frees up our afternoons to work on our alpha x projects which are our Olympic level masterpieces um mine for example is researching a way to prevent and cure breast cancer through the connection between epigenetics and food super happy to answer any of y'all's questions this is true what I see from working with people with movement habits patterns is that once the new habit is perceived as valuable by the brain more efficient especially more pleasurable easier more fun it automatically overrides the old habit over time but inevitably your brain will select to do what feels good and works better over the old way few hours days enough uh I haven't checked what has happened since but this book is over 10 years old uh but but it talks about over 100 years old institution uh making customized learning for students already many many years ago and uh Diana thanks to Diana uh you can check out the alpha

1:22:43 room here as well they are already doing that so so high school students uh uh learning lessons uh from apps a few hours a day and that's enough what happens when AI pots the llm models will start to take over the back offices of municipalities governments a lot of bureaucrats work that may already now been outsourced somewhere but there uh Outsourcing companies figuring out how to do it without the employees scaling so we're getting a lot of people who are pushing papers who are not needed anymore and the interface is just a Common Language you explain and consider it done if more than half of the early subscribers to the Twitter PL have already cancelled this doesn't promise too good for the cash flow of the company so is there like a real risk that it's going to go bankrupt this year will the public learn if you can query any leg body any pills suggested or passed already you can check the whole history of anybody who has been elected whether they actually do what they promis to do you don't have to trust you can verify you can just ask who did this bill benefit how has this candidate been able to perform have they actually ever done anything they promised to do where did they oppose uh etc etc whatever you can ask how does this go with the think tanks uh do you only need a think tank pot in the future just add your bias and you are good to go no the founder just drop an article where he's U pointing

1:24:46 out that uh blue sky isn't really a protocol in a sense that it's not open uh a lot of things are centralized there's only one instance um there's a lot of good points in there I wouldn't call it a scam but uh having open AI as an example going out all good and uh for the benefit of people uh what not open source and then close it up uh in this case um yeah you can you can say and I think that's what they believe they start with with this instance and doing this way but when going to they open it up when is that moment there's never going to be a perfect moment and longer you wait uh harder it is to build something this this is the same thing what's happening with the nost at the moment um some of the clients are building centralized backends there are few big uh relays nodes at the moment which pretty much everyone is using so it's becoming centralized def facto even though it's open but it's just more convenient and and this is the way it's going to tilt uh the further along it goes so the idealism Fades away the true decentralization never happens because it's not built upon from the beginning and the inconvenience and imperfectness of building the hard way uh from the decentralization up um never happens I think if anything Blue Sky was more like a philosophical Cry for Help like it was um a noble intention and maybe a a means of encouragement to get the ball rolling and to display uh demand for this type of thing from Twitter from the administration of Twitter but um that's just kind of the wrong way to do things though when it comes to like

1:26:50 really getting a good result you know it might have encouraged other people like me and uh viat jaff and stuff to take a shot at it so it was successful in that sense but it failed in the sense that you don't start with funding and then work on a protocol you start with a protocol and then you get funding for it you know if you are expecting to come up with the hard part after you get funding then of course it's not going to go well most of the time um um quite a quite a moonshot so it's just common sense you know I mean this is not something that money can buy if money could buy these decentralized protocols uh then billions of dollars that's been funnel into the space would have gotten us somewhere but it's created mostly garbage um so I don't think money is going to be what does this it's going to require some Blood Sweat and Tears true cyppher Punk ethos and even then we'll see the real challenge is making these protocols costly to attack especially the part where you're trying to discover honest nodes because if you go to discover nodes and just a bunch of Cil nodes then you're in trouble you know and if creating a bunch of siil nodes is easy or requires trusting a third party to find uh nodes then then you're in trouble you know and I don't think Blue Sky uh creates any cost when it comes to discovering these aggregator notes um I don't know they're very vague about the whole thing and it just very I will uh I will restrain from commenting on on it further but uh in terms of Noster like you know they rely on the mutual relay model so I just find a friend and see the relays he's connected to and if I trust him

1:28:53 and he's found some good relays then it's all good you know it just that doesn't scale very far but it works for now so I think noer definitely has the leg up here and hopefully there will be better ways to discover noes in the future that are extremely costly to attack and scale far better once we have that um and ways to store large files and stuff uh I think we're going to be Off to the Races steps in a stream this is for some historians check but and verify but I think that's how it is with protocols they start with some a few people who are maybe outliers in some sense and it's it's messy in the beginning and then it catches up but you you cannot like you said you you cannot put it with money has it ever worked uh that way has it ever been the case that you can just buy your way to the protocol and if you buy your way to the protocol it's already Twisted in the sense that well aren't you asking for some kind of a um payback you need to return the money somehow or you have some ideas how you want to go about it there's only pretty much like the founding of Bitcoin not too many of these uh completely uh selfless uh acts uh of uh putting up great stuff which actually catches up the other thing is that if you truly want to have it decentralized at scale or and in scale um it needs to be so easy to do that it's basically part of every client uh the whole thing but can you actually build this type of a monol from the beginning because the value chain will uh build up there there will always be middlemen they'll always be these aggregators they'll always be these ones these are the weak points and um through decentralization peer-to-peer kind of

1:30:57 worked to a point till it didn't and then now we have hosting and aws's and whatnot so the file sharing didn't catch up CI is another take on the same same thing but how to solve this issue that uh we don't really need anyone anything which is in the middle and make it so unattractive to that middle party to build up their position I think that's what's required so it needs to be so decentralized from the beginning that it's already game lost everyone on this app already knows this but I I think it's a really great Point uh so if you're working only synchronously you have hard time doing it uh very concisely and asynchronously but those who actually Master this uh short art form can do the long art form the same thing that if you can do really good short tweets having a long article is no issue but it doesn't go the other way around so this is going to be a big change or I think it's already a big change uh for a lot of people and maybe it's a generational thing as well some people just cannot cope this and uh I'm really looking forward actually having more people doing Asun stuff but it seems to be really that lot of people are rolling back to the ex uh post preco times where things were just uh happening slowly second I push everything asynchronous and unscheduled that's why I built this app or at least one of the key motivators why I built this app but do you ever feel that the more uh features you add in the more cool things in it becomes more of a distraction rather than you know a change how do you kind of battle with that stability AI founder CEO EMR just posted his first blog post he said it's pretty lonely to be CEO and listed also some

1:33:02 other reasons also stating that uh uh speed is the easiest then comes text and and then the visuals is the most uh complex form of communication uh in regards to how much effort it takes and AI will flatten that out but nevertheless I think he would benefit from a that as a medium to express his thoughts and I think we would benefit from his insights as well anyone want to throw him an invite this is already some half a years old uh podcast episode but uh I I think he as far as I recall I I listened to it when it came out but I think he he was really opening up and there was a lot of uh background to where he's coming from from and why did he establish stability Ai and he also talks about his personal life and uh the health issues he's having I think this is a pretty powerful concept uh the founders are here I can put a cheat on on on a comment reminds me of uh when I'm when I used to play some console games Xbox games for example a lots uh uh huge uh games with huge Maps uh and I was either listening to some podcasts or I was listening to some music so in practice I was more or less like walking or I was having a visual uh Concepts and and places and when I visited those later on and the music was not on or I was listening to that music uh without playing it was the same Association uh if I would actually be like doing that uh in real life like talking with somebody over the phone or in real life and we were walking somewhere so I can attach the conversation uh to some physical locations and uh you can also remember at least I can remember the discussions

1:35:06 uh just by going forward so this visualization I think is a really powerful tool if your brain happens to uh work that way uh this response by rle who is a Twitter co-founder and been building this centralized networks for a long time and also know a blue sky team and is friends with them uh check it out I will in the next chits uh elaborate a bit more this is an intro if you want to check out what's been happening before uh it also mentions the rles article I posted previously this is a good um state of the no and blue sky in 60 minutes or 30 minutes if you do it in 2x uh it summarizes uh the weak points and what's going to happen or what can happen in the future and why things are as they are at this point uh rubel points out that uh at the moment Blue Sky which is the company holding uh the client and holding the keys physically the keys as well because uh that may be news for someone but not your keys not your data they are only one pair of keys and and those are by the Blue Sky which is signing all the uh all the messages so you cannot really at this point take your keys and go away it's not possible so um it is in a in a way it's uh similar to the signal even though um you have signal and and you can copy the code and run your own things but you cannot interoperate with the other signal instances or actually with the official one so what do you do with the uh protocol is it the protocol if uh if you cannot actually play with other people you just like an isolated island which doesn't communicate with anyone this is the state of uh uh blue sky at this point he calls this as a cathedral model where there is one single entity uh team of people maybe

1:37:13 one person like in mardon's case uh who controls the building of the infrastructure and the ecosystem system uh you can even ask whether it's an ecosystem if there's some single point of failure or choke point in a way that you have to appeal to someone Lobby them to do something hope that they actually will um come through to their promises uh is that a a protocol and can you build it later on he also mentions that uh when is the good time to do anything when is a good time to step back um I I mentioned early on that I don't think there's ever a good time and it's getting harder and harder the further you go along the uh development process um the opposite view of uh Cathedral model is a Bazaar model where nobody's in control uh you have only soft influence uh if if you don't like something you can build it anyways but usually it's better that you uh try to get some other people also to cooperate with you so you you you have a more appeal to the market and you have a better product what the other people can also use and uh hopefully also that other people will see what you've been building because if if it's it's not a protocol if at least it's not standardized if if if it's not working properly for everyone those who are building creating writing something and those who are reading it so that's why the soft influence is so important and that's why it's so chaotic that's uh why it's also fun and he mentions uh the the fun point that um the crazy stuff is happening in NOA because you can you can build fun stuff and people love fun stuff and and that's sort of part of the appeal you just don't do the Bing back end and uh and and documentation and and whatnot uh which doesn't appeal to the end youer you can put some crazy stuff on it and and make it uh really appealing and uh worth your while even

1:39:19 though it may not be the the the most important thing from all together but the biggest takeaway I think is that uh it's not about the people because Rael is saying that the People Are Awesome who are building uh Blue Sky they've been handpick they've been doing this for a long time they know exactly what they're doing they they knew what they were getting into so it's not about the intentions is not that they are against the vision what nost is for example building the common goal is the same he doesn't even believe that the the persons are trying to build something uh which is a business for the for the blue sky so it's not about that but I think the real point is that alignment matters the structure you have is very important if you have a strong wrong structure in a sense that it's not aligned with the common goals the public goals uh the public domain goals it's really hard to do it right Blue Sky was uh C of C in sense that uh they wanted to do this uh in their uh offices without too much attention and and build it properly but yet uh the invites went out and somehow it just uh exploded in a sense I think there's now 70k people in the network and they just released as well that uh no heads of States so they don't want to get certain type of uh attention at this point yet um can you build it and when is the right time to to release it to the world uh if you are de in incentify

1:41:22 Builders who are seeing all the bus who are not part of the core Loop of the decisions which are now made uh about the protocol and the future of the ecosystem the protocol capture is a real danger meaning that uh it will just be one nice instance and it was also mentioned in the video that uh people really want just a Twitter clone they just want to have it like a Twitter and just yesterday I thought I was riching something but it actually was risking so I was actually doing it in Blue Sky even I was just confused with the uh interfaces they look the same uh and those people may not care about uh protocols and all the fancy stuff underneath they just want to do the same thing over and over again and that's absolutely fine but um the builders at this point and some other people may have more deeper ideas and way to save the future that we are not repeating all the web two uh 2.0 mistakes and it seems to be that nosta might be a better chance for this having funding cions um building blue sky at this moment being overwhelmed by the demand by the a lot of people there having attacks by trolls and um needs to build uh moderation and all these things coming with the running an instance distracts and then the appeal comes that should we take a little bit more money should we just try to keep this all afloat and because the structure is not done in a way Rubble shows a do doent when Blue Sky was uh they were discussing about

1:43:25 blue sky that how it should be structured and he suggested that there should be a foundation uh responsible of the code and then a private company running an instance having the business model of itself but it's just one of the menu like an example of the possibilities uh to be built upon the protocol because this didn't happen this conflict remain and it's up to the people to keep the integrity and sounds to me that it's harder to do the split if there ever will be a split the further along you go David asked Rubble has there been in the history some successful cases where a top down build of a protocol uh successful and he gave some answers and I don't I don't want to spoil them for you check them out in the video but the outcome is that um it doesn't build an ecosystem yeah they can remain they can still be sustained but for other reasons so this Outlook doesn't look too promising for blue sky at this point maybe a time to check nost how much much of your private life is going to open AI without you knowing it there's a lot of services there's a lot of uh companies using J GPT and in order to do the job they need to send information are they anonymizing are they taking off the private sensitive information which can identify you I doubt it some are doing probably most of them M maybe not are they excluding all that information from the uh learning models how big of a bomb is this

1:45:30 soon um well maybe you have to as always assume that anything you send out is out of your control and in public domain but yet just use that uh we really desperately need some um open source models and mulard uh LL L type of service this is a huge loss for well I think for everyone so Twitter is dying as we speak this segment at 4 hours 4 minutes by wol from catches a bit what I was about to say or trying to say that there are certain aspects of the universe so you know we think space is made timately these atoms of space and the hyp graphs and so on and we think that uh but we nevertheless perceive the universe at a large scale it be like continuous space and so on um we uh in quantum mechanics we think that there are these many threads of time these many threads of History yet we kind of stand so so you know in quantum mechanics and our models of physics there are these time is not a single thread time breaks into many threads they Branch they merge and but we are part of that branching merging universe and so our brains are also branching emerging and so when we perceive the universe we are branching brains perceiving a branching universe and so the fact that the claim that we we believe that we are persistent in time we have the single thread of experience that's the statement that somehow we manag to aggregate together those separate threads of time that are separated in the operation of in the fundamental operation of the universe so just as in space we're averaging over some big region of space and we're looking at many many the aggregate effects of many items of space so similarly what we call branchial space the space of these these Quantum branches we are effectively averaging over many different branches of possible hist of the universe and so and in thams we're averaging over many configurations of you know many many possible positions of molecules so what if you have infin compute doesn't it look from our perspective which is obviously not correct but just to give some mental picture static oh we started to see a trend where YouTube and other content creators who have hours or in this case for example 2,000 hours of uh video content rich media content they're going to be deleted from the public consumption and they're going to be

1:47:32 turned into a chat G PT or some kind of AI pot or whatever format which you can subscribe and you can ask questions or entertain yourself so if you couldn't make it enough uh picking the YouTube advertising monetization model maybe you have enough to turn it into some other type of uh subscription based service or if you have a website and you could do the same so is this the beginning of uh once again silad internet behind pay wall cardens is it nobody knows M no modes big small who knows here's some basics of uh four kinds of universality required for human level intelligence uh first is computational can run any given program two reachability can eventually generate any program idea three stepwise can take any kind of step into space of programs ideas for comparison can compare select options using any criteria there's um 8 minute video in the top of the Twitter thread check it out it's way better explain there du Punk was so ahead of day time check this video from uh 3 minute Mark onwards till the end and you will see familiar stuff which is appearing more and more nowadays thanks to image to animation models a CH is like a box of chocolate but it's also Pokemon type of Treasury hunt if you know that that you said something and you would like to find it in somebody else's room where you replied good luck same way if you seen someone uh whose uh content you

1:49:35 like but where did they reply all those things they are not in the wind but they are somewhere in the air chat so you start to roam the rooms you start to wonderer open the chocolate box and once in a while you find these gems they were just hidden in different corners of the Box can you afford not to use these augmentation tools available already now just you p a link to a project in this case it's a GitHub or repit code base and it will give you suggestions ideas maybe Point some errors optimization things you could probably even get a better version of something when you just ask the right questions and reest what you would like to have versus uh you hope that somebody else is interested of your project if it's an open source project and will contribute and and point out those things uh what the AI just did for you when big search engines are starting to answer your questions without going to the original sources and this will not drive traffic and obviously it's not good for the content creators why should I do it I don't gain anything somebody else reaps the benefits the monetization out of it this is the same question um for a lot of other Industries as well and I don't think there are really good answers uh at the moment but uh I think this is getting to the low level in a way that if the fight has been more like in the news organizations uh like we've been seeing what's happening now in Canada happened earlier in Australia where they threatening uh uh well Facebook is threatening to uh close Canada off from their new sources if they're not paying the similar thing happening now pretty much to everyone if you have a good block or if you have a great content maybe really needs content what happens when you can ask it

1:51:39 and the answers are coming to you without you ever knowing who was the one who actually created you just care about the answers what's the incentives what's the future for the content creators nothing new Under the Sun so if you're coming from a small uh language um base uh that doesn't have too many potential customers and you're writing a book in that uh language uh and and then there's a local uh publicly uh organized and Run library system which is buying a few copies of your book uh that the rest of the population more less is buying so you are not actually selling those books so you are not getting revenues for writing uh all that content for the people to read what's your incentive of doing that if you're not looking for the fame if you're not looking for those other qualities uh boosting your uh carrier and using it as a calling card or uh similar meths why should you do it well you could argue and say that well if you want to write it write it and if not don't and that's a fair point that's always been the case so it's been like a hobby for many the similar thing with musicians there's only a few who can make the uh make living out of it and that's the power law but this is getting uh more and more obvious uh to the more and more larger part of the life which can be codified and put into the llms and it will spill out the answers and fine incre ingredients the raw material is all that human creativity uh which is then synthetized and um offered as a basic uh service for the rest this is a really good point although in those other use cases both the library and the search engine there was attribution so you were trading

1:53:42 Discovery uh in exchange for monetization you had more people discover your content sure some weren't paying for it and we getting something for free but most of them them would come in or many enough of them would come in and buy your content that was overall a win-win no such case with AI with AI there is no attribution there is no driving of traffic both libraries and search engines drove traffic AI does not drive traffic AI assembles it all together and makes it its own and that's why I think it's going to be a unique case that will be litigated and thought of differently by Society how to prove it yours that's going to be really hard as well because it's not direct quote it can be roundabout way the same thing uh in some really specific cases there's nobody else coming up but usually um contributions are incremental and combining already existing things so it's become almost mission impossible to prove that it's yours originally especially when it's a black box what comes to the libraries uh I'm just taking an example of Finland for example which is just plus 5 million people so public in any non-fiction books is already difficult even without the library system when you have a really efficient and good Library System people are assuming that you get everything for free they're not buying anything more or less unless is really popular and that's usually fiction so you cannot get there's no Market you cannot sell it outside of Finland well nobody speaks Finnish more or less outside of Finland so there is no Market there's no incentives to do this and uh when you have this type of culture that you get everything for free it's really hard to to turn the tide and and I think that's what's going to happen also nobody's paying for searches Google searches are for free same thing with AI but thing so it's getting really difficult I I would assume in the future and even if you're going to go to the court with the copyright infringement or that's what you're claiming it's going to take years maybe a decade and what happens in

1:55:47 the meanwhile well the market happens so you already lost one way all the other I really like this Amplified intelligence because uh if you're already an expert on something you can do so much more you know exactly how to do and and how to get the most out of it which you cannot do if you noob in some area because you just don't know your way around and this also means that a lot of other people can do the same thing what are you doing if you because they don't know either but if you already are really good at something yeah there are some others who can do that as well but this kind of keeps your progress uh advancement and for the rest uh yeah they can catch up but uh you can still uh be in the lead and and just do greater things be careful whom you're speaking with the your deepest Secrets might be going to the places you never figured out and that's a fingerprint as well which probably can be combined later on is that the V dream of Google and others who are having the advertising model I had a quick test run and it seems to be quite okay and nice uh the token size was a bit small but I think you can even change that so ask some questions like uh explain me the coherence and it gave some reasonable answer also what is 42 and it it got reference to the hiters GU but I guess these are pretty basic questions to ask but yeah it works a very informative dashboard for the far cast network how it's growing and what's happening in there uh the links would show the dashboard but here are something what's happening now uh one of the things which was really interesting was that there's like 13,000 total users uh out of which only plus 900 has been costing so posting more than 100 times um 157 new users but uh this is capped because uh there's there's no

1:57:54 open um sign up at the moment average cast per day 2K uh what else is here monthly active Cas is 3K so 3,000 out of 13,000 uh there's also connected addresses which is 10K so something has been dropping off total casts if I didn't know say 669 th000 I have some uh text on my website which has been there for I don't remember when I wrote it last year probably but at least the archive or uh capture is from the January this year so Mark Randolf who is the co-founder of uh uh Netflix wrote almost exact verba him in his tweet and it's quite a coincidence so I don't know what to think about it but yeah these things apparently happen and I'm not saying that he he was visiting or copying my copy it's just a coincidence why the large language models uh text is so good and then there was this explanation that it's using the most efficient language which is more sophisticated because you know the dictionary is huge so that's the the least resistance path and I was like hm is there Beauty in the llm model else is this the thing what David dos and Prett Hall are talking about what are you thinking I liked another explanation that was in the same thread where the writer said well this is David Deutsch's thing that truth is hard to vary so if you find 100 explanations for something the 90 that are wrong will all be wrong in different ways but the 10 are right that will be right in the same way so even if you take the most likely outcome it can often end up being the truth now in this particular case it turned out that that the example the original poster was giving was actually very easy to vary just by asking GPT the question

1:59:57 in a slightly different way you would get the wrong answer and then the original poster had to fall back on well the syntax is really good even if the content isn't so anyway just kind of passing along those two observations this sounds pretty amazing not sure though whether it's going to be open sourced or whether it's going to be still quite resource heavy did you hear about Google's paper on soundstorm um no I must have missed it what's what's it about well it's a parallel decoder for efficient audio generation it can even be used to generate dialogues oh interesting yeah yeah like this one was generated by soundstorm wait what but make a summary of this PDF here's the URL Bing make a summary of this URL here is the PDF through results were the same today I did it with b and um maybe one or two weeks ago I did it with Bing they are supposed to do that so they should be able to retrieve the PDF and read it but they were completely totally hallucinating just pulling some completely different whatever statistical usually people are asking these type of uh well these are the results people are usually hoping to get for type of answers and uh in this case a part also showed three different ops options and they were three different hallucination articles not too impressive it's like the Boston Dynamic CEO's interview in Lex fredman where he's talking about uh how llms are basically sophisticated uh curve fitting machines statistical machines even though they could perfectly fit what Humanity has put together in form of text and ideas in Creative

2:01:59 form uh yet how can it become conscious out of that one and how can somebody think that it is uh a proof that the prediction machine can do stuff by itself and it can have a will of its own refuse to do things and act upon like little child for example who needs to learn language as well without any grammar books or anything we just implicitly figure out the rules of the game how to communicate with other people and now we have managed to make machines who can simulate the same things thee when you already it's the good advice such a and Who would thought it's bigger the same thing has happened many many times in the history for example when we were mainly having agricultural societies and a lot of people were just working on on fields and and then they were suddenly kicked out of them and going to the cities and this is still the situation in China and some other places where a lot of people are looking for even the horrendous uh working conditions in in the factories but it's still better than what's in the countryside so this has happened over and over again and it's not going to stop and always we've been finding ways and as long as we have creativity as long as we have uh ideas uh things we want to uh create and there's a demand for whatever we come up with uh we will provide the supply for those things and that's what's creating the value value is subjective it's not about producing

2:04:02 things uh so um the last 3 months uh Boston Dynamics ce's interview he talks about that even now in the especially in Europe I think he mentioned it's very difficult to to find adequate amount of uh people to work on Lo Logistics it's just not possible to get the people there so it's getting becoming a real issue so the robot the the Boston Dynamics and other robots are going to take over those uh field work which was done before the tractors and other machines took it over so the similar thing is now just happening and I think it's just a really great thing uh but obviously the transitory periods are hard for those who are not so fit to uh adapt to new cult and and Technologies I'm not worried about kids at all um like di you mentioned they're so quick to adapt the things uh even me we we I'm not obviously a kid anymore but uh just playing around with the AI and all these tools and it doesn't take many minutes or I was like okay why can't you do this and why we can do it that way oh why this doesn't work yet and and oh I have to do all this thing myself so um there's also this demographic thing uh uh that a lot of the population in the at least in the west uh is is going to the wrong direction we don't have new people coming taking over the old jobs so we're going to have a sages of people soon so the robots and the AI are not coming minute to early to uh solve that issue as well have you noticed that uh a lot of the things we discussing and what's happening in the world require to take some kind of a principled position so we talking about ethics we talking about value values um or lack of them and that's something I think we going to be defining the next uh maybe five 10 maybe a longer time as well we

2:06:09 need to go back to those because they've been in the background for way too long and Humanity tends to forgot things it takes three generations to start the wars again when your grandparents are not telling you about the firsthand experiences anymore the same thing with the values we forget that actually we need to uh fight for them they need to be always on your mind and you need to be uh always uh educating teaching and living by them or they will be gone and uh now we have been tested in a way in a lot of front economical political even in security levels uh what our values there's no Switzerland anymore you cannot be just neutral you need to take a position and uh that's pretty interesting times but it's also very revealing times also a lot of true colors are coming up and also whether you uh swimming without if there there's a generational cap or it's not about age uh it's more like type of people uh between those who are using the latest AI Technologies and that's a really small amount of people who are really efficiently using them at the moment there's a enormous Gap also uh between those who well I remember it's co times when they started there was a lot of accounting companies just a lot of back office stuff which was done at office there were desktop computers so there was no laptops and these companies smes were having even big companies were having difficulties of making everything remote and that was sort of a shock therapy treatment for that but a lot of the things are going back to the way

2:08:12 they were even though things are changing but the remote asynchronous way of doing things going back to the mode of going to physical meetings traveling to different places with airplane and these these means take a lot of time it's not efficient it's not uh how things are done in the most remote only asynchronous uh companies which tend to be most likely software companies and others where they take advantage of the latest tools and aware of them so there's a huge difference also in the business life uh between those who are used using the the latest uh tools and using the latest methods and those who are just still basically back in the uh 40s 50s 60s ways of working uh it hasn't changed that much obviously the school system is in the same level as well I mean it's on GitHub so it's headed for some sort of Open Source who knows which license model but also it looks incredible it's really fast speech generation I mean they talk about the matching um they talk about the matching piece but it sounds like the generation for like 30 seconds worth of speaking is 2 seconds on uh TPU 5 symbol V uh so this looks awesome # symbol uh sigil symble sigil conversations needs to happen more meetings never I think that's a great line another notion from the segment and I intensively kept it long because there was in the end a a slip or I don't know maybe it was a plock about the automatic word presses new chat gbt feature which is not public yet so if anyone has any information happy to hear put your comments in this doesn't uh project confidence uh let just say November 15 2022 a firm update cannot extract

2:10:16 private keys from the CQ element and yeah yesterday 17th of uh um May this year technically speaking it's it is and always has been possible to write F that facilit key extraction you have always trusted let not to deploy such firmare whether you knew it or not I guess there's this saying about uh digging yourself deeper and sometimes it's better just to stop maybe that applies here some of these uh tgpt API rapper apps are making hundreds of thousands per month and obviously it's a fat it doesn't last but it's still uh good money and some others taking the opportunities available isn't it getting a little bit expensive uh when each and every time you're sending the prompt template which is in for example snap's case like 1 A4 it's a long long list of things and it doesn't even include yet the history and the actual question the user is asking and since the as far as I understand it's stateless the API so you have to remind it and provide the context each and every time this doesn't answer the previous question about prices but gives some ratios and and some ideas and I will put the link on the comments so you can check the source yourself my guess is that you just focusing on more on the AI because there's so much happening and it's so hot space so every hour seems like there's a lot of things happening but yeah things are moving personal Voice is coming to Apple later this year so you can type and it will talk with your voice maybe soon in air chat near you the Supreme Court ruled that Andy warhall has infringed when he was using a photograph of Prince the artist and that was the basis for his artwork which was too similar and fair use did not

2:12:20 apply for that reason because he was famous he still needs to stick by the rules and why this is interesting llms if you're training with other people's art and if you're using these llms and generating pieces of output I wouldn't call it art and they are too similar as well are you as a user in fringing as well for the productivity nerds who want to use uh Apple shortcuts but always been cursing that Mac OS doesn't allow the automations which are available on mobile devices well here is the thing which makes it possible uh it's it's a paid app but it's 10 10 bucks so totally affordable and can make your life easier with the time based or for example if you're pluging in some USB devices or take them off or some application becomes active and so many other triggers you can use to uh well make your life easier if Instagram allowed anyone to become a photographer share their pictures even though they were not really good or they didn't have a professional gear Twitter and Facebook allowed you to express what you feel and share some moments with your friends and family and strangers air chat is giving voice to people if if you're using social media for conversations like here in air chat you care about the discussions that's the main purpose and the benefit you're getting who views it who listens to it it's a nice additional bonus but that's not the main goal if you're looking for attention the content doesn't matter that much you just want to monetize it you want to gain power or influence or some other other means and there's a lot of that happening in Tik Tok attention videos a lot of places doing this then if you want to broadcast your ideas uh or whatever you putting out there

2:14:25 Instagram pictures or tweets then it's one way communication you just want to State your piece and that's it so there's at least three ways you can as a content creator use the the media then I'm sure there's also similar way you can consume it totally passively you want to just contribute some way or you genuinely want to want to actually further the discussion or the conversation if you export your air chat discussion it's not the same conversation anymore you can put it in Twitter you can have conversations there but it's different people it's different format is it doesn't have the same feeling it has completely different context the same thing if you put it on the YouTube people can watch it people can comment it but it's then broadcasting it's not the same thing so the place where the discussion happen really matter as well and you cannot expect to get the same results from different places where you put in the content even though from the outset it looks like the same thing what do you really own I mean I have a voice you can hear it it has particular sound and personality I have my thoughts ideas which are in different forms All Around the World in different places in websites and air chats and tw Twitter and Instagrams and people's private archives I have my DNA which is also all over the place there's a lot of things we're leaving behind and when we leave who owns those yeah I can State something but uh Shakespeare stuff is still there there's a lot of stuff living forever and now there's also the voice prints and the appearance and all these things which can be used for whatever reasons later on so do we really own our on habitus whatever composes of us or are just

2:16:27 basically creating it for the rest of the world to use it afterwards a is your Muse your secret weapon if you use it that way you go into rooms you find some fascinating discussions you jump in then you realize something you haven't realized before because you needed to think about it and that just inspired you to LEAP into some direction you haven't been before and now you have a new idea which you can start to explore in this episode David dowis uh talks about the example of a tennis player and the decision he or he needs to make when returning a ball it's not a conscious thinking because there's no time for that but it's not a reactive uh reaction either because it's a it's a decision which made needs to be made and he talks about that later on that player has a story about it so uh they can talk about it like yeah I know the opponent usually does it this or that way or whatever you know the reasons behind it why the response was what it was but that's not what happens in your mind and there's no time for that so where does it happen and I think this is the interesting part this is the Quantum Computing part this is the part which is I haven't seen too much people discussing but I think is the key for our Consciousness and everything which we cannot explain at this point and it it's a bit well I would love that um there would be David do would go more deeper in that discussion maybe it's coming in the next episodes I hope so we're talking about these things it's so easy to uh go to the labels yet we don't know what Consciousness is so it's it's almost like redundant to talk about it but I think the important Point what he is making or trying to make and I'm not

2:18:30 sure he actually makes it clear enough for for many people to understand maybe I'm confused as well but about the implicit and explicit part and I think that that's really the key but it's it's so subtle when we stuck into the words like uh and taking the example of our mind is like yeah subconscious conscious but we don't really know what those terms mean either uh so yeah we have some context we' been putting those things but it's not the Consciousness it's not the awareness and I think the really brilliant part is when he's talking about when you don't use words but you still can communicate so even speaking words thinking in that level is not needed for that implicit knowledge which can then be made explicit but that's really the important point and uh that's really deep rapid hole I think we should go as soon as possible because that's the key for the things we really missing at this point here's a little thought experiment on quantum physics so if you start from the premise that everything is known there's no uncertainty there's no risk everything is just boring in a sense that um nothing flows everything is at the end state from the beginning but you would like to have the smoke and mirror show The Sensation that and the illusion that things are moving uh there's uncertainty you don't know what's going to happen next how would you do it wouldn't it look like uh the everan world where there's a Multiverse uh we only have the one collapsed uh state which we see and the rest is well in the wind in the Multiverse so it's there and and that's basically how you

2:20:35 do it so it's just an illusion how we experience time and the flow in order not to know the end game from the beginning aat creates his own problems uh you know know the feeling when you're in a cocktail party and you are in a range of uh two three groups so you can actually follow three discussions at once well here's the same thing if there's like three people actively uh dialoguing but they need to do it as so uh the tits go one after other so there's discussion one the next one discussion two the third one is discussion three and uh you just want to maybe follow one of them so how do you do that or if you're following three of them and they're completely different they are even comically uh separately differently intervin it's getting confusing related issue so if there's a huge threat while you were sleeping and now you're catching up and there's so many of them and it's a long discussion going going for tens or maybe hundreds of chits so uh if you're in the beginning of the infinity or beginning of the chat uh threat you haven't listened can you comment on something while you're listening to without knowing what's in the next uh 100 trits and where the is discussion is going uh and should you put it into the reply section of the chit or can you put it to the main thread if you have the access rights so what is the best way or is the best way or it doesn't really matter this is also the consequence that if you have a few of these uh did you want to reply which are further along the path into into the past maybe few hours or doesn't really matter and then you as a Pur mode reply to many of them for those people who are already at the end of the thread uh will just see that you are spamming the thread with a lot of um individual chats clustered and there may be even like a cluster bomb in a sense that they are replying to separate people in separate discussions within

2:22:37 that same uh chat when centralized media is dead well that's already no news uh we can trust a few part parties uh this is what we need to do in the past because technology wasn't Advanced enough and well we are having only a low band with ourselves so we needed to trust something and get the news from somewhere but now everything is fake everybody is producing content nothing is through you can trust any of the centralized sources so we have to go to the uh Netflix mode uh turned into a news Flix where pots uh AI is gathering from the fragmented sources verifying sources by the authentic visting uh the source and also double checking with thousands sometimes millions of info pieces coming real time in and uh distilling your perspective and take on things without trusting any particular single item from anywhere or even dozens or thousands of them but just getting the aggregate view of things that's a great question and I think that here's how I think about this I tried to record like a long chit and then decide to dismiss that that draft but here's my uh next thinking when you reply horizontally which is to say in line then your reply cheat gets added to the end right so if there's 100 chats between what you're replying to and uh and your eventual chit then the reader let's say from the side after a year will only get to your reply like after a long time whereas uh if you apply vertically down thread you you're freezing that reply in time so the reader in a year will get to that chit and then see your reply if they click the replies button and they will because they're interested in the whole conversation vertically and horizontally and they will see your reply right there and then uh the additional benefit here if the author who you're replying to decides to reply back to you what they

2:24:40 should do and maybe this should be like a more straightforward thing they should restitch your uh reply into the main uh horizontal thread and then immediately respond to you with that like connection of of the reply uh this way when the eventual user after a year gets to this cheat they can then skip to the reply and then they will hear your reply and the the authors reply like together in sequence um so that's still very long but that's how I think about this that's the way to go the only downside is that not many people are checking the comment and uh you can only use a few of the options um live video audio you cannot uh put links or text and uh you cannot generate the background image uh but I think these are coming later on so yeah maybe that's the solution at this point comments are terrible we need to redo them entirely and make them much more palatable much more first class citizen if there's room with the eventual rehaul um other platforms have eventually converged on emoji reactions versus just the just the like uh they give a lot more freedom in expression for folks without overloading the chat stack necessarily with like replies because people would reply with that emoji and given the option to like Emoji react they would just react with this Emoji leaving the replies for conversation uh but it's tricky to get right especially given the Simplicity of the current UI yep it's on the list here's an idea for the hopping back and forth and uh dedicating and knowing which thread is what or within a single chat with multiple cheats um you need to when you're replying or making a new chit you have to uh tag a person you have to put it to a person if it's in reply to something if it's not then you have to

2:26:43 make a tag so you can actually have a new topic so either it's a tag or in reply to and there you go now you have always something where you can go back and forth all the options are there but we can't make it that mandatory there's a new reply to interface coming that's more flexible and more powerful but again we don't want to make it mandatory get rid off the comments and make them the threads so they can just start to stand on their own uh feet and start to have their own threats and this way the conversation can go on that's how it happens in the dinner parties as well in in cocktail parties people are going from one topic to another smoothly without any trans transition basically and this is how it I think should be in as well technically should be possible as well yeah I've been thinking that too which is each comet is its own room or each Comet thread is its own room but it gets messy pretty fast you can get lost very easily and uh then nesting within nesting do you keep spawning off Inception rooms every time how complicated do you want to get this is the perfect opportunity for Microsoft those who actually were with the windows 3.0 95 uh and all these older versions and the blue screens and whatnot they are already old and the new generation just have experienced maybe Xbox or some other stuff or they see the Bing generating AI images and whatnot and hey Microsoft is cool and they haven't been in the team's trap yet you already have a Threading withing r rooms chats and that's tax so if you're using tax you can play those tax and basically now you just uh represent all these individual tag lines given that people are using them uh as uh intertwined but if you just unbundle them and have them as a separate uh threads I don't think that's a bad idea I'm going to have to think about that

2:28:47 tags aren't getting much love but yes if everyone was tagging everything all the time correctly it would work I see it mainly as a UI issue so now you have the uh chronological order which is mixing everything but if you just turn it off then you have the threading option where each and every conversation it's its own thing so you can consume it from from the beginning of the conversation till the end and if you have done that then you just come back and choose another thread and I I I think that's uh the regular way how a lot of these other uh applications are having the threats going as well whether it's a emailers or some group messaging systems and stuff it's just that you need some kind of a subject line which is compulsory or some other way of keeping the topics together rename them as topics or subject ts are great I I really enjoy the fact that you can like filter by tag essentially I think the it's not super clear that consequent consequent chits are under the same tag until Changed by anybody in the room that's not super clear uh and also when you're viewing the second chit onwards from a change tag then it just says new tag like it doesn't show that this is kind of under same category because it can assume it cannot assume that right unless there's like a direct reply so maybe a post cheat reply selection like if I just posted she is like oh actually I meant to add this as a reply to over there somehow to maybe being able to connect chats after they've been created kind of like we can stitch cheats to different chats but maybe like attach cheats as reply comments post the recording flow we actually already did and my build is called topic everywhere maybe

2:30:49 also on yours just check that feature is already built out shipping really soon just putting the finishing touches on it like everything it's more complicated than you might expect so you boiling the Frog slowly next is make it compulsory to have a topic M really don't like compulsory then it's not time to do a chit yet if you don't know what you're talking about and you cannot label it yet not even the slightest becoming good at something usually takes a long time so it's a long game doing things yourself practically knowing how to do things is a pretty good scale often times if not always better than trusting on other people's opinions uh or secondhand knowledge when people are taking out of their family routines and confronted with real human needs the best aspects of human nature come to the Forefront this disproves the negative assumption that humans are inherently selfish and demonstrates our natural abil ability to emphasize and connect with one another Rebecca solnit in Paradise built in Hell uh paraphrased in this article which I recommend you to check how does this sound with your AET experience so far AI Builders are the new redu ists that's the thought I had when I was looking that chit so if in the 20th century people were splitting atoms hoping for finding the smallest components and then there would be that c particle which makes everything thick and work yet we haven't found anything just the Le of blocks which doesn't become live when you do something with them the same thing with AI no matter how much automation is put in there it doesn't

2:32:54 become the living and breading machine Japan is taking the stand that it doesn't matter where you're getting your training data for the llm whether it's illegal or legal sources copyrighted material somebody else's intellectual property it's all the same spoiler all it we're going to have a lot of boring content coming out no I don't mean Hollywood one I mean automated llm generated animated content is already starting to appear in YouTube just check it out I haven't tested this yet but it looks pretty promising you can just drop your documents and uh off you go without need for additional gpus or expensive Hardware might be slow though yeah it's a great consumer app I have this running with MPT and vicuna I believe waiting for them to Falcon they said they're working on Falcon and it's pretty good it's pretty good it's not so slow as well surprisingly fast on an M1 oh and also they just released plugins and they have a checkbox in there that turns it in the server so you can actually kind of you know replace your open AI code that calls gp4 and call that directly uh it's not great for everything but it's pretty cool to play with is there any particular use case you would say that it's it's really good and you can already replace open AI with with that one maybe quiring your own documents or something which you don't want to leak out the rest of the world no not yet so the use case uh is likely just having it on your local machine and playing with this it's still the set truth that you know gp4 is way better than most of this however they do keep updating with a bunch of stuff all the time so if you want to try out different models you want to get a feel for llm like a difference between a vuna and a falcon whatever uh this is a great way to just

2:34:56 run this without messing around with python environments and and collabs and all that jazz um I don't know if there's a very good um use case right now the article is nothing new new and fantastic in itself it's just shows you how complete your profile might be in the worst case scenario check it out if you like and input your email address in there but um I'm just more thinking that we should scatter so much info about us uh like Amada was saying the other day that uh you have to input fake information so that the real one doesn't stick out so I wonder how how we can do that and are there already some llm spots who can help you to do that and are there any legal consequences of uh spreading misinformation about yourself in order to protect your yourself so is this self-defense when your old airat content reappears in new rooms and it stits there and it's emerges for new audiences or the new ones are digging out or maybe the old ones are digging out and discovering new rooms that's a sign of good quality content if you like to uh get some piece of History here is U fa HX uh personally assigned copy of uh wrot the Ser serf D to Carl poer and uh it said is fellow struggler for Freedom with friendly greetings from fa haak so I think the starting price is 400k fyy this is an example of something what happened to me I was listening to this acquired podcast for many years and I always did it in 2X and then for once I realized that I never heard the song actually how it's supposed to be played and here you go who got you is it is it is it got welcome to season is it you is it you is it you who

2:37:00 got the now is it you is it you is itra welcome so which one is better the init knowledge which is not cified it's not in the any U tangible form that the llms can use is the gold standard basically which required so a lot of the things which look simple are actually just uh simplifications based on complexity which is earned by experience knowledge and wisdom by either practicing the the field or some other way acquiring that knowledge and these are not in the llms so they cannot do the same thing they can basically predict but it doesn't mean that it's actually correct because they they are approximating something which most likely is not exactly based on the right fundamental analysis of it it is like real life but without interruptions there's so many things going on here uh that I think we can bundle it the whole next thousand chits uh but I just give you one idea that it's also a generational thing um people who are old enough to remember that you actually have to pay for each and every photo if you develop them so um you know it it cost to take pictures they still uh store and keep all day photos usually and they they have that mindset but young people they just take snaps and they don't care about it and and they can take thousands and and then just pick a one choose one and and throw the rest away they are in a completely different mindset the same thing here digital is more tangible than than the actual physical so they treat it differently but it doesn't probably go as far as uh thinking as as a asset or value it's more like a natural thing with different meanings than what we

2:39:04 used to bundle and I also tend to think about that uh we here have all these multiple meanings for things but most of the people are just completely ignorant of these things they just see it as a one thing and maybe just something they throw away tomorrow yeah the concept of notar is is nothing new so in that sense the proof of uh Providence uh is is an old concept and uh yeah it shows something but obviously it doesn't prove that it's not a fake or anything else it just proves that you know it existed at that time most of the people are not using strong passwords most of the people are not using uh second Factor authentication they're not backing up their data it's not that deep so we who no technology we understand how it works how they just point us for nothing or there's nothing underneath it or technically doesn't make any sense that knowledge doesn't exist for most of the people they just buy those things without uh that knowledge the ignorance is place I think this Academia is just too slow and uh it does doesn't work in this environment stuff is happening so fast that you actually have to be in the middle of it and you have to execute and you have to do things yourself and I think this has always been the way uh David dos U Einstein they were not exactly in the center of the Academia when they were doing this stuff the other thing is that uh deep knowledge that's specialized really deep but the narrow knowledge is also the specialization is is something which make get you in trouble in these days as well so in a sense the generalist may work better if you actually have the Hands-On knowledge uh so you can actually apply and use it to the to the new fields and uh do that but but that's where the shallowness comes in if you're just uh superficially hopping from the way of the next thing and maybe you have a deep knowledge of something but it's not applicable to the next one it's using different kind of knowledge space then you in trouble and I see that quite a lot nowadays there already a lot of signs people been saying that they've been getting better talking and practicing

2:41:06 here uh you can find those in the some of the rooms uh who were shy before didn't speak that well and now they're articulating and I can actually come up with at least three persons who've been saying that in the last two weeks so yeah yeah just participating you become better uh at at this so and and this is not a competition you can participate or not uh I I think people are always social animals we want to look good to others we want to be articulated we want to be heard we want to be expressive we want to be understood nothing changes that but those are skills you can become better I've already seen some comments by Master users because that apps would have a Fed verse uh activity up support uh they're saying that they will block the instance so they already t b it's a good idea or really bad idea to meta to to come to the mustard on and decentralized networks game so it's going to be interesting times yeah the problem is that uh those people who are actually visual uh that type of people are already in Instagram and now Ina is saying that hey either we want to pull all the people from Twitter uh who actually want to get rid of Twitter but the social craft is not moving to come to the visual place or then the other option is to actually have the visual people turn to become like idea people and I don't think that's going to work then there's Facebook which was in the early days at least more like a text based thing as well and and it has more tendency of sharing text than being too visual but that's not how they're going about it so uh yeah I don't think it's going to work there's also the last option uh which is WhatsApp which is mainly text and a lot of people are using it but they're not spinning It Off from from WhatsApp which I don't know just now thinking about it would maybe be the most neutral place to start something off I think there's something into the

2:43:11 projector thing especially if you can project in your palm for example some kind of a touch screen uh in a way that you can tap something and and do primitive things so you don't really need a device then obviously we have our iPhones on our phones and in that way it's a bit redundant but uh given that we wouldn't have a phone anymore uh it's liberating if you can just talk when you know we meeting in you know in person we don't have any gadgets wigets we just talk and see each other and it's all buil-in devices using I like the clubhouse uh first experience where the volunteers Clubhouse enthusiasts were running these rooms and helping each other out uh that that was really powerful in the early days um the community was welcoming and it was fun um a lot of people left so the experience and the magic disappeared and that was never the case with when Spotify launched they live I think it was now called now they just close it down as well it was about was it diamonds or something and that was a horrible experience from the beginning so when you're going into a new environment uh culture is something you observe and the behavior as well so if you grow too fast it's really difficult to keep the this uh in invisible boundaries these ways which tell people without explicitly telling what is okay how do you be behave here what is the um nice way of communicating and interacting with people on no there was this similar um nice fuzzy feeling in the beginning and was it early March when there was uh some mustard on trolls uh who want it on purpose to make havoc and they tried to troll really hard and

2:45:17 uh well cause cause disharmony and I think uh the community managed to turnone them down and probably they left or they were left their own devices they didn't find the the reaction they wanted and it it went down the same is now happening in Blue Sky partly it is that I don't know whether Us app build are attracting the people or the people are choosing um the app and you just have to partly hope that you get the type of people you want to get in there because if it sets the wrong tone you may never recover like in the case of the spotify's life seeing sounds as colors is an interesting I would say superpower and uh I had a chance to talk with the South Korean uh world famous uh classical free style pianist uni and here's a tiny snippet uh from the discussion I had with her the thing is um I was in the kindergarten and then there was a piano but then I just played it there it was so nice feeling but at home I didn't have a piano so I put I just draw in a long paper and then I colored every Keys then I played on the piano and then I imagined how it would sound like and then the next day I played on the real piano in kindergarten and then it sounded great oh yeah it sounded like as I imagined and that was the really fun time when did you make the connection that you can see the colors and you also have a perfect pit was that from the beginning the very first time you draw and it was yes yeah so uh my hobby was every sound which I heard it from the kindergarten bus or like um animation theme uh songs or TV uh pop song from music from TVs anything which I heard I always wanted to play and then it out as I thought and that was really

2:47:19 fascinating so when you hear something maybe bird singing you can easily reflect and do that yeah yeah and then when I was like seven and some piano teacher told me that this is called a perfect perfect pitch and okay are we actually now living the last times of really great uh games which are basically written and everything is human made more or less before the AI will take over everything and it becomes more plastic in a sense that uh something is missing because it's just generated and it doesn't have the same feel I hope not but uh just an example that this game came out 2015 so it's been development earlier on and and yet it was quite sophisticated when it came out and it's still nice to play I wonder how the AAA games are going to change in the future due to the AI um how the story lines how the characters how the depth of the game but also the gaming Styles can change and and how how they're going to look so much different in in let's say 5 10 years time and I I don't mean just the graphics and you know the Outlook but really like the does the AI change the fundamental nature how do you uh work with the games are we going back to the 80s where you prompting just an example from completely different world uh this is an augmented reality game um where you can skate anywhere with the physics engine and uh you can change the scale and and you can you can basically imagine what you want to do and then you can accomplish it and it's not limited to to the VR world it's basically everything and everywhere around you and you can also multipay uh with different people even when you're not physically in the same room so what happens when the AI is completely combined with this and the compute uh enables way more uh surroundings and understanding of the uh

2:49:24 environment in a way that uh it can really start to interact I don't think it's about features you can add all the features in the world and it doesn't make you the the app or the or the the killer app the app presents something Facebook is for grandmothers nowadays it's the Boomer app no cool cool kid wants to be in Facebook snap is still I guess some what's cool in in some circles but those are probably already in their 20s Instagram is for certain type of people LinkedIn so the app presents something it's not just the features the features doesn't make the app it doesn't make the user experience it's a statement what you're using as well and sometimes you need to start fresh meta as the whole thing with all its Legacy and all its apps presents something and it's also something not everybody wants to be associated with everything is fake indeed and even your relatives are now giving excuses why they not showing up on your functions Madame Claw is the latest superstar in the meme world so check it out uh this cat from the KE bom shelter is really making it round at least in Twitter I wonder who's going to make songs out of this you just mix and match the individual cheats uh maybe in random mode maybe in some context and make songs about it also all the new audio books uh the best of uh this and that chat room by these and these characters all the best of March 2023 in this and that particular topic a lot of Readers Digest coming up I wonder how rap paddles would work if you have to wait for the opponent to give the next rhyme and you have to hold back uh it's like test moves you have time to think about but you have to respond but you cannot just keep on going also you could actually uh play turn-based games even chess uh with air

2:51:32 chat maybe even visualize the thing uh if it's specialized enough fun times ahead in the last 10 years there's been so many YouTube artist and you can actually see them growing up sometimes they are teenagers maybe even 30 13 14 when they start their career and now they've been already 10 years doing stuff and you can see them grow Their audience grow their talent and becoming even famous and and getting recording deals and making their living out of the music this just came out uh it looks rather extensive already but not sure what to think about it yet because I don't have any personal relation to the company yeah CH is a conversation app for ideas Apple mentioned that they did some 5,000 patents and here's a really interesting um insight about for example that before you click something or you do something uh basically they can already read it what you're going to do uh check out the the Tweet is a long one but in the middle there's a interesting stuff how Apple actually has been uh trying to figure out what we do by our perceptions uh by our gestures and things like that and what they have figured out hey so while everybody is looking at the Vision Pro thing I just want to read this uh I don't know why this is the room I chose it says the latest and this is the latest a Former Intelligence official turned whistleblower has given Congress and the intelligence Community Inspector General extensive classified information about deeply covered programs that he says possesses retrieved intact and partially intact crafts of nonhuman origin what the actual I believe it I uh I have a video actually I should upload it of uh those those black balloons that were recited everywhere my friend took a video of them way before the story came out and it's like hard to Fathom what the hell they are I got to ask my

2:53:34 friend's permission though I have I have the video extraordinary claims require extraordinary evidence I don't see much evidence for this being true I think it is much more easily explained aable in many other ways yeah I agree uh that article went hard into his credentials and how much he's like a trustworthy Source uh but I haven't seen too much evidence Beyond his credentials and waiting for more but wouldn't be cool though to discover today that you know there's life in the universe could be cool we all have 4K HD C in our pocket Now show me the videos H res please we got denied damn it now the alien's going to get him before he has a chance to upload he needs a dead man switch somewhere if he shows up acting all happy and strange check for lizard skin underneath can we get meta data about this video at least what it Shou in 4K HD like noal requires or is there some grainy [ __ ] that he's going to release for some credit internet points here it is um it's not 4K he might have it on the original I don't know that but this is what was sent to me through iMessage back in July of last year this was many months before the whole uh UFO thing that was going on um the thing I'll say is like the objects were so far away that again it's one of these videos that you can interpret a bunch of different ways but it's definitely a super weird video when you guys see it but I don't think anybody is going to be confirming its aliens uh this is pretty big uh consider that almost 50% of all the websites are done with WordPress and now it's available for the wordpress.com but also I guess with the plugin you can put it to any website using WordPress 20 credits for free and 10 bucks per month this is pretty extensive thread I haven't read it all

2:55:38 yet but it's worth checking out it's in real life antisocial thing to have that on and and you not having one so you will not want to communicate with people who are actually having those on so you have to share the space so in that that way it is the factor when you're alone or you're working or you in some productive or some other mode which is like don't talk to me mode like every new catchet you love them in the beginning and you use them for the novelty and then it will wear out depending on the on the person whether it's a day week or month but eventually it will happen anyway so then you start to think about that uh recording this [ __ ] should I put the headset on or should I just do it with my phone what is more convenient what is more easy and this is how usually it goes so before it's actually like a eyewear or something you can forget that it exists it becomes like a seamlessly part of your existence your hit of stand basically uh yeah you don't have to think about it it's already there but before we are in that stat maybe it's 5 10 years out I see education as a huge opportunity like part was mentioning how to play piano but it can be anything basically arts and crafts where you have to physically do something and it's basically mimic making or you have to follow some instructions you have to break something apart you have to build something which requires skill and usually this is done with tutoring somebody showing you how to do it or looking after you and and and guiding your way now that can be done automatically also in the schools uh it's easier to demonstrate things and and do things so allog together I think the education self-learning is a huge Market b2p um maintenance services is an area where AR is already now being used quite successfully and it's reducing costs you don't need to send people over specialist uh from Another Side of the planet uh to fix something which is Trivial in a way that if you can just guide the person who is not a specialist in the field so they use it in the remote service centers AR tools uh

2:57:43 having a view for the local uh spot and uh then guiding The nonp Specialist how to do the uh proceeding whatever that is I don't believe outside of tech sector that the regular zooms and teams uh going to go to the VIS on arway anytime near soon uh the reason is that if you look what happened to the co uh remote working time a lot of people didn't actually have a proper cameras they didn't have a proper mics they didn't have a proper setup even though that was the primary place of work and they needed to adjust and it was a good excuse and not even excuse to upgrade their stuff and maybe the company even compensated for all those upgrades you you still have people who are not actually properly using the cameras mics and the audio is horrible um so it's going to take a long time but I don't think the incentive is there for people to to do it better if that would be they would actually buy good mics uh for the sake of the audience but even that is not happening if there's one use case which might be interesting and and this is what I was wondering how soon we start to see people business Travelers using uh Vision because they're on the road can it be actually the next privacy thing you don't want to expose your laptop with all the cameras and everything around you don't want to maybe type with the keyboard because everything is tracked and seen so if you can just use the vision and and uh nothing leaks out you don't need to speak at least with your uh outer voice if there's a way to to speak without uh making sound maybe that's even a way to to be in the in the zooms and meetings but I I think that might be interesting use case which can be used by the sales people people who are on the field traveling and uh that could be something as well where AR headset even though you may look ridiculous but still it's like

2:59:48 you working alone uh and it has the benefits because you don't need to carry that heavy equipment with you apple is coming to the gaming finally so direct X12 support without virtual machine allows um a lot of the games which are existing to also be playable in apple machines here cyber Punk uh 277 running on uh M M1 MacBook 16 GB if you want to click the link any of the trade skills where you actually have to use both of your hands and physically do something I think that's a tremendous thing where where a lot of things can be uh improved it can be obviously something like a surgery it can be something which is just Plumbing it can be hairdressing it can be basically anything where you need your hands and the virtual testing or practicing is really handy but it's Al like in education I can imagine that you could just uh visit the the solar system put atoms in place and do a lot of things just by visually having them in in front of you so I I think that there's a huge potential in this uh in some particular field the Disney uh collaboration cooporation there were already somebody justes think that Apple should buy Disney I'm not sure if that's a good idea I just come to think of Nokia in the early 2000s when they started to compete with the customers with the operators maybe that's the same thing if Apple goes too deeply to the content creation being the manufacturer of the devices finally when we start to have a good audio books uh read by AI uh they may become redundant not redundant but the next thing might be actually storytelling in AR um I already heard about um some theater Productions and these type of things happening in the covid times but imagine that storytelling plays can be animations any type of content can

3:01:54 become live even your history lesson in your living room or wherever you are and I I think that's a really fascinating new chapter as well is it actually the next level of uh consuming uh your your content fiction non-fiction content as well apple is in a really interesting era now uh with the gaming platform Unity cooporation also with the AI created um movies animations which are coming alive becoming full full scale fullblown uh trailers short stories uh Netflix type of a series Almost in a few years out of a text prompts or whatever the multimodal first uh origin source is and I think this is really fascinating if you start to see all these things things converting to AR space this Runway version two model is uh quite nice looking already so text you prompt in and the video you get out it starts to get to the point where you can tell stories and they start to look like regular movies yeah maybe it takes a few Generations still to becoming actual realistic but now it already is uh good enough to to use for some purposes 3D capture on your device um it's easy to see why this is uh becoming a huge thing well with in Pro surprise surprise so you can take any any object in your environment you can scan it in and uh then you can work on it with your vision we are getting there when there's going to be hundreds of uh millions of these objects scanned I guess llm models will help also to produce those 3D uh spal videos as well just a few observations while listening to the episode he repeats so many times the safety of the language models that um I don't know it's just the only thing

3:04:01 I almost remember from the episode The another thing is that he emphasizes so many times that uh friends and being close to friends and basically the early days of Facebook if they ever was that it's a thing where you communicate with your friends and it's important to to connect not just in a superficial level but hey the whole business model is to monetize the connections push stuff to people use the content and the people as a means to chunk more monetization out of everything happening in the interactions so I think there a fundamental um this connection here then in regards to the smmes how it's so important for the small and mediumsized Enterprises basically every business in any level to have a access and possibility to do stuff with the aots and llm models and yet he did not um open source the the models he actually denied for each and everyone access to use them commercially meaning that they need to these smmes he cares so much about needs to use his products or the big cops products because you cannot use those models commercially so how how does that work for the Innovation and for the little guy he cares so much about yeah he says that it's about the safety to putting the models out but if you release it as a MIT license isn't it like hey we don't take any responsibility of this this model do whatever you please with it but it's not our fault if something goes wrong why cannot they do that they have all the legal persons and all the money in the world to uh refute any claims uh about releasing it as it is text based media he just touches the new initiative

3:06:06 within the company uh against fetty ver muston and Twitter type of content but it's interesting that he focuses on the um medium text not the content per se when we now start to have multimodal language model so it doesn't matter which The Source format is and you can switch between the formats uh like you can have text and it can be converted into speech or it can be converted to animation similarly why Focus so much about that it's in text and not in the substance itself that's interesting and maybe telling something this Ben Thompson's article is a high price for Apple's Vision Pro and it's worth checking out because he actually managed to get a hold of the device and use it for 30 minutes in a controlled environment but what he's saying also is that Apple claims that it takes 12 milliseconds for our mind to realize that something is not instantaneous so that's the hard limit any compute needs to happen before that uh 12 milliseconds is Tick tocking this is a subtle shift but it will change a lot of tools a lot of workflows and a lot of ways of thinking if you start to think about and create first 3D object where you put the skin you put the characteristics the variables and then if and when needed you can generate 2D images and the a similar sift in perspective is happening also with all the content you could say that now we're going back to the atomic World in a sense that first there was an idea and then you think about what's the best way to Manifest this on this particular case is it text is it like a tweet is it a picture like

3:08:10 in Instagram is it a video like in YouTube is it like a story story like in Tik Tok Vis Pro or the further Generations enabling it it becoming like seamless experience that you forgot that it exists plus the line between what is real and what is virtual if you can project things which look like real into your perception of the reality but the other way around as well if you can capture any object any event any sequence of events and you can save them modify them remix them match them manipulate them this also causes Havoc with your memories because visual is so much how we experience things and if you see it seeing is believing it's easy to memorize things if you can walk around them you can perceive them and you get a strong emotional uh attachment to the event somehow it raises some emotions some excitement happens so this imprints the memory yet it never really happened if you had a neural link or direct connection without any external catchet eyewear in you then it's becoming even more I would say hideous what is real what is not real would you like to have a roommate to live with pick and choose the character you like pick up the trades you like how

3:10:14 many months would you like to have a roommate subscribe here these are the steps to the Matrix type of reality which feels real you are immersed into it you can experience the wind in your face you can feel the touch everything is so real is it like a chit like air chat air chat is the new content rapper one idea one [ __ ] autois ideal content rapper let's see whether we can keep it that way or it's going to be inflated bloated [ __ ] in some sense it's like uh Nokia building excellent Hardware but you already entering through the software world and uh that's where they're not good at so the meta is in a game where you need Hardware you need your customized chips and not even that is good enough because you need the Deep integration with the things only Apple can at this point provide even Google is struggling because they don't have the whole ecosystem and and the value chain in a way Apple can integrade check the chit first for the explanation and uh after that uh here's a little bit of more about the second uh proposed law of uh Quantum complexity so in essence uh there are five stages uh where the first one is initialization uh when basically things are starting to get bit more complex uh thermalization is where particles are bouncing and uh uh stuff is getting to the thermal equilibrium and after that when that's reached uh things doesn't stop there there's a comp complexification as it called and it's getting more complex but uh still the entropy is invisible uh so it's a Quantum effect happening and then comes the stage where the complexity

3:12:18 equilibrium is reached so there's like a saturation point where nothing can be uh more complex anymore and then the last stage is recurrence where it basically collapses back to a simple state which is really um against all the odds and then the whole thing starts again rinse and repeat so is the second law of quantum complex complexity basically about the distribution of the quantum effects um when there's no time so everything already is it's just a matter of uh and we we know that there's no uncertainty either it's just a matter of the distribution where each and every of these states is in which uh Multiverse uh universe or Multiverse uh segment everything is and that's when from our perspective it looks like time it looks like a transformation uh when we going through these different uh slices but uh from the total system level it just is and in that sense um the complexity is also saturated and basically you can say that from the total system perspective it's always saturated to the max why it's going to be pretty hard to do the hi no matter how much compute we're going to put in there here's a segment is the difference between inexplicit ideas and explicit ideas according to you so it's very unlikely that ideas like a painting are stored in the painter's mind as bit Maps like they are in a computer they are stored in some other way which we don't know but then we don't know how anything is stored in the brain it's certainly not the same is how we do it in a computer so the both the equation and the picture can be

3:14:20 stored in a computer when the equation goes into a computer it can be translated directly into something that a person can read in in the case of a picture it you have to see it to to appreciate it I don't know you know you're probing the fact that I don't know what qualia are but nor does anyone the reason that I'm digging into this inexplicit ideas thing so much is that I expect that this is related to what Consciousness is and how we come up with new creative ideas and what is the thing that is different in the human program compared with a a dumb computer or an animal and it seems that you and I have different ideas about what inexplicit ideas are how they function and so that's why I'm interested in this do do you also have have the suspicion that inexplicit ideas is possibly the the secrets of Consciousness or what do you think of that well a a secret of Consciousness since since since we have no idea how to how to encode things like Consciousness and qualia in a computer there's something we don't know now it and definitely and if if that something did not include qualia and and so on then and Consciousness then they they wouldn't be sufficient they wouldn't be sufficient to make an AGI and so on so whether understanding that is enough to make an AGI I don't know and whether understanding it once we understand it can be translated back into the form of things like equations I don't know that either equ so here you go there's pieces missing and like the reductionist atomistic view breaking things AP didn't make things work uh it's the same here just putting more compute most likely doesn't resolve anything doesn't result the AI people are hoping for aat is a community uh even though it's also a social platform and uh this David spin's summary of His 11 thesis of communities um is an

3:16:25 interesting I would say opening to think about what's happening here in aat and where it's going and how things have been going so far and why they've been going where we are at the moment so check it out and if interested give your thoughts something is not right in the world uh four week old a startup races record 104 million in European p and yet the uh valuation is 240 million and this is a Paris based mistal AI which just started something is not right in this news I don't know more about it it's just the tweet I saw but I I think the sentiment is there I do agree it's probably excessive but I think people are now looking at AI as being a hundred billion do plus space and so in investing at a couple hundred million valuation for AI experts out of meta and deep mind doesn't seem like a crazy bet you know it might be a 1 in 10 might be one 100 but if it gets there it'll pay back and the scarcest thing right now is to find Engineers with AI experience who are driven and have vision and are willing to build base models but it's probably inevitable that we will go through an overfunding and a hype cycle and then crash this just the nature of how technology and Technology revolutions get deployed often the startups who doesn't have a real founder it's like a corporate from the beginning U corporate people coming in high valuation a lot of money from the beginning but there's no real ownership no agency and this seems to be exactly that type of case where already 50% or more of the equity is going to the investors so yeah big promises but is it really going to uh push the needle I'm not so sure oh I didn't catch that I assume two 40 million was their pre money and they sold less than a third of the company which is still quite high but they got a nice chunk for it and AI is very expensive if they did indeed sell half their company or more then yes

3:18:28 it signals very poorly the founders aren't really in charge and who is the core founder anyway and was this team slapped together or was it really organically grown with a leader at the helm these are important questions now the lead here is Light Speed so it's not necessarily a classic corporate play where they're just throwing money at it uh lighted is very profit driven they're very smart and they've had some very big hits so hopefully they did their due diligence on the people as well as on the idea is it any big surprise that uh it seems to be that these big corporates uh have been using each other's uh data for training the llm models uh for example that uh open AI has been using YouTube and broadcasts uh to train the whisper model and uh probably each and every one of these have been using data which is actually most likely users data for these big models and uh yeah now it's a matter of regulating and preventing the commercial use of these models when they have been used and trained by other people's data air chat never feel alone again and here comes to context AI is going to make us more lonely people will ask AI for help instead of our moms dads friends and communities people will get AI girlfriends and boyfriends instead of actual relationships we're going to need each other less which will make us need each other more it takes a long time to the Future Evans is also taking a step through the metas expenditure of uh doing the uh well day version of the VR and he SS a R&D of Apple before the iPhone came and they are completely different ballp numbers there it's a good piece and very insightful if you read the end you can see he's withholding his

3:20:31 criticism because he hasn't actually tried it but overall he's skeptical and he does think that the iPhone is the correct Mass Market device all the way through the form factor in these things is just too large it is the right device but in the wrong form factor and getting to the right form factor might be literally decades away due to battery and weight issues something that even Apple cannot solve indeed he said that Apple started from the MVP and that's the minimal the high resolution um screens and the compute needed have you seen anyone uh actually digging into the physics of things where are we with the compute where are we with the uh the required components and min isation of all those things I think the Battery Technology is probably the key but if you really want to have it like a eyewear eye classes and if it cannot have anything else so no weird wires dangling somewhere everything needs to be more or less transparent needs to be miniaturized how far off are we given on the Leading Edge technology we having at this point I have not it would be an interesting analysis but I think anyone who's been around the tech industry long enough can also just see the speed at which battery Evolution moves and this is mainly about batteries more than almost anything else yes uh then the thermal effect I think is the another one if I recall it was the magic liap who had issues with the uh burning in your face Sensations because the units were running so hot so maybe it's also a little bit about the uh computer chips that they need Cooling and that's that's going to be an issue so the transparency cooling uh compute and minorization with the enough choice to run with it those are probably the exponential law is hard so who knows or maybe it's something we beam it elsewhere the energy and the compute with some weird ways and that way

3:22:33 doesn't really need that much but if we uh step it up also that it needs to be not in your nose but within your eyes so you have some kind of contact lens that's a another level of difficulty again I don't have much hope for wireless power and bandwidth you can think of Wireless bandwidth as the entire air around you is just one cable which is shared by everybody so there's just not enough bandwidth relative to what's always available with a cable so people will always choose the cable power is even worse all of these power uh startups like UB beam and so on eventually collapse in a Halo snake oil obviously for wireless bandwidth we do use things like Wi-Fi but you're not going to use it for an interconnect between different parts of a computer you're not going to use it to drive things from the GPU to the CPU that's where you're actually just pushing maximum resolution and uh you know maximum Fidelity at all times so you're not going to take the massive Fidelity hit that comes from going Wireless instead of wired some 20 years ago the startups trying to build uh CPUs with Optical Computing uh having lasers or some other diodes emitting light and building Gates uh from those bases but I don't think it went too far at that stat too expensive too complex too difficult I know I was part of that movement the first company I helped start was building semiconductor Optical amplifiers and use vertical cavity lasers for that purpose um that's still not going to go in the open air open air scatters so it might work in a vacuum if you had line of sight but other than that I wouldn't hold that much

3:24:36 hope something you learn every day so for some reason I didn't know you were part of that one so pure coincidence I took that topic up uh I was also involved with the [ __ ] nitrite uh building templates wers uh building the blue light then eventually obviously coming to the Hy prus lad so I was part of that movement in the early parts of the 2000s as well hilarious yes looks like we both spent some time investigating gallium nitrite and indium phosphide and vertical cavity surface emitting lasers and herum dope fiber amplifiers and all the like life is long yeah and after that one I was involved with the p and f to Second lasers doing optical codings without uh termal effects this chit has to be part of some sort of movie dialogue please please how to turn things extremely toxic and I don't know what there's a way going back in this case Reddit the facto carried out a hostile take over one of the communities Apple subreddit by saying to the mods that if you're not reopening the uh room again did I just call it the room I guess I did uh they will invite new mods so if you do that it's like a poisoning that well isn't it uh there are a lot of these claims that um gbd4 is so brilliant with all the uh tests it passed and here's the latest one where the MIT Underground computer science curriculum was passed with 100% Mark uh yet it seems to be that uh it was only 62% so it's hallucinating actually less than a putting in a so-called economic Forum in St Petersburg but nevertheless 100% is in a total different range than 62% if your business model is tied to the U the user interface and the user

3:26:42 experience currently even though you may have a complex uh system and some modes underlying it may result in some difficulties and here's a segment but there are incumbents For Whom the current ux or UI Paradigm is not one that will sustain effectively in the new environment so let's get into uxui so as everyone knows a chat interface hence chat GPT is a common way to interact with these llms there are going to be some B2B use cases and certainly many more I would say b2c use cases where a chat interface is superior to a pointand click interface and if you an incumbent who has built their entire stack on a pointand click interface you're going to have an innovator still in a problem so I don't know if this is true but let's imagine a world where a chat first interface is the best way to build a CRM if that's the case it's going to be really hard for Salesforce it's not that Salesforce can't afford to hire gr product people to build that it's that they have an install base of millions of daily active users they're used to the point click interface and they can't disrupt that business so there's an innovator's dilemma issue that could arise from this ux paradigm shift when llms while training are looking at things reading is that a copyright infringement and Evans is saying here that uh because it's not reproducing work it's just looking at things the existing copyright doesn't hold so we don't have a legislation uh in this situation and if we start to consider that this is the the case then what happens to the hall of humanity who's been the entire time replicating mimicking and copying each other open source llms are not exactly all the same and uh you cannot check the weights uh so you don't really know what's what's going on what's inside and it's impossible to check everything so it's not like reading code here's an example uh from the Falcon somebody asked whether it not anything about Abu Dhabi

3:28:46 and uh surprisingly human rights abuses are not something it knows about but it does know that there's a good governance going on so maybe you have to Source a little bit deeper who's been doing and what if you want to use these uh weights and systems uh ethically and who actually knows what type of uh Easter next they are in all these blackbox models anyways muston is uh peculiar place if uh some of the fediverse servers moderators that is are blocking just by the spite of uh meta project 92 or was it 96 just because it's coming from a big corporate that's one take but you are not just uh blocking your particular user account you're blocking everyone who's using metas instance and you also blocking it for everyone who's using your instance and uh isn't the whole purpose of uh Federation to keep it open but uh this way things are blocked and are we any better than the rest of it if if that's the way it goes well it certainly is a way to keep it fringed because who has the time and uh patience to investigate what type of instance you're getting and whether the moderators owners of the instance are still in line based on what they say said when you joined it and uh sounds like a complete mess I was testing today a little bit uh T2 do social which is a new Twitter clone well it's not a centralized it's just a good oldfashioned centralized database and it seems to be that I'm uh

3:30:54 a number I'm a person 6236 ID and I think that's the most unique thing about it it states the number saying I guess how many before have been in the app then you can also verify yourself by making a schedule a call with the cofound Sarah and paying five bucks for the pleas of it and uh that's pretty much all I can report about it when the llm model has put in more or less the publicly available knowledge we have so far well there's a lot of things which are not in the net or in this platform but given what we have what we do next you can optimize the model but how far off are we from that point and how many of these big compan is going to do it and when it's becoming just like optimization gain and not any more huge exponential or incremental gain what is the next level is it rush to the ex uh existing but non digitalized rest of the knowledge they were also discussing in that uh episode marri and Alex Redman about putting llms AI Bots to discuss and argument issues from both sides and when you do that in a sufficient level and sophistication are they able to produce new knowledge the machine just uh predicts what's uh good for the uh next token but uh probably can it come up with something which uh is already existing and available but we just haven't figured it

3:32:57 out because nobody has been looking it from that angle I would say that yes there's a lot of that stuff which is already there but we just don't know it because there hasn't been anyone who's been looking at that corner so when we put these automated pots to discover existing knowledge how big of a gold mine is that given that uh the data uh this BS can spill out is verified is factual in the sense that it can be factual uh they also discuss that what is truth and is there truth and who says it is the truth but um besides that point how long does it take before we come to the point that uh humans are out of the laop in the sense that it doesn't make any sense for the public policies for bigger decisions uh to involve any of the humans because it's just evidently wrong like uh you cannot accept any of the Marxist theories into the public policies because they just uh are punkers they just are not relevant so anyone advocating or suggesting anything which is just widely accepted not through is just becoming R ridiculous and uh out of bounds so how far off are we from the point that basically politicians lobbyists become so naked that it's becoming ridiculous my guess is that the deception goes to the deeper level if previously governments were owning the media you have a national broadcasters and they were controlling who can tell

3:35:01 how it is now the same is starting to happen in the technology layer so the national llms are going to have a conveniently twist ing the weights in a way that you don't really know what's underneath and what's hidden so the Chinese pot will not tell anything about certain events and Indian pot will not talk about something Russian but Pro probably don't talk about almost anything about anything purely fictional story so it's just a level of degree and then the fight goes to the open source versus approved llms and the same fight goes on it's still a big capex capital expenditure issue it's still the incumbents versus the open source Pirates hackers who are trying to keep the world open and fighting with Wikipedia and open source software to keep the balance and and the check in place this RAC is on issues because the soures the so-called facts will go more to the background there's no more books to verify those things they only digital recordings which can be also manipulated yeah there maybe are images there are speeches there are maybe text but it's more and more in a digital form which can be manipulate to whichever shape and form so how do you verify anything in the future and how do you verify the past who has the original copy of something the original book which has it the way it is and originally was how do you trust anything

3:37:06 and how much does it really matter I'll be going back to the basics in the sense that is it the right thing to do and that's the guiding principle so it's ethical question and uh facts per se are becoming less relevant if we are not talking about something which is physics where you cannot negotiate with nature if you have access to whole human knowledge space applicable to your context right now and every moment forward what do you do isn't it then not race and basically combining anything existing to just a different form is something everybody can do so the ultimate Frontier is creativity creating something which the BS cannot do because they can mix and match and uh recombine things they can extrapolate they can do certain type of things but they're not creative in a sense humans are so every person needs to become a creative person if you ever had a open book exam you may may be the first time think that hey that's really EAS before it hits you that it's actually really hard and this is probably also the way it goes in the future when everything is an open book exam but you just have a shallow knowledge of this situation which is given to you as a suggestion but you don't know your history you don't know the context behind

3:39:09 it so that way you need a neural link to get the context and the depth which now takes a lot of time and effort uh to to learn these things and experience and experiment maybe there's a way to expedite that as well in the future but that's probably where the compeed advantage and Ed will be you might be a expert in tactical prompt hacking in the future way where you know your way and you can ask the real smart questions without having the whole context behind you autal models photos but that's not the way to go about being really creative because that's usually the combinatory effect of having a wide knowledge and deep knowledge of inter related and sometimes isolated issues which can then spark the creativity and create something new or if you're not creating something new what do you do maybe there are some islands where the pots and the robots and automation doesn't still reach but they're getting sorer and sorter in the future at least uh fulfilling the basic needs so is it really that we have to go back to the history arts and uh all the soft Sciences unless you really want to push the edge in the physics and the hard Sciences so these are your options because the rest is more or less covered the new level of future idiocy is that you can only ask things but you don't

3:41:13 know why or what you just uh repeat what it's told to you and you just move your hand and do this movement because that's what is shown to you and you don't know the cause and effect of the things behind it well somebody could argue that that's how we drive a car we don't really know what happens when we push these buttons or we don't know what happens when we push any buttons in the computer or mobile phone probably that's just a different level where it goes way deeper I went through the list of um rooms which are open to me and I noticed that there's a lot of those who doesn't have any new updates the list was very long I would guess in hundreds and that puts me thinking that what if aat will not have this Mega influencers and it will not have these Mega rooms huge rooms in other places like huge groups with uh tens of thousands of people or thousands of people because it's so easy to Fork here the room make a smaller one and if there's too many discussions going on at the same time it becomes um difficult to read it's like reading Elon musk's tweet comments there's just thousands of them coming immediately and you just cannot follow the discussion anymore not taking any uh any comment on the quality of the comments but the idea that there's a too much content so it's becoming uncomfortable or impossible to to follow it similar manner it's so easy to copy the success of a room and not just copy but make it more comfortable more intim by just making another one

3:43:18 even though the topic is the same we already now have here the AI rooms a few of them prominent ones looking at you naal and Kei so my thinking today is that uh maybe the big influencers are not coming here anytime soon because there's not going to be a lot of followers uh in in that sense and also maybe the there will be a lot of rooms which are small and the discussions happen in these more intim room or house or block settings which are limited to few T of people per room uh this looks pretty amazing uh you will see all the stuff it can do the price is quite Hefty as well it's 25 100 and I don't know whether it actually even includes the additional keyboard but uh the way it can be used uh the OLED display looks also pretty neat um for the first time it looks like that wow uh maybe I would like to have that one instead of a iPad but uh I don't know it doesn't come with the Mac OS anyways uh if you want to check the original video it comes with the German um description of the of the device and and the functions uh I muted it but you can check it from the V ofs even well in his Tik Tok it's not in his website so check it out if if you like to to hear the full description he also made a blog post about it maybe I can put it into the comments the actual blog post but anyways looks pretty impressive yeah I agree with this and

3:45:30 I'd take it a step further although I'll listen to the segment in its entirety it's not even going to be a text interface it's going to be a voice interface and models that get get good in a hurry at dialing in what's best for the customer air chat could actually be really powerful here and like a quick chat you know as an update to the t2 experience still haven't seen R2D2 but um the traction is actually better than in Blue Sky I got something like almost 20 uh reactions to one of my posts uh yesterday and it seems that even though there's only a less than 10K signups which doesn't mean that they all active there's a some activity obviously everybody is new and everybody wants to connect and everybody's walking around this thing but uh still it's better than in some some other of these new emerging networks and social media experiences Mar Reon is back in the broadcast talking about his latest essay about Ai and here is a segment uh checked the previous one first it's uh more longer and I would say more comprehensive discussion but this is like a next one on the line if you're still interested of the topic and want to get a little more nuanced perspective on the on the matter sumption that basically smart always beats dumb right um and the other form of this that you'll hear that the doomers will say is they'll say well if anything sufficiently smart can convince anybody sufficiently you know less intelligent to do whatever they want right this is this is sort of one of the things that they do one of the superpowers that they kind of give AGI to kind of say why it's going to like Escape you know constraints whatever um okay here I would ask like look at the world around you right would you say that the smartest people are in charge that feels like a rhetorical question which we move past like like like how's that going right so like so like the again here the core assumption that like smart people are somehow in charge of like less smart

3:47:32 people like not so much yeah I would call I would call that one easily falsified that sounds like a cult which well there's been a lot of those cult um yeah so there's a lot of these um uh back and forth about U why AI is not really going to destroy the humanity here's the third one I've seen now in a matter of days uh di missing values obviously uh not much new uh about the different angles uh what is interesting he's talking about covid and uh after thought of the co and um how not saying the right thing and not thinking what you're supposed to be thinking and uh that discussion I found fascinating also he talks a little bit about um the way a16z goes about now and sticks to the founders and supports them in good and bad times and how the head count of uh his VC is so 500 people I thought they were a little bit scaling down that model but uh maybe I was misunderstanding that but um I would love to hear if anybody has some insights how they actually are are they really sticking to the uh what they claim to be supporting in good and bad times and really on the founders side when things are not going right and and and good for a long time Clubhouse I don't want to point them but it's just uh something which came to my mind that obviously uh hasn't gone right even though they raise three around of funding I think within a matter of a year with a good valuation so how is it going for them they clearly still alive so yeah third one on the row maybe there's a fourth one as well I just haven't found it yet Elon Musk may have finally managed to kill Twitter if you're unverified meaning you don't have the

3:49:36 blue check mark and you haven't PID the eight bucks Chuck tax per month then you are limited to 600 tweets per day that's what you see then you will have a rate limit meaning that you cannot see more tweets and 600 tweets is not that much if you consider that uh it's all those things you skip over you checking some thread and there are a lot of small comments all those account so 600 Pats if you want to use the call product sounds a bit silly users are penalized for using the product and the cor product is about reading the tweets and also it's a advertising business he's supposed to serve ads for the users in order to get views and uh now there's a limit which also also reduces the amount of views impressions for the ads I'm not sure what to think of this uh the easiest explanation is that he's running out of money and he needs to limit uh the expenses and uh views are expenses uh you have to serve them and uh this way uh play in the uh AI scraping uh which is not really any news there's always been web scraping doesn't sound plausible explanation oh I think this is absolutely about AI scraping I think he wants to pres preserve that data for his own AI company that he's building and he doesn't want it lifted by open Ai and by other crawlers I think that's exactly what's going on so not an

3:51:41 accident he'll change the is needed that could be the case but um I don't think that's the most uh easiest answer here because Twitter also limitate the access to to the tweets if you're not logged in and this theory in this tweet says that U feder started to do alling itself uh because some kind of a bug so this would explain why even the rate limit was put in place hastely and also increased just a matter of a few hours later on so it was not like it was anticipated event like if you well he already uh limited access to the API earlier on that was already expected uh Behavior this seems like a rush and uh quick thing yeah the 1 of July is something as well which could also affect this but uh well these are the theories I I found out obious L limiting the lock in requirement also limits the scraping so yeah well obviously you would rate limit accounts uh that are not logged in or prevent them from even seeing the site if you're worried about AI scraping you so I think that's all logical and consistent and for all the hatred that's flying around on Twitter there's just a whole bunch of people who hate Elon and will automatically attack and call him a [ __ ] and try to make it sound like the worst thing possible yeah there probably some unanticipated qu consequence of doing the rate limiting they probably had to implement it in a hurry because they probably saw AI Bots just going for it and once the data is gone it's gone forever and now they kind have to clean up their other processes that are relying on that which I'm sure they will it's not rocket science but I think all these people rooting for the downfall of Twitter are just the people who were demanding censorship and we're annoyed that Elon opened it up to free speech if anything I think you know I'm not a big Twitter user these days but from AF far it seems to be doing reasonably well maybe even better than ever and of course the preview is also

3:53:44 gone now from Air chat so here's a picture of the Tweet yeah I saw that announcement today too I also saw one immediately basically from Elon mus saying that rate limit is going to go up I don't understand the reasons exactly why this is happening but whatever the case I think that he's going to change it uh this is him you know marketing whatever you call it uh in order uh generated a bit of Buzz he's he's very good at this it's a form of trolling I think maybe there's an important reason behind it maybe not but I wouldn't be overly concerned at this point and also he says that it's about um reading posts now what constitutes reading a post is a post absolutely identical to a tweet I don't know is he talking about that long form thing where you can write whatever it is now 25,000 characters or something or other you can include video and all that sort of stuff is a regular tweet a post and is scrolling past one does that constitute reading do you have to pause on the tweet for a certain amount of time does that constitute reading it's a bit slippery he hasn't defined his terms there so I don't know then again I'm not an expert in any of this stuff but I just wonder I just can't imagine until we start getting reports of people saying hey I'm an unverified account and I can't read any more tweets that this is a problem maybe it is I'm speaking from a place of ignorance yeah I agree with everything you're saying there you know it seems to be haphazard uh going off half cocked all that sort of stuff but um let's put the optimistic spin on it it's just iteration right he he's checking stuff out he's experimenting hey more power to him to do that stuff and in the meantime people go ah you're ruining Twitter Twitter's you know falling apart and everything else he's sitting back going look this is just temporary don't worry about it

3:55:48 in his mind he's just thinking that people can yell and scream or they like he's he's he seems to be completely immune to that in some situations couldn't give a damn about what people are saying the noise the buzz he ignores it all just plows on through trying stuff and hey if it goes bad then he just changes it back he reverts back he try something new try something different so I think that might be what going what's going on people don't like it they're very protective of Twitter they don't want their platform to be experimented with and so they're going to just say how awful it is and how that this is the end how many times ever since ever since musk took over the place how many times have we heard well that's it that's the end of Twitter Twitter is just you know going the way of the dinosaur the whole place is ruined I'm leaving Twitter that's it you know everyone has to announce you know when they're leaving Twitter no one does the place seems to go from strength to strength so I am not at all that worried I'm sort of looking at this just just as a you know that old meme the person with the popcorn just eating eating the popcorn while you know the people on the side of mus are saying yay go Elon do this thing and then the people on the other side saying that's it I'm leaving Twitter it's all gone terribly there's all this hate speech out there more power to him to try different stuff I don't know I'm not aware I don't think many of us are aware unless you're inside Twitter of as to what exactly is going on uh it seems to me he often knows what he's doing better than the people who are outside the company and he's playing a game I mean if anyone in the world can be said to be playing 4D chess it's the world's richest man who is out there actually building rockets and and going places so he he knows what he's doing the rate limit was very real at least in my case uh I couldn't see even even on my own tweets um what I noticed is that you could uh if you had a logged in session from some other device or browser uh it didn't apply to that one so you could if

3:57:50 you were logged in with multiple devices then you would effectively reset the um the limit but nevertheless it prevented you from using Twitter I agree he's really good at playing games uh if you remember the CEO uh appointment he already knew that beforehand and he was playing these balls that should I resign what should I do and almost all of those occasions where he's asking something he already knows the answer or he's already decided something and he's just playing the game in this case I think it may have been that there was some fault in the systems and uh he just spined it this way because it was evident that things are not going to blow up so he needed to respond it some way and why just say that hey we failed with something better to spin it in some fancy way and make it to whatever marketing event because it's going to happen anyways and he just wants to have a little bit of fun like you describe and he's already now shuffling it uh maybe it's already a third time he's been changing the limit it was a known fact that uh he was not paying for the cloud services and Google was the one I think he cancelled and it took effect uh first of July and uh so in that sense if there's a lot of uh services in the back end which are dropping off and this is starting to cause issues uh it was a know thing he was trying to ramp up the alternative services but it wasn't fast enough so in a way this is a passing event nevertheless uh one of these whales which we don't ever see anymore in the Twitter but uh well stuff breakdowns at times but the biggest thing is that uh there's no really place to go and this is once again another attempt for a lot of people to try to go to Mardon some of them went to T2 Blue Sky people are

3:59:55 scattered around but there's not enough impact uh a network effect to go one place and this is quite unfortunate uh because well you have to be in all of them pretty much because some people just choose one place and others the other place and now you don't have a un unified place to to get anything this is a very interesting take about how uh musk has E3 and wrly or purposefully basically cut off the uh censorship uh industry by cutting off their access um so worth the listen I think it's like 8 minutes uh worth listening to it all cuz it's all highly interesting this is great um but I think the unfortunate reality is that most the value from Twitter comes that it's uh fully public and the de facto broadcast medium for the internet and I think the thing that I'm afraid of is that by locking it down um it's basically ruining the reason that people use Twitter I think that people think that Twitter could also be for conversations and with pay walls and closed off that people still come there for these conversations and you know I I find that suspect because you know there's platforms like air chat which are far better for conversations um and Twitter's highly geared towards broadcast and making you famous um for that one tweet where it's totally innocuous but uh it gets put on CNN so I'm not sure how the platform will evolve as a function of these changes um I think it's great that they're going to put the death nil on the the censorship that's if you will um but I don't know yeah seems complicated exactly it's uh it's not Twitter's content per se it's the users content who's been generating it and uh unfortunately still today twter daily

4:01:59 users user hours are 40 million and top 25 News sites in the world aggregated daily use hours of 15 million so uh it tells where are the news well they are still in the T and twio is basically originating the news foros uh top uh 25 News sites in often times as well so in that sense if uh Twitter goes down a lot of the news go down as well I really think this is short term as a matter of fact on the other side of this I think 99% of the people that contribute to Twitter are coming back and uh probably at least 10 or 20% of the Bots that make Twitter crappy are not coming back I think it's a brush off the rapid poles the Optics are there for scraping and that's the reason to do it but I'm sure he was just thinking sitting around I was like wait a minute if I turn off the scraping all these damn Bots that are constantly like bugging all the people I'm actually here to serve are going to have massive problems and we'll just see what they do remember it's a privately held company right now he doesn't it's actually of no benefit to him other than fluffing certain metrics that frankly have been fluffed for a long time and therefore served the business poorly because there's no real feedback mechanism to how many views they getting so it's privately held company anymore Maus dius doesn't really matter unless they're kind of either bringing value in content or bringing value um in consumption and all the Bots and Rapid buyers and the people that keep coming back to the wild just try and scan like thousands of things um yeah that those aren't profitable things to concentrate on much like not so much with airchat but like if it's only about listeners do we navigate so we can repeatedly get back to where we were listening only for weeks on end uh I'm not sure that you know End customer that cohort is actually longterm what makes the platform like a wonderful place for sparkling conversation I mean this is my bread and

4:04:02 butter of course the customer is very important but there's lots of people that are in your vision and in your workspace and in your path of existence that actually aren't your customers that person that's merely reading Twitter and doing nothing and scraping a thousand what did they doing for your business they're not your customer good marketing good business is saying [ __ ] yeah to Great customers and I don't care about the other customers in fact some customers out outright need to be driven out for the business to succeed which is why there is this constant like jokey griping about customers in every profession including the restaurant business and it's not about all the customers it's just like it's actually about you know the customers that are [ __ ] that should be dunked on right like just it is what it is that's how it happens you know um so those aren't mutually exclusive you can exclusively serve the customer and then look at another potential customer that's not a real customer at all I mean just doesn't add or contribute anything to your business model So within the confines of propriety and serving your real customers you got to get rid of the lousy ones so you rate limit him bye you want to leave fine bye uh this uh Douglas half interview is quite revealing and U I don't know what's the best word for it I would say puzzled puzzling confused confusing depressing depressed so I think he's all those words and uh to you the Watcher may be feeling the same way and uh in a way it's quite a sad interview uh he has written such a wonderful book and uh his life work is uh so fascinating and yet he somehow seems so lost it's really interesting that he's

4:06:09 been trying to find the meaning and uh crack the puzzle of the mind and Consciousness since the 60s and he was even doing a rudimentary chat GPT experiment without really obviously having the the capabilities at the time and he was really puzzled about the the meaning and in that segment he explains that it's just when you have you can extrao curve fit perfectly enough so that it seems that uh it is uh having meaning then it has meaning it's like when you have a really good Vision Pro which is mimicking the world and you cannot see the difference between them then it is live and it has meaning so there's intent behind the words there's creativity behind the machine because it's so good at giving Illusions so there's an interesting twist here where sure that's El Elon can protect Twitter Twitter's data which is probably protected in their terms of service along with many other things um but terms of service can be thin Comfort indeed like nval do you consider all your tweets elon's property now because they were on Twitter I mean it seems like this is the same version of like you know um who has the rights to train on this stuff I don't think Elon has the rights NE I I think he has legal rights I'm not sure it's clear that he has we're talking about not Distributing stuff but like what these AI models can do and the of data that goes into them and whether or not that should come cost free to whoever's training the model I think it's an interesting area to dial in on WE and then it trips us into part two which is if it is the Public Town Square if it is the world's Town Square which he said and I believe well I don't know if I believe it all the way but it can be axy for that in that environment do and extend that to Facebook extend that to all these other places now what are people's underlying rights if were you

4:08:13 know pragmatically without the terms of service nonsense just in a common way we kind of all knew we were talking in public and we all know that if we're in public like it's a public statement um so yeah it's an interesting rabbit hole and then last one I mean you've been on Twitter like all due respect I mean these a lot of these Twitter accounts they're lucky they're going in some big model and getting averaged out like I don't know there's a good amount of accounts on Twitter that if they got dragged into the light for their [ __ ] um they would deserve a good amount of public ridicule uh so I think those folks ought to be happy they're getting averaged in there's some real [ __ ] there I used to resent uh not resent but the type of regular that would come in and be like oh I remember this place like no one was here was awesome we had the place to ourselves I loved it that it was so cool now this place Total H everyone's here and you're watch and you're like yeah dude like I remember that like that's when we opened it was brutal right no one was making any money we worked hard to bring it up here you are wishing like wasn't busy anymore and I say that because right now I'm like man I can't now thinking of it like this I don't want all these broadcast media like multi-level marketing people in here I hope this air this air Chat thing just kind of flounders along for 20 or 30 years while I have incredible relationships with smart people sorry Nal ooh yeah this is a rough one actually careful what you wish for threads what can you do with that you cannot do an Instagram but more importantly the other way around why should I update both of them if threats enables to put all the pictures videos and it it has a larger text form and it's the novelty thing now and um if I having both why should I update both of them so I think these are some of the questions so is meta actually canalizing itself and maybe that's the intention as well because they are well in a bigger

4:10:17 competition now with all the other social networks and maybe that's on purpose but if you think about from the individuals perspective why should I have po because I can even import the whole Instagram social craft so does it actually really make sense and if same people are following me on both are they supposed to get different content the same content and why should I see it yeah exactly it really makes no sense when you think about it if threads can do more than text they could do photos and videos then isn't that just a competitor to Facebook but Facebook is owned by meta just like threads is and yeah if I'm posting photos to Facebook and threads why would I post in the Instagram too or vice versa it's just like absurd it's amazing to me that all these companies are owned by one Master Company meta I mean these people are lost in the trees like they need to zoom out look at the forest for a second it's just like people are making products just to make them and they've lost sight of everything that matters in some way it works with the physical products like uh uni liver and some soaps or whatever can't C or Cosmetics you can find the NES and you can canalize products even within your own um brand Brands if you have a hundreds or tens of them and you want to capture some Market segment but because it doesn't really matter from the user perspective you buy this or you buy that but uh if you're producing content or you want to use use the product as a distribution so now you have so many places to keep it up and you're not really getting any traction properly in any of the places unless you already accomplished and you just push it out to everything and somebody just put it out there to all the possible channels but

4:12:19 if you're building something you have to properly focus on something and there's so many options already nowadays just to get a little bit of perspective on this thing let me try to um list some of the options you have as a social media manager or if you're creating content you want to do the distribution you could start with YouTube Twitter Facebook in Instagram threads T2 Blue Sky forecast nost and the list just goes on and on it's just impossible to keep up and uh you need to focus on something and even within meta you have already three options and the social crafts are pretty properly aligned but they be completely different in Twitter in YouTube and in some other places and I already forgot Tik Tok not quite because uh Facebook is a friend graph and uh threads is a follow graph and those are completely different and lead to different networks and different outcomes um Facebook also tends to be private or semi-private amongst your friends whereas threads is very public so even if they add photos and videos I don't think it's B it it change yes it changes the media types but it doesn't change the fundamental interaction type and the fundamental interaction type of threads in Facebook are very different that's fascinating Petri so yeah I don't buy that line of argumentation either I think you're being skeptical about it correctly so which is that just because something is higher Fidelity doesn't make it real um after all even the Vision Pro the Fidelity there is just enough to trick humans but to an animal with a different form of perception or to a measuring instrument The Vision Pro is not rendering reality it's just rendering reality to full our eyes and our ears and the same way uh a chat GPT might be uh conforming or pointing to the real world but it's not pointing

4:14:21 with the same understanding that a human being who is actually embodied in the real world has so yeah I don't buy Douglas Hof's arguments at all and I think he was Far smarter than that so we must be missing something I mean obviously either something came across as incomplete uh or there is uh a further underlying intention or new Nuance to what he's saying that he was not able to convey in that video exactly right which is why I think that threads and Noster and forecaster and others are very very difficult because they have to displace Twitter so if you're already good at creating text broadcast content you go to Twitter you focus on that if you're good at images you go to Instagram you focus on that if you're good at short videos and music clips and so on then you go to Tik Tok you focus on that if you're good at long form video you go to YouTube you focus on that which is why a lot of these U networks tend to be created around a specific content type if you're a content creator and you're good at a specific kind of content you go to where that content is best showcased the largest possible audience and that's it full stop so having multiple players in a single content type simply does not work all the numbers 2 through n will go to zero indeed if you watched um the video I think at least I I'm reading it in a way that he's really depressed he's lost he has probably lost his uh original purpose and meaning or forgot that so I don't think he there's a angle mission in that sense I think that's how he genuinely feels uh today and unfortunately that's where we are at this point um a lot of people already have left Twitter is not the same place anymore and they just scattered into wind um so there isn't really a place so in many ways you have to now go to the multi- channel approach if you want to reach to people because it's also risk if you just in YouTube YouTube does something that's whole of your Social

4:16:26 craft or like Twitter if if they do something uh silly and you're being just putting together the engagement and the following for the last 5 10 years and you have a significant amount of people with they there what do you do uh checking my threads feed today after 12 hours or so in the app and I couldn't help but uh a laugh out loud uh because it felt like a parody account like a real parody account because everybody was pushing their content there repurposing their content and trying the best to pit people and uh some are some obviously were doing it as a joke but many were doing it uh as serious as they can but it still felt so unreal so I don't know KY in a way uh that uh it's like okay I'm too old for this game it seems that um the Colossal fight has already started between meta and Twitter by Twitter sending a letter to meta saying that stop poking our IP and followers and uh our data the letters to PES you can see what they're saying there so it's like do check them out and check also our robot uh text uh file in our website and uh you know you should know what is allowed and what is not allowed game is on you can follow public figures and big companies pages on Facebook but yeah Facebook does have that friend model is that really any different than just having a Twitter account that's private where people can follow you and you follow them I mean it's really the same thing under the hood um and when you really think about it like Instagram does pictures and if threads also does pictures then it's

4:18:31 competing with Instagram and they're both owned by meta so it just looks like a mess to me and I can only imagine these social media managers posting from 15 accounts trying to publish everywhere like you were saying the n number it just starts fizzling out to me what would be more interesting is if you unified all the mechanics and you found all the things that worked and you put it on one social media like one decentralized social media that had all of it in one place and wasn't a ux nightmare somehow and actually pulled it off because if social media is truly decentralized then why do we need 15 different ones especially if at the end of the day it's just pictures video text and so on on you could have it all in one place and then you decentralize the data across thousands of nodes uh in different varieties of noes you don't need multiple social medias technically so I think that's a much better experiment worth having at uh and maybe it would be the piece that all of these clones aren't bringing to the table something that has never been done before not boring phone I posted this yesterday I don't know what it is but anyways can you see and guess AB bcde E I think there was even seven or somebody mentioned seven which one is which which one is Twitter and which other clones so this is the area of everybody cloning each other and having the same brilliant idea except uh a chat and few others who are going completely different way but nevertheless yeah it's the same text block going going to different places and for some This is the End place and for others it just it's just the distribution and I think then the only difference is that uh are you really sharing any of your family stuff or extended family stuff in any Social Network anymore I haven't done that for many many many years and it's sounds weird

4:20:36 and strange to give your family photos to some of these advertising driven companies so maybe that's all going offline anyways all end to endend uh protected messages or some other places which are not run by these social media companies anymore jgpt uh code in the prer is now available for everyone either via API or if you're paying subscribe so you can start to use the llm model uh for um data analyst purposes for example so the B can now write python directly so writing programs and then analyzing and outputting inputting data and manipulating it and outputting and you don't need to know how to code it previously we needed to do the python coding or copypaste at least the code run it and Deke it and put it back and that why it was a compass and if you didn't really know the language it was harder to do because you didn't know what went wrong now that's all gone and done within the chpt and that article is pretty interesting which I will put on the comment uh and uh I will tell you in the next [ __ ] uh what I'm actually thinking what's going to probably happen or the thing what I've been pondering it's not exactly related to but uh more like a future prediction now that it becomes clear and evident that um and I read somewhere that almost 50% of all new code is not written by llms anyways um what happens to the pythons and all these languages which are human made and it's a huge achievement to do those and it's something which is almost like a basic infrastructure but uh nothing really happens if if that is not in place so are there still in the future a lot

4:22:43 of python core developers uh because it's becoming less visible if even the these open- Source languages are written mainly by llms is that also something which becoming like a luxurious uh and uh sort of press this thing but mainly these languages will also disappear like uh what happened with um language translations mine language translations they started to directly create their own language so there was no go between language anymore in the later stages for example when Google uh was developing the models and uh is this the same thing happening soon that uh pythons and these are going away after they become enough sophisticated and now it's good for the error correcting and it's easy to see that if there's an error and it's a feedback mechanism for the llms but uh in 10 years what going to happen next it's funny literally the only social media I can think of that couldn't be merged together into a single one is air chat like Facebook and Twitter they're so similar like if you want the privacy of just friends you know just you can have a private account and they have pages and I mean all the stuff is just so similar I'm pretty sure air chat is the only thing I can think of that isn't so air chat is saved I would say though that Snapchat is something that probably couldn't be decentralized because you're trusting the servers to delete your messages and you cannot trust decentralized servers to do that I mean maybe you could have like endtoend encrypted Snapchat but I don't know how you could force the

4:24:46 client to delete it either you know so um maybe you'd have to do like tees like proof attestation that you deleted it but then it's just like bro going pretty overkill for Snapchat but uh yeah I mean everything else man just kind of seems like DMS or posts of files under the hood it's all feels the same so that's why I'm intrigued by by air chat there really isn't anything like it yet it's pretty pretty clever yeah I haven't used um enough the simple x to know whether the client can actually Force to delete my message if I send it to you but uh probably this is the closest we have now uh which I think there are already some initiatives to integrate this to the nost as well but uh yeah maybe there's a replacement for Snapchat as if we would invent social media now and we don't have anything before but we have uh BS chat GPT type of user interface how would it look like and if we have the multimodality in place so you could start with the text uh you can have video you can have audio you can properly draw you can draw in a pixel do it it either way uh convert picture to text or whichever way so basically everything is possible from to translate transform from one set to another how would a social network and social media app look like I would go for this Atomic uh content unit which would come from that multi modality and U if optimally the philosophy the values go for the

4:26:49 decentralization you own your own data then it would probably be something along the lines of air chat meeting nost or some other decentralized network and what comes with the storage is then one thing but the other is that uh it's and every piece of content is their own content unit which can then response to some other or they can be combined so in in Twitter way text way it would be called probably a tweet here it would be called a chit but it's a the smallest content unit and it can then respond to some other ones or can be merged together with the other ones and uh they could could be distributed Maybe maybe even in some other form the text form can be converted to video or audio and this way it can be distributed and it's up to the user The Listener the other end the re receiver to decide which form uh the intake will happen and also probably which which way the response will happen as well and that's pretty cool future way if you start to think about it this way then the interface is just uh user preference user experience and uh content becomes more like a backend storage unit in a way and it doesn't really matter which form it was in the beginning now after playing a little bit more with threads and connecting with people and and uh posting a bit more and noticing that it actually allows you to do things Twitter doesn't allow you to do anymore Twitter's video

4:28:52 limit is uh 2 minutes 20 seconds uh threats goes now for five minutes you can put links I think you can put multiple ones the character limit is uh bigger than in Twitter I'm not now talking about Twitter without paying uh for the F features and it seems that fto is going more to the paid route for everyone and if threats will keep being free and it basically offers good enough service I think that might be the biggest thread pun intended for all the other networks who are now trying to get the users from Twitter and invent the wheel again the good enough is usually the the biggest appeal because it just works and if it's not irritating and annoying enough and there's certainly the appeal and the possibility for the network effect and um in a way I start to see that already now happening at least within Tech and few other circles which are starting to post things it's not funny Kobe it's a fact cuz I feel the same way I can't compare every child with anything what exist and it's very hard to explain my friends uh what it is because they're they're asking me to compare it with something and I can't because it's impossible cuz it's something different it's brand new unique it's a fresh air IR the definition of a loyal fan has uh now gone to the next level it's not just that you have to like or engage within one platform now the expectation is that

4:30:55 you do the same for the rest of the platforms as well even though the content is the same an elaborative U summary of lot of the points circulating now and also why it feels like that we are now at the end of the user to user communication and networks where the actual value is to try to reach to people and uh well interact and not to be milked uh for views or data mining or uh considered as waste if uh users are not paying for the service so premium model doesn't really work anymore if it's ever was working is in this concept there are a lot of points and it also doesn't uh paint too good of a picture for the future of threads this predicts the end of threats as other products early on as well and the main points are that there will be seated with Instagram users well who will quickly get bought because they're not really Twitter users so they don't must the uh that format of text and they are to influence not to uh well to do what the Twitter crowd is doing also threats will try to attract Twitter users but will fail because uh Twitter gathers for specific niches which are not appealing or meta doesn't want to have them and this way it will will not succeed for example news meta already has mentioned uh the head of one of the main designers of threads has already stated that they don't want to have the similar type of news um service going as what's happening in

4:32:59 Twitter and third one threats will be shut down less than 12 months uh to protect the Instagram's score business the upside of getting even all of the Twitter used and the ad business is marginal for the Total Core Instagram revenues so in order to protect not to get uh mixed messages and getting people off the Instagram which is the moneymaking machine they need to do something and uh the outcome will be to merge threads as a feature within Instagram which makes sense because now you can start to do more and more stuff in Threads what you can do in Instagram and it's sending mixed messages there's more into the uh Twitter thread but I will put it into the comments I have noticed that I haven't been too active in threats anymore and uh today I tried to a few times post something and uh this was the result in the all of them so I guess it's not a blow whale what is on threats uh when things are not working is is uh black to new Blue Screen meta has starting to take some more drastic measures to block uh users who are in EU so even if you're US citizen and you supposed to have the threats app uh if you come uh to the EU they will block you so you will not see the replies you will not be able to post anything you probably see the feed but uh not even the profile content of users so you are effectively just kicked out of the app um vbn may not even function at that point or this point so here we are looks like people uh from the EU or whoever is in the EU while accessing threats may not be able to use it uh this year this is intriguing and paradoxical as

4:35:03 well it's not a mass media social media if there's a lot of people in there but when you start to find for the community nominators you eventually come to this point and uh that's what the pop culture is and that's what the uh the most famous people usually are um well what's the option then if we want to have a social media in the future which is not exactly like that then I think air chat is an um example of that here it's not matter how many people you know and how many people you follow because you're not seeing everything and it's not the point of uh this thing you just cannot even absorb all the information so yeah social media of the future if it's a unique customized and uh bite size dinner party block party since 940 we have four times the population 100 times college educated people but where are the 10 to 100 times x finman's Einstein and all the other clever people that's the question are they actually now in startups and in other fields why is it that we don't see so many of them or is there so much going on that uh we are just uh not seeing them as we used to see them similar way that uh in the 80s and 60s and 70s and earlier on musicians it was easier to pinpoint the the stars now there's so many of them so it's just that we have so many of those already that now no single one pops up and it requires quite a bit of uh specialized knowledge also to recognize them so they just don't get to the social media pople anymore college education has nothing to do with scientific innovation and may even be negatively correlated with

4:37:07 scientific innovation as it introduces group think so it really it's only population that's an issue and I think that is very easily explained both by redirection into other fields as well as Innovation being below certain threshold and being recognized as great in this time and finally I think the group think effect of colleges negatively contributes towards Innovation and I think has been responsible for the stifling lack of innovation Innovation as a finite quantity that you mine I just find that highly suspect and uncreative there's always been cup thinking there's always been the popular PS which uh tend to give you the best carrier advances so I'm not sure whether these are the actual explanations there should be also more private funding available it's easier to get your message out so in a lot of fields uh it should be easier uh to be outlier and get Innovation Innovative stuff done so is it just that our expectations are going exponentially higher or is it that there's something else explaining this maybe regulation some other things which are hindering the actual progress it didn't take long for to have a real example of a Marvel movie or any other breaking pad for super conductivity type of TV series plot coming up um going back decades and really have these outliers people people who are not part of uh Academia or the regular institutional other the institutions with the proper funding who are making the discoveries this is not obviously proven yet but uh if you read this catch-up thread and uh check the background and the different stories uh it starts to feel like that yeah it's not always that easy and it can take a

4:39:12 long time to make these discoveries even though this threat claims as well that you could have done that hundreds of years back is it the losing battle to fight against uh llms which are mimicking imitating your art uh the stability AI introduced in the 2.0 version the feature or non-feature of taking away the possibility to use somebody else's style uh but then the open source Community train the model uh and uh now it's back there uh so is this the future if you're an artist um it's no point starting to fight against because it's the same thing like uh pirate pay you're not going to win after all and who you're going to sue if Tesla opce community and it's all out there as waiting models for the llms so what is it what are you suing against and who and how and are you going to just have a stri and effect if you going to go Pi with your claims hopin which was the competitor of Zoom for the event uh which was actually a really nice product uh they raised 1 billion with 7.6 billion valuation in mid 2021 um now they sold their business in low Millions to zoom their Core Business and return half of the investment money to investors and now the compan is valued at uh 400 million which is a quite significant haircut also the founder who made a secondary uh exit or so his sales uh during that 1 billion round who did 200 million I don't know whether he needed to return some of the money but nevertheless he's out as

4:41:14 well um now it seems that the company is uh starting to do something which they bought uh at the same year streamyard which was actually a nice product as well but it's not exactly the same as hopping so times they are changing and some of them were just temporary like the remote work as it seems if I use a piece of Hardware which have a code running in it and I point it to a direction and then I push a button and then there's an algorithm who will sort out the details and uh from dots make an image that's protected by copyright if it's called a picture I take with my iPhone or with a specific camera but the principle is the same if they're digital it's copyright protected but if I instead of pushing a button with my finger and pointing the device I prompt using a different method I put together a series of Words which describe the C and what I want to produce and then I push a button maybe with the same finger and then the code will within a hardware box it will spill out the pits and the result is a an image and it is said at least in that article that AI can support a copyright claim if a human selected or arranged it in a sufficiently creative way that the resulting work constitutes an original

4:43:18 work of authorship so what's the difference between me taking a picture and using another form meaning putting together some words and then pushing the button this are latest Google tool for rappers writers words me and poems called text effects I think it will set the power higher for everything the total quality of stuff people put out should go up if they're using tools like this but it will also set the power higher for anyone and everyone who is already at the top of their game and really world class it becomes so much more about uh creativity ideas behind whatever you're doing because the tools give you so much powerful ways to express yourself but you don't have the ideas it just STI outs like a bated moving few years back this this was solely in the millions the Drone operator systems and how do you do it so it was within military but now it starts to come to the consumer level so just uh if you check that video and even that the post suggests that uh this person needs a twitch channel so this becomes more entertaining and uh you this this could go in so many directions easily it's becoming a bit more blurry because it's so easy now to do that you just take your Xbox controller and uh what if I donate quite a bit of money for the drones and then I and try one of those that obviously has a lot of uh uh legal and ethical and moral and all the above and possible questions but uh this

4:45:23 is what is coming to there's going to be hundreds if not millions of drones flying around and it's becoming just everyday thing and they're going to be weaponized and there's going to be illegal stuff happening as well why it's very unlikely to get U radical Innovations coming out of uh universities and Academia well here in sound Carl's episode at the end of it he talks about it at length but here is a snippet for you 10 people you don't hire 100 people or even 10 people you hire a person every 5 years at best maybe every 10 years or longer depending how big your department is you don't get a lot of chances at the prize okay you have to hire someone and if you're a department like most departments were probably you there's a big pipeline squeeze bottleneck between being postto and being Junior faculty but most Junior faculty will eventually get tenure so you hire someone they're 30 years old you're going to be with them for the next 40 years right so there's a so a conservatism Creeps in is what I'm getting at you don't get to hire many people when you hire them you might very well be stuck with them as colleagues for the next four decades and so it would be nice to take some take some chances take a shot here and there try you know roll the dice a little bit but there is a natural conservatism that Creeps in because you don't have that many chances to hire a theoretical physicist and you want to be relevant and Department uh going forward and you know maybe you can say well a very good person will start off in Paradigm B but then eventually realize it's not working and switch to Paradigm a sure that's possible but it doesn't it very often doesn't happen I don't want to say that it doesn't often happen but I should say it often doesn't happen how about that right people stick with the the basic Theory the basic framework in which they've been working for a long time the people who believe in the steady state Theory kept working on the steady state theory for decades after we discovered the cosmic mic great background okay so what happens is that because you have individual departments doing hiring and they don't get to hire many people if there is a feeling in the that one approach approach a has a 90% chance of being right and approach B has a 10% chance of being right roughly speaking you hire 100% people working in approach a that's not always true because people are quirky and etc etc but there absolutely is going to be an enhancement of the bias against people working in approach B it's exactly this is very familiar from voting theory in a

4:47:26 democracy right yes the latest uh with this copyright infringement uh Court actions which are happening now with the AI llms whe they are actually infringing somebody's copyrights either when they are trained or when they are being used and uh this article has quite a many points in there so it's just uh interesting to see where this is going one of the things which stick out was that uh open AI claims that they don't have a direct financial interest in uh allegedly infringing the copyrights of the writers or whoever the AUD are of their work so even if there's a derivative Works which are done with the models uh they're not benefiting that's what they saying we starting to approach a point uh which has comparisons in history for example let's take the movie industry previously there were a lot of sets and it was expensive to build the physical um special effects and uh settings and now a lot of that can be done with the FX it's just computer graphics and uh it starts to look real or even better than the real similarly previously films you had physical films and you were using chemistry to make the pictures appear and if you need it to fix you need to have a physical fix and then became the digital here's an example where previously you would tried to do some mechanical means uh even though they're digital so you keep the original and you just enhance somehow modify but you try to be through to the original you not doing anything new you just uh change a bit what's happening but uh this is not what happens now all those things um oh

4:49:31 and ice pop I actually decided that I would treat myself and buy an ice pop and an edit like that there's no audio engineering degree in the world that can fix that it's just a total intonation mismatch so sorry you can't make that edit got to find another path or just re-record that's where regenerate comes in I just click the gap between the words and click regenerate and new audio cells are grown that make it sound absolutely seamless I actually decided that I would treat myself and buy an ice pop I hadn't had one of those since I was in like grade school so we can now compose the current U thing the art content whatever it is into ideas and then we reconstruct it and we just take well the idea and the rest is uh completely new but this way we can make it even better we can enhance it way more because we are not limited to what was there before and this can can't be one to one with the original by the looks of it or the sound of it not yet smell of it when it becomes more coste efficient faster easier to do the Reconstruction regenerating the original this brings new opportunities and possibilities as well we not exactly in the B up Scotty situation yet or for a long time for that to happen happen but the idea is pretty much the same you can take the old archived analog movies for example and completely reconstruct them you take the audio Fingerprints of the actors you transcribe it you take the scenes and you redo it in full HK and all be suround of whatever audio you

4:51:35 want to put and put the nice special effects and all those cracks and Pops and weird visual artifacts are gone this is already happening uh with the advertising world it's cost-saving faster uh to do just uh AI generated pictures and you don't need to find the location the models all the things which Go With It uh the makeups uh hair clothes there's a lot of uh Logistics there's a lot of uh economies behind this as well we don't need that many resources anymore because previously you needed to physically do those things you need to acquire them it takes a lot of resources to do these things and now it's just pixels and it looks as good or even better and we're not even talking about the visual protype of uh AR world where you can look around and you can walk in it and it's a 3D World so you can touch it and you can interact with it and that's the completely different level and with uh decomposing into ideas those movies old ones for example could become 3D worlds how cool is that maybe in a decade or two it's the same situation as with the LPS vinyls and C s and downloadable media streaming audio what is the original is the lp better is the original actor with the intonations and the cracks and pops and uh weird thing is happening the real thing or is the digital AI regenerated version which has the intonations and it has all those weird things regenerated

4:53:40 the real thing the better thing an example of tremendous help which is already now usable with AI is translations if you're having an app software app and you want to localize it to the native language of your users usually that's done either with uh paid services or with volunteers doing it uh with the help of some platforms where you do it manually or you have some deep Google translate assisting in the process but it still takes a lot of time to do that now you can do that just uploading the file to these um llms and boom it's done it's 99% in there you have to check it out but it's done and the crazy thing is that if you need to translate your website or you need to translate some user guides or stuff obviously you can do that the same but then you can take the Lexicon from that app if it's special terms you're using special way of the words you're using that llm model is going to remember all of those thousands of strings and that takes a lot of persistence and work if you do that manually each and every time checking that there's a consistency in place and you put it to the LM model and it happens in seconds that's from the two Appo from last year they are performing live audience they are usually fully booked and being already making probably 200 million same amount bit less to produce the tool and what is unique about it is it's completely virtual I mean the actors avatars Clubhouse was almost dead after they initial success in a covid era when millions of people and uh the huge funding um manage to capture the perfect uh storm in a sense that uh everybody

4:55:43 was at home and that was the place to be yet that's not the thing anymore and they were really quiet for a long time and just now they released the latest um change uh which is interesting uh to to say the least the people it from the live room synchronous audio into asyn Chronos friends and Friends of Friends Model and this has left almost everyone in this thread by the founder completely don't found it and the response is almost 100% negative even before the latest release it was extremely hard to find the room which US even have had uh bookmarked so that you will get notifications when it's live I remember earlier this year there was one room which was scheduled to come uh at a particular time every week and I was at that time there and I couldn't find the room so it seems that people been using other social media to synchron to syn themselves to to find the the rooms and talk about and and share the links for those rooms so they made it extra extremely difficult and hard to find and use the main function of the app which is which was the live rooms and apparently they didn't care too much because of this new release one of the big things uh which is a bit surprising is that they completely ignored the following people were building and I think that was one of the appeals of uh clubhouse to discover new people discover cool people discover and learn new things from people you didn't know and that that's was the magic of Clubhouse getting to know nice people

4:57:51 and New Concepts and being positively surprised and now they completely ignored all those followers the following base you gathered over the years in the blog post they mentioned that uh they will make it so that you have a room where you can stack all your followers so you don't need to friend them so they're not you your new friend list yet they didn't uh make that in their release available so there's no way you can do that you just hope that it will come someday they were quiet for months and now they come out with this thing it's kind of working but it's not really working so people are really upset and if you're using uh intro walk through guide for the new features uh it makes you to go through the list of friends um I don't know how it was designed and then you can choose and select the ones you want to keep and then you can accept them all but after that there's another list of friends which says friend requests and you have to go through them one by one and if you go through them one one by one there comes a point when it says that too many requests and they block you so you cannot even accept those people there so it's really weird what's going on there are they not really friend requests is it just some suggestions named like friend requests I test it a new feature it's easy to push the button and start talking and then you have a new chat room it says friends or friends but I couldn't change that to only friends uh so apparently friends of friends can see it and uh I don't know who can talk in that room some other chat rooms uh it say that invite only but I didn't see anywhere that option and I couldn't

4:59:53 change it so I don't know how it works the new release is almost uh 101 carbon copy of air chat they've been using even the same vocabulary which is in the essence of air chat uh you can see in their release text there are terms they use friends of friends friends dinner party house party there's a lot of exact same things and it's hard to say that this are coincidences it's not hard to see why people are upset you were having live conversations which happens right now with everyone who is in the room yeah you can reply them later and this is the whole concept of Clubhouse then suddenly they change the whole thing into Snapchat type of casual with friends and only friends type of conversation social app but it's a completely an entirely different way of communicating if you do in publicly you don't know these persons here some of them may become your friends but it's more a public forum where you're talking it's not your friends Forum it's it's not your closest friends where you're sharing your details and uh your everyday life this is not what is it about and suddenly you switch that it would be the same like Linked In turned into your private uh communication platform you ask who are these people well I know them from the professional side of my life they're not my friends some of them may be but not all of them and this is not the way to do it and you just cannot switch that and expect that everybody would be so delighted about it so it

5:01:58 might be that they are obviously losing they already lost because of Co went away and people started to go their ways and they're not in their living rooms anymore so the live um room situation was already going away and but there were still cor people who were really loved the concept and and were there the true fans now they're losing these people most likely because it was already now really hard uh to to use the the software the app and now you completely switched the model so who is the new Target group who are those people well I'm not sure that you can count on the original users because you somehow they feel betrayed and they may have some memories and feelings about the covid era and they may not all be positive or you may want to not touch them uh in the in the years to come so you have that Association as well maybe it would be better to change the name of the app as well they're also dividing people to two categories those who are your followers the made us Instagram's broadcasting channel and then those real friends uh you casually chatting asynchronously and having good time so clearly it puts the app a bit in a conflicting situation in a way that uh who who wants to have those live things and I still believe there's a place for the live synchronous discussions Twitter space is is still alive and it has a place the founders are and were in a really hard place they had a instant huge success and uh tremendously successful funding grounds as well early on and then the times changed the the success went away the demand sifted and

5:04:05 they needed to do something uh with all the money they have and the company they have and the vision and and they've been a long time in the field so it's not just a whim of doing this thing they were doing things along those lines for a long time so this is their passion and now you need to do something else we will see whether they can pull this through but copying it at and completely disregarding the original model with the true fans and loyal fan base even though it's been shrinking by the months and most of the people have gone away but you had a lot of Goodwill there was a lot of uh credibility there was a lot of trust uh a lot of good memories and executing it a bit well you you can fill the gaps it might be hard to ear trust again probably you need to sift the company away completely and have a refreshed start with the new user base and the money in the bank you probably can do that but uh it's interesting to see where things are going especially if you're just uh copying what some somebody else is doing because that's not having your own Vision it's following somebody's else's Playbook uh and you're always just responding and you don't have the knowledge and the experience of executing your vision and learning from there and and creating new things it's hard to make it to work when you're copying someone just the recent examples and uh we are still in the middle of it so there's no Final End uh results yet but for example Tik Tok uh was the one who didn't come up with the idea but made it

5:06:07 popular with the s videos now YouTube uh is feeling threatened with the long format videos because the swords are taking over and the long ones are going out of fashion and obviously they did their own variation of uh uh Tik Tok shorts with their own SS similarly meta has the short format reals but it's not exactly the same as Tik Tok they cannot copy the culture they can not copy the the feel of the Tik Tok the latest copying case is threats getting 100 million users but losing most uh if not almost all of them in a matter of months after coming out is an example again that just copying something which works for the original crowd which is different uh Instagram posts are not exactly Twitter type of uh news and instantaneous discussion of ideas and talking it's completely different format so the parallel here with the clubhouse that you have live audience at the same time in the room talking about things and then jumping to different thing is not obvious that you can manage with the resources you have if the idea is not supporting the use case and where you've been known before what feels fair and what doesn't uh is difficult thing here is an example where book 3 data set which is copyrighted material for example had made stale or those 33 books were part of that data set so now they are part of the large language models I think open AI was using it I might be wrong so these are part of now the the results you can uh get out of uh these uh

5:08:10 llms but ideas are not copyrightable and over the course of human history we always been copying things it doesn't sound fair but yet if you're not claiming to be the original isn't that the way humans are but what these copycats usually are missing is that uh the thing which makes it unique the thing which makes it successful you cannot see it from the outset it's something in the process something in the learnings which is hidden intact in the knowledge of doing those things in some Fields uh it may give the advantage in others it may not but if you're an artist uh would you like to be known only from your old work or would you like to innovate and create new stuff so the copying somebody copying your old stuff doesn't really matter that much app software are also living things in a sense that uh the market demand is changing people's preferences are changing the set which is working now may not work in the future so whatever you're copying and if you're only copying somebody else's work you're just reacting and you need to and you cannot copy the the culture the use case the the audience and the feelings they have and the connections they have with the product and with each other those are the things you don't see by just copying something and nevertheless what do you see is already months if not years behind what's to come and if you have stopped innovating if you have been just uh freezing and and Frozen in time and just hoping that you will be successful in the future without uh doing something new you're already out of the game earlier today T2 which is trying to get uh the Twitter audience or those who are leaving and and doing something better

5:10:14 but still uh having the similar features as Twitter uh they sent out an email which they apologized a bit later on saying that sorry we didn't intend to do that but they've been saying already early on that if you're not using your um social media account uh enough uh you will lose your short handle if you have have one of those I don't remember maybe it those five characters or something and you will have your longer one which you can Define early on so they sent that email already and uh I'm not sure that that's the best way of reminding people that hey better be liking and posting or else uh because you already telling them that we're going to take it away just a matter of time so keep on posting and putting up stuff not because you like and want to do it but because otherwise you're going to lose something what else could have they done instead of copying airchat one of the customer groups which was really successful loyal and still is and I just picked in is completely confused is artists musicians who want to play and connect uh with the the fan base say audience and also get more discoverability and obviously new fans Spotify uh in the co era was in the perfect spot and place and they even have the platform to do live audience uh music related things and even for just music consumers you could share your music lists and talk about music and Licensing should not be an issue because Spotify has already that covered but they're completely pumped they launch to the market and they just close the whole up down so this was one of the quite obvious but uh maybe there's

5:12:20 some reason they didn't go for that Pats clubhous could have extended and gun because if you want to do live virtual concert you could have a monetizing model for that one as well and there isn't really too many platforms who are doing that properly so there's a huge gap in the market for artists who already know and play Live which was the clubhouse audience and the clubhouse musicians and if they could make money if they could reach out and if Clubhouse would have gone deeper to that market and really serving these artists and musicians they could have a chance and the future it would be interesting but for for some reason they didn't want to pursue that route So currently the artists are a bit confused because you cannot play music with both of your hands if you have to push the button and if it's asynchronous it's not the same thing for example two musicians playing at the same time that doesn't work if you're doing it as if you have a house you probably can have a room and do the live thing but it's quite confusing at the moment so seems like uh it's not really the the market to go one additional point is also that the club deck which had over 100,000 users a lot of the users were the pro users who were having good rooms and also musicians because you could use your audio interface more easily plugging it in with your computer and this was also something which happened around Clubhouse and became part of the ecosystem all right so I'm on a bike I'm a bit winded what I want to say is like the obvious statement I think which is everybody is copying everybody there are many apps well before air chat well before clubhouse that are exploring social audio one that comes to mind is cappuccino this is an app that actually functions very similar to the new

5:14:23 Clubhouse except the big difference is it was a private group with friends where you would record a message and then it would batch it together almost like a mini podcast I also invested in a company called TTL that was all about more like phone calls like airpod based phone calls with friends that ultimately didn't work they pivoted to Paparazzi but what I want to acknowledge is like yes if you're copying someone directly you're probably not going to win but we're all copying each other to some extent so we're all taking ideas from one another the patterns and even the muscle memory that people learn using these apps like you don't want to fight with that too much so anyway just had to say it while some of the patterns are the same the actual difference between air Che in clubhouse the new Clubhouse is quite different you know I think somebody else mentioned this in another chit but really air chart is about finding new people it's more interest based clubhous is really focusing on friends friends of friends in some ways these are just different use cases in different networks entirely of course if both grow and succeed they're ultimately going to compete friend graphs move more into interest graphs interest graphs move more into friend graphs there's also an argument to be made that perhaps the new Clubhouse help air chat if you think about it like you know Instagram wasn't the only photo sharing app there are tons of them out there there many of them exploring different filters and a lot of that led to maybe the adoption and exploration of these types of apps and the exploration of more social audio might lead people to try more apps and the people that want to try Clubhouse might also want to try air chat so might be a win-win at least in the short term yeah we all learning from each other and that's basically how we build new stuff as well because I was trying to come up today with something which is not based on anything I know already and I'm not sure actually whether it's even possible to come up with something entirely new which is in some shape or form is based on some Concepts or ideas that you

5:16:27 already know I was also thinking that u in some sense air chat could be the new Twitter and by Twitter I mean the Twitter which was up to date with the latest things happening where people were really analyzing talking responding like we now doing about the clubhouse there were a lot of other Vision Pro happened we jumping and talking and analyzing and it's it's really interest based it's it's idea based it's uh about those things and then you find like-minded people to talk with and discuss and this is uh What uh air chat is is is about so yeah you can do the FR stuff but uh and you obviously start to talk with the people you you are likeminded but it's also about new people and new ideas so it's a curiosity app in a sense I think it's actually win-win because it's also going to segment the market uh not everybody likes the discussions what we having here and there are certain type of people who are more about different type of things which uh been happening quite a lot in uh clubhouse and uh happened also in spotify's uh thing and that's probably not a bad thing to segment the market to different uh interest and different people yeah I think they copied a chat and even the interface looks the same the async messaging is uh the same and I don't know why they did it maybe the audience was not retaining on the platform I figured it out uh how you can actually make a live room or you can change the settings of the chat feature whether it's friends or friends of friends instead of pushing to talk you tap and then you get a selection menu where you can choose the settings but you cannot change them later down so you have to do it before you start the

5:18:32 room another buzzling thing is that when you're building something and then some people come unless you're really specifically ex excluding people and just picking and choosing meticulously who can come in um is it the users who are choosing the app or is the app choosing the users and this is evident in some of the apps I mentioned the S Soundwave and wisdom as well they had a certain use Bas in the beginning and I'm not sure they actually managed to break uh to the larger market after the initial users and when you have certain type of reputation like rep restaurant may have a PA it's hard to try to attract different type of audience probably the same a little bit with the apps as well which direction do you think is harder to do layering like relations on top of a interest graph or extracting interest from a friend graph right I think of creativity as being able to apply one concept or insight into a different discipline or you know bringing two concepts together that haven't otherwise mixed um that's also what created Homo Sapien sapiens when we first realized that we can improve upon the tools that we had developed over the last few thousand years um and that just kicked off the birth of innovation the idea we can just we can actually improve things and you know to improve something you have to start with the thing at first and hence it might be called copying but I think it's creative so I know a lot has been said about a CH and Clubhouse but there were some things I was thinking about through the day and I thought I'll add them here um I'll start with just my understanding of Clubhouse and how it

5:20:34 has evolved uh one of the core things is that Clubhouse was built on a very different social graph and very different positioning uh when it started uh famous people with influence came and since they have external Social Capital which is on Twitter uh or on Instagram they brought in a lot of users who want to listen to these people just to get a feel of oops being in the same room as them that was how Clubhouse got its growth now what we need to understand and remember is that um these famous people got a lot of users on the app but they were not users they were mainly uh listeners who were coming to tune in uh as the co and pandemic thing sort of uh calmed down the famous people went back to work right and the amount of time they had earlier is not the same amount time they have now they had later and eventually they started speaking lesser and lesser right time liquidity is something that matters in a social network and uh the busier the person the less of the time that's one of the reasons I think this struggled in after the initial Spike another important Point here is that um all these people who came in because someone famous for speaking were only coming in and listening and not really starting their own rooms uh and remember at that point was just sink spontaneous rooms right uh mainly because they had no one to talk to no rooms to start people tried got bored tried got bored um rooms uh sink spontaneous is a very was a very new behavior and not everyone is a natural creator and so it wasn't just

5:22:39 a natural thing for people to keep going when the first room was awkward or alone or lonely Clubhouse structure was syn spontaneous so it was a spontaneous togetherness problem that they had this is a term coined by someone I know called Samir Singh spontaneous togetherness so you have to it it was a spontaneous room and it only worked when people were together in room otherwise it was just a boring room right um then they went from spontaneous uh sorry then they went from uh uh to syn synness and schedule rooms because spontaneous rooms were not working and schedu room started working a little bit more so it solved one problem it solved the problem of uh everyone coming together at any random time uh but it wasn't enough now creators and people who saw value they kept on going they kept running rooms uh in fact there are some rooms which are very high quality and they do very well um and they kept creating they ke kept creating rooms kept inviting guests lot of these things happened listeners tuned in right so people listening stayed there just because of these high quality rooms right then clubhous said oh hey we'll introduce houses which made lot of sense right uh because they wanted to Target the friend graph and this is where the friend graph concept really started for clubhouse at least in my understanding uh if one of you has another theory please to uh correct me if I'm wrong uh it's it didn't work because again it needed people to come together even though scheduled but it needed people to come together at the same time for magic to happen so that means you are limiting the chances or you're you're really really playing or really creating a very tight situation

5:24:42 where conditions to derive value out of the product depend on a lot of variables right U so that's one of the other thing they did introduce something called replays but there was too much work it's too high friction for someone to go and listen to replays it's like a podcast I understand it's cool it's nice but hm will I spend 35 minutes listening to a replay uh of a room versus listening to a podcast uh no I won't you know uh so that's something and because it's not like the replays you you actually really knew what's going on in the replay and it was too high friction and it didn't solve their problem right they were bound to go async by the way Clubhouse was going to do this because this is the only way for them to make audio work right clubhous structure uh which is the graph that they were working on was became friend based friend based graphs are very interesting of course everyone here I think knows about it but there's no concept of uh there's no problem with data Decay right that means uh when you're looking at Instagram stories of friends Instagram stories actually get archived or get hidden but you don't care because what you really care about is what your friend is doing right now what they did last weekend what they did last week what they did last month maybe that's about it right and uh that means that well okay come together have fun and move away this is something that I think they're trying now let's see how that goes so air chart is different in many ways it is an interest graph although uh there is an element of finding interesting people and then sort of working with them but there's an interest graph an interest graph means that chits or messages or anything that is created here uh will remain valuable for a very long time the challenge will

5:26:48 be to figure out a way how to keep surfacing the data that is relevant to different rooms uh so you know how do we do that and that's a product challenge that can be solved edart is also async from day one so people who came here uh I think we can of put them into two or three uh buckets number one for example I think I came on about a year back and when I came on the product looked very different so I've been looking at I've been seeing aart evolve for a long time um so one of the core things is that um people who came here were either explorers just looking at a new product or are people people who are active creating rooms engaging listening doing things and then there's a third coin which is uh let's say I invite a friend and they come on just to just because I invited them right um so it's quite different um a last Point here is that it's very very hard to introduce a new interaction to the same network of people um I'm not saying it's impossible but I can't think of a good example right now for example uh threads was a good one right I mean uh it's a good recent example of why it doesn't work you can't just induce a new interaction to people who are used to something else um Al I'm not saying that Clubhouse won't work or can't work all I'm saying is that it's very hard because this is the mathematical thing uh people have been coming together for x and now suddenly we are giving them y will that work will that not work who knows right social apps no one knows these are some thoughts I had I read uh the other chits uh and sort of heard other chits in this room and other rooms about this and this I'm we all know how this works and uh but you know not to beat Clubhouse or say oh it

5:28:52 won't work or anything like that it's just a very very different game uh although it's interesting to see them moving into this now but that's the thing uh social apps uh are a lot of luck lot of art but it's all they're also math right you define certain conditions uh and you say hey if you make these conditions happen then the prod will be wonderful now maybe we can force those conditions to come true once and maybe second time but can we create a product that replicates that again and again it's very very very very difficult and then again is a concept of bordom date which means most social apps get boring so that means they can't just be a status game or a novelty game or a fun thing they need to become utility sooner than later for it to keep its U value but yeah that's about it these are all my thoughts I hope they help feel free to uh counter or add anything you guys think at forgot to add a point here when I say air chart has been asynchronous from day one I mean that in a way where while everything else has changed in air chat and evolved the S the nature of the core purpose of air chat was async from day one and that hasn't changed that's because that is the value of the product uh and principally and mathematically that makes a lot of sense to people who are here because we're using it for that there's no other reason right so that's the point that the current network of air chat and the current graph or current sort of cluster of users who are using it giving feedback uh learning

5:30:57 about each other learning about new topics we are learning a new way of interaction uh which didn't exist uh earlier and that makes it very valuable because when this network and when this this cluster grows it'll become more and more valuable for more and more people well it's true that in Tech we do uh quotequote copy from each other I think it's more about Concepts it's one thing copying Concepts because you just build up on Concepts and the concepts are you know stuff that people have been thinking about for a long time many people uh the actual implementation cop that um is I think unethical and I've always thought that's unethical uh you know but there's a lot of money involved uh look at some of our brothers in Germany you know they made a fortune uh uh you know just literally copying us startups um you know ethics goes out the window when there's money involved Kamal what do you think is the ethical breach here copying I mean isn't that what humans have always done I think that uh copying is fine it's uh the execution the originality the constant iteration I don't think that for a second that uh clubhouse will be able to compete against airchat I I think just because the people behind airchat are much more creative they're going to move faster they're going to be uh more iterative and they're going to learn faster Clubhouse has been around forever it's like a dinosaur yeah they can strap on some Aon uh uh characteristics but they're not going to get to be Homo sapiens sapiens I don't know if I stretch that met met for too far I hear you I'm not a big fan of ripping someone off coping something one for one in fact that's if that s strategy you're probably not going to win anyway but the challenge is where's the line there's certain things certain patterns certain ways of interacting with technology or apps let's say that become so normalized and expected by users it's actually in some way similar

5:32:59 maybe to you know certain brands like Kleenex for example people call tissue paper Kleenex uh but Kleenex is a brand and so there's sort of this this progression perhaps of new patterns or new ways of interacting with technology that also become the norm they almost become cultural and uh in the public domain well in that case of Kleenex versus tissue that's just branding uh which is a a different exercise um as far as people blatantly coping you know there's a like like in legal terms the sniff test you can sniff and know if it's done or not you know it's just common sense there's nothing really complicated about it there's nothing to be done about it but they copied the microphone button they copied the push to talk they copied the chit line they copied the recording indicator around the microphone button they copied the way we coin stack the faces to indicate how many people are have spoken before you on the chit line uh they copy the multiple speaker heads they copyed a lot man it's actually hard to see what they did that's original other than just put in a bunch of viral loops and share your links out they copied friend of friends they copied friends they copied dinner party house party terminology and they passed it off as something new they literally said we made something new so yeah you can say what you want there's copying and then there's photocopying I came back just to hear the discussion about this cuz when I saw the new Clubhouse update I was like oh my goodness this was literally a complete r off of air chat so the example that comes to mind is um perhaps one of the most trendy Industries which is the no taking uh space I think that the I think that what Rome research did uh when they populated the whole idea of bidirectional um links and all this

5:35:04 concept of knowledge gra which I'm not really sure they invented this concept but obviously um they introduced it to to the masses and kind of the wave afterwards that brought up to life apps like um reflect obsidian or logc um kind of seemed like everyone were copying each other so at the beginning I was really convinced that you know one copied Rome and then the other one copied copied Dam so I used to think that this is an equation of The Winner Takes all and we live and walk in an industry and an environment that sanctify um the big Winners but I think that there is a room for um many apps and many companies in the same space so for example um Rome is a VC funed company uh whereas um obsidian for example is not funded by investors but rather by their community what is what they cannot copy what is unique about this place and what makes people come back over and over again Clubhouse coping air chat is pretty hilarious but ultimately doesn't matter Zuckerberg copies things all the time Microsoft Google copy things all the time ideas don't matter even execution of new quote unquote new ideas doesn't matter uh the only thing that

5:37:08 matters in the long term is real continuous innovation obviously this is more nuanced when you talk about things like YouTube being created at a point in time where there were no other options and they ultimately got critical mass and scaled and you know built networks Network effects around that data now the the barriers for competition in all forms of Digital Services are lower than ever so um ultimately I just keep going back to the only thing that matters is a real continuous sustained pace of innovation and copycats uh will ultimately just accelerate the pace of commodification did they steal my faces did they steal my face no they everything may be a remix but you still have to remix it that is some blatant uh ripoff really o o o come back here with that no no no no no so Petri I apologize but uh my new best work ever to stopped and uh I've watched this at least 70 times today no big deal I love it Richard what's wrong have you been crying hey which one was it it was church candy wasn't it you just brought piss to a [ __ ] fight you little are you get in your [ __ ] house nowadays people say copy replica whatever L of this [ __ ] um no they stole they stole everything even the phrase we made some something new but you know what the thing is to Ste the interface of air chat

5:39:18 without having the philosophy of airchat equals zero profit and suicide as a social media so we'll see not that it would have like a huge impact on air chat's total addressable Market but it would be pretty gnarly to see air chat maybe you know take on Live Events um to kind of just steal user base away from from clubhouses thiso it would be like the ultimate jab also to add if um you wanted to make it even cuter you could probably you know reduce infrastructure costs and bandwidth costs by only making it scarce for that feature in terms of like how many Live Events can actually be hosted on the platform maybe at a per user level or at an app level per week um so that you can keep the quality of those live audio discussions good such an interesting discussion uh greetings from New Zealand everyone I'm always um a very interested and curious listener um um but I think that there might be an important distinction to make uh in this particular discussion Ryan's coming from a from a perspective from a market perspective where we discuss the duplication of idea methods right so there is a problem that's solved by a common idea and then you end up having with similar products as a product uh designer and Industrial designer myself I I can I can empathize and resonate with what navala is saying which is more in the case of pretty much onetoone carbon copying or duplicating a designed experience within the realm of that same idea problem so

5:41:24 uh the examples that were that were mentioned before um like uh when Facebook copied uh Snapchats uh first invented story mode and then they basically took ownership and nowadays when people when people think about stories in their minds I'm pretty sure the most common image is that one of Instagram because of size popularity exposure they basically hijacked the design for for for themselves um so there is a there is a big difference between stealing someone's experienced design as sort of like a property and then embedding in into yours and claiming it it's of your own design versus that um inevitable point where different projects different companies different brands trying to solve one same problem end up with a similar product um when when you start hijacking other projects or other other products experiences and then embedding it into your own that's when problems start because that that goes beyond that thin bord line that separates we're all copying each other we're all drawing inspiration from each other uh to a a more dishonest way of building or rebuilding your own product based on the experience design that was created elsewhere so uh I think nval has a very valid point in terms of how the experience the designed carefully designed experience of air chat being copied um and then stamped on um to clubhouses new design and then the following claim that it's new and

5:43:28 that they invented that that is that is an issue um yeah there's so much to be said about this but in summary I would I would say that there are two big stand points here one is what um Ryan uh explicited with like you know we're always drawing inspiration from each other and the inevitable fact that we're trying to solve a common problem we end up with common Solutions versus nal's point which to me uh being a a product and Innovation person myself is a bit more critical because it speaks to a more Insidious character in that uh duplication it's funny because when I recorded my initial objection I hadn't even played with their product that much and now realize that there's even more that they copied like floating speaker heads and tap back and forth and even my launch phrase we made something new so yeah anyway we're going to go back to building I don't really care about them they're they filled and uh they can chase and copy there are significant differences still between the products and the approaches which is fine and you know maybe at some point members of their team will realize that they have better things to do than clone other apps uh gain some self-respect and go build something actually new and for the record I know that Ryan you're defending them uh and it's okay to be inspired but it's a different thing to straight up carbon copy elements and then claim that you came up with them in the first place it's just defies Beggars disbelief uh in any case I know you're an investor in clubhouse and so you have to do the right thing and defend them and you're morally obligated to do so but at the same time I think their claims are disingenuous hey nval weren't you also an investor in clubhouse I that was the impression I got I could be completely mistaken uh I got that impression because you were so active on the app I

5:45:30 think in the beginning uh and I expected that to be some kind of sort of promoting the thing that You' invested in but I don't know you tell me hey uh leond leond I'm sorry if I'm mispronouncing this I just wanted to comment quickly not about the subject itself but the fact that you decided that you're a curious listener but you decid to come up please come up more I really enjoyed your St style of delivery and speaking I definitely followed and am expecting your content elsewhere as well um yeah we have several rooms that I think you know your expertise and and uh the way you view things will be definitely appreciated Apple's about to release a bunch of things you actually involved you know Vision Pro uh yeah join us this is lovely I followed with pleasure hey Alex oh that is nice mate thank you thank you very much uh yeah you got it is leand perfect um I've been I've been a uh an attentive follower to airat for a while um I I made an acquaintance with Josh little he used to be the founder uh he is the founder of a now defunct project called volley uh which in my opinion is the definitive uh format moving forward board which is US truly exchanging you know our ideas via voice video or a combination of both um I got nothing but love for for a chat and I I just wish more people here in in my country in New Zealand uh used it um yeah lovely to engage thank you very much for the warm uh warm welcome um I'm following you back as well this is awesome thank you

5:47:36 hello all at a16z I just wanted to start with red and see who is uh still on the app um and to say hello um how's everyone doing today I did a little bit of U live TV last night about AI whether it's an existential threat do we need more AI regulation and obviously all the bad ideas are mine and the good ones are coming from the AI rooms and the brilliant people here and elsewhere do you really want to compare this with airchat come on guys I think that's um an insult for airchat because for me as a consumer this doesn't like anything uh on airchat and uh let's go into a room asynchronous room advice Alison to be The Rescuer can skip to the next speaker and I can tap to hold than to you and I have to wait till all the speakers are done before I can actually add something let me show you hope it's not too long okay and that's the microphone well that doesn't look the same so yeah that uh is what I call a very cheap copy then because no this UI cannot compete with the UI of edet get your backup onoro if you're still on

5:50:07 X uh sounds like that it was an inside shop or more likely a hack within the system or maybe one of the racks from data center moved by Elon Musk uh was tipped to another truck uh King aside ROM got his account back um he knew somebody in the company and uh they put the account on lockdown and uh the recovery process was successful yet you may not be so lucky so what do you think think of this what should we do except do the backups most of the business communication will become automated in a few years human interaction will become the exception it's already a bit silly to read any business posts or articles or blocks or anything of a kind because most likely it's been created at least partly by Ai and in a similar manner people are already decrypting that AI text into short summaries and trying to get something out of it when you're sending an email uh to request some further info about buying some services or you considering of selling something the same thing happens uh the whole transactional nature of business will become completely automated human output will become the new API so it doesn't really matter whether you're speaking in which language you're speaking if you're speaking whether you're drawing a picture whether you acting are you using some other means to communicate apis Those computer apis are no longer needed any form of of communication can be used as an input and it can be put as an output so in

5:52:12 this way the whole business becomes quite interesting from the automation perspective how much of the bureaucratic nonsense uh for humans signaling can be taken away all the waste can be reduced because it's just computers talking to each other and it's just about the getting the business done this is the ultimate antispam because why would you generate any content or anything of that kind because you know that nobody's going to read it it's going to be just filtered away anyways so why waste time to do it at all and when you as a human do something it's most likely going to be recorded in and captured in some form so if you say something there's no need for you to record it uh personally it's going to be there the transcript is going to be there so if somebody asks what did that person X talk about in that event uh you can get that idea or these are the main points he was talking about and uh these were the counter points and he was referring to this and that from that that bit of in information from history referring to these and these things so you can put a lot of things in context and it can be even drawn as a cartoon if you like that way so we can start to focus more on the ideas the important things all some of us hopefully can focus more on those things standing on the shoulders of giants is how Humanity has been building the knowledge base to experience to know how and when all of that more or less becomes in the AI models and you can access it and you can get all those

5:54:16 things which came before adding something new becomes an interesting Quest how do you actually do that and what is it what do you need to pass on to others to take the button and uh run a bit further now when you're doing a peer review article you have to put a lot of stuff into references and establish your point and so that what is the new thing but in the future do you have to do that because you can imply or assume that it's all known and maybe you need to put some reference points but uh the AI can do the rest so you only need to State the really exact point and uh the rest is just waste in a way and it's just something which needs to be explained for the humans but otherwise it's just U vectors and reference points and weights in the combined knowledge of humanity how we reinvent our creativity our contribution and we I'm focusing our time effort and energy if and when most of the things are already done by computers in a way that you don't need to put yourself time to gain that knowledge or find it or learn about it it can be given to you and you can observe it watch it from different angles and just input the things the computers cannot do the Creative bit the common Universal explainer what makes us so different from the rest most

5:56:22 of the current creativity is mixing and matching and uh combining existing things in a little bit different way and interpreting them in a different way way but it's just using what's already there and sometimes by accident or some other coincidences there's something completely new but most of the so-called creativity isn't that creative in a way that AI can do it so mixing and matching and reproducing derivatives of those will not become that interesting in the future because you can just ask for it and it shall be given to you if it's based on something that's already existing and there's some reference points it probably can be reproduced or produced digitally the same is not through in a material sense our Material Sciences are not in that level but in the digital realm which becomes more and more of our day experience there's no need to go to the analog world to the material world and produce something and then digitalize it that step is going to go away and that will fasten so many things but it does not matter that much in the future because your sensory experiences are going to be augmented anyways so the digital objects start to look like real objects they're probably going to even feel the same with some heptic AIDS so in this sense what is reality what is really needed in the physical sense and when the distances

5:58:28 become vast meaning light minutes or light hours in a little bit further his uh future then it needs to all be digital if you want to communicate with different solar systems for example and then we are already in The Matrix or altered carbon or in a metaverse which is already starting to approach this happens uh too often to even consider thinking about it but uh funny enough yeah that's how we reinvent the wheel but uh Sometimes some good and crazy ideas come which could be worth pursuing as well hey would you rather invest on Legos or would you prefer to invest in a VC fund is actually a tough choice unless you really man to get into the top quarter or the really top VC funds otherwise you may be better off with the Legos our news have been broken for a long time but now globally the information sourcing for the news is really broken as well when Twitter came along it's became pretty early on as the global Telegram and no not that kind the old one where the information sources quickly uh across the world and you will hear the latest and directly from the source it was uh tweets but it was also Periscope which was giving the video and authentic voices for people around the world the power of Twitter was that you just had an account and then you could

6:00:35 uh tell what what do you see what's happening how you're feeling and people could read it and obviously there was a lot of and is a lot of noise but also authentic voices and you didn't need to be in English or in any of the global languages this worked for a while and it was a tremendous power which obviously the regular major media was also using as their sources but then it turned into a current manipulation and agregator and something else this original truth has been lost in the fight of spams and clickbaits and revenue profit motives people who actually have the news who have something to say and are just happen to be in the place where things are happening they may not have the verified account they don't probably are the in influencers are the ones with the large following they just happen to be there and if they would have a account and they could share their voice that's the the one people want to hear they don't want to hear about someone who is sitting at the desk thousands of kilometers miles away but this is what the current trend is and where the X for example is going I don't find this a particularly hard problem to solve in fact I think most of the people who are whining about it are mainstream media who just don't like being disintermediated the reality is there are plenty of great OS in accounts on Twitter who figure out what's real and what's not and I've tweeted out lists with these OS in accounts and anyone with half a brain can find them so yeah misinformation gets spread for a little bit but no less misinformation than was being spread by the mainstream media they're just annoyed at losing their Monopoly on misinformation but it's not just X or social media it's the old media the

6:02:40 major media which has become the Legacy Media losing if not all most of it credibility as well but as well the journalism they're mainly just opinion pieces nowadays M media used to have the resources to do the original journalism and that's what the news is supposed to be about not about uh secondhand telling what's happening and coloring with your preferred opinions or filters but that's mainly what's happening now and it is an opportunity it's a good thing but it's a change where are the news where are the opinions where are the people you're following well they have always been individuals you notice someone and you start to trust their judgment and their expertise and then you follow them and it doesn't matter where they are writing or reporting whether it's a Twitter whether they do it independently whether they have the newsletter whether they're writing in a wasy journal or some other major media but it's about individuals and this is where we're going back and we should be going back Major media is dead and they had their place but we have to move on so what's coming next llms AI customization everybody has a voice 8 billion people plus uh a few trillion pots so if you're not starting to get customized view what's happening today from the sources you trust and it comes in a way you like and in a in a manner and in a format in a length and in a discussion you're like if that's not happening in a few years few months then we doing something wrong we need a automized uh AI layer to sort out the spam all the noise from the authentic voices and what's happening and then

6:04:46 bring us the good news the bad news and all the news and the analysis but it requires that uh we can verify the sources there's authenticity proof of human and a track record news have always been automized pieces bits and pieces around the world gathered together and filtered analyzed by somebody but instead of that somebody being somebody else now it's either AI some algorithm or some people in between and then you do your own judgment but like in physics it's all relative there's never been any absolute roots and now it's just more in your control previously it was Major media who decided and was pretty much binary or few Alternatives now there are so many different grayscale layers it's not a technical problem it's a network effect problem currently uh the latest uh developments in Twitter actually have made it and the wired article illustrates this pretty nicely it's becoming almost impossible to find out the uh the voices the real sources because of that image thing and few other changes in the platform and currently there's no alternative blue sky has been catching up but it's not yet there the problem with the Alternatives is uh uh first of all a lot of them are still in paed so you cannot get access to it so it's not open for everyone so that's not an option mustard done is open but um it's really hot work and it doesn't have the network effect people just don't like to use it it's too clumsy to work blue sky is the best alternative but it's not open you cannot get access you cannot get invites so there are no real Alternatives Noster is open for everyone the clients work but

6:06:51 uh it still has only some tens of thousands of active users and that could be an alternative but it hasn't caught up yet Blue Sky uh doesn't have the capab abilities of showing video which is a huge issue givs and the memes is less so must done accepts that but uh the instances are isolated and interoperability is not a really good use experience nost does have the same problem as blue sky so the image uploads are difficult and it's a little bit too hard for the regular people to to work with hasn't had a perfect opportunity here yet they didn't want with their threats to focus on the new segment also they have blocked uh EU so some 300 million people and a lot of things are now happening into Europe so that's also completely out and there are also other reasons people don't want to get involved with meta right now it's the opposite of Arab Spring where Twitter brought people laser focused literally into the same geolocation now it's such de ation of truth that I think Nala is right about the list being it's the the agency the personal responsibility to create and carate those lists in order to find the truth and we have to seek it we can't passively rely on pushing some buttons in an algorithm to provide it for us I think for too long I became spoiled by the algorithms and the tools it was just magic I made a list and I deep dived and I learned if somebody else was smarter ahead of me in a field I borrowed their lists but I think now more than ever it becomes personal aggregation and the skills of research and discernment are very very important these are things I used to teach and now I see where we need them today I'm pausing because I want to ask

6:08:57 myself if my argument is that it's worse today the dissemination of misinformation and the inability to find the truth or if it's equally the same throughout all eras of history I'm thinking predominantly of yellow journalism I'm rewinding my brain to the American Revolution where Sam Adams wrote fake editorials in the in the the paper so I think it's equally incumbent upon us to come up with the truth and now while we're waiting for AI to reliably push that button for us after the age of magic algorithms I think I have to do it myself I think that might be my uh my conclusion the speed and the availability of both news and propaganda reminds me of a conversation I had with my dad oh about 15 years ago maybe even more I don't know a long time ago dad retired from uh he was a banking CEO and I asked why the markets were so crazy all of a sudden and he said well they're not crazy all of a sudden they're impressed in my day I had to pick up the phone call a broker then the broker would do something it would hit the market it would clear all that took time if I did something with the bank it would take days for checks to clear and all the things with that had to happen what you're seeing now is it instantane instantaneously hits a lot of times you skip the broker it's automated so if you took that graph that you're seeing and calling crazy and you stretched it out to the original time period that I dealt with it would be similar but now you're just kind of squashing it like an accordion and it's crazy and I always thought about that in in terms of news cycles and the availability of news too that we get so much so fast and I wonder well how can a human keep up so I'm waiting for the that magic time where we leave the magic Al algorithm Zone and everything comes to me and the next step being the magic time where it can

6:11:00 automate and curate to my liking to filter out that truth like like you said Petri that's it's a little long but that's what I think about this sounds awfully familiar the fact that it can go around and talk to every person on Earth understand their exact preferences a very deep level you know how they think about this issue and that one and how they balance the trade-offs and what they want and then under understand all of that and and like collectively optimize optimize for the collective preferences of humanity or of citizens of the US that's awesome even rather intelligent people have a hard time sorting the information there's just so much of it how do you tell the the true from the false the uh the clear signal from the disinformation um it's very difficult I think this AI layer that you talk about is kind of going to be where it's at uh AI in everybody's pocket but that AI has to be trusted right we have to be able to look inside and look under the hood and know that it's actually programmed uh and set up to be looking out for us rather than looking out for whatever Apple open AI Google whoever sold you the AI in the first place open source on device personalized models I think that's the future that's where it's got to go I think we going to have a huge change and we have no idea how the world will look like in 5 10 years our modern platforms and technologies will start to look like as modern as the 80s TV now because what's about to happen is not just a gradual shift incremental step but a change in the way and in the mode how we perceive interact and do things

6:13:04 I think it's going to be bigger than if you were in the late 80s and you were lucky one to have a modem and a PC and you could pay your bills by dialing up and then typing with your computer and doing it that way or in the early 90s and you had an email and now looking 20 30 years later the services and the ways are still somewhat the same they look more nicer today they're more convenient to use but the change is not that big there was already Windows 3.1 and uh interfaces were already there so you could easily use them now if you just sift it back from the future or the other way around the big disruptive change is the multimodality combined with the AI capabilities which are just getting started the 80s television was pre-programmed scheduled you take it when it's given and what is given your only choice is whether to watch it or not we still work the same way somebody designs the platforms and the services and you have a choice whether you use them or not but uh that's pretty much it the future ways don't have a interface they don't have a design it starts from a intention desire or idea and it's generated on spot for the purpose in a way and at length what is needed like the chat gpts of today which don't have the textbooks images in them they were just

6:15:10 the learning material photos and when you make a request everything is generated on fly based on those learned models the future interfaces are hard to imagine because they can be so different there are some familiarities maybe you prefer to get your information via rap songs your neighbor may not prefer the singing format maybe poems are better and your other neighbor is a boring old text type of person so they are like written messages and the younger one downstairs prefers cartoons and The Neighbor Next Door has some other weird way I cannot imagine they all get the same message they all are communicating they are getting their business done but it's done in a way in a format and at length they prefer and that can be unique the magic happens when all our current boxes that have shiny screens and buttons can disappear they not need it anymore so we could live like a village life in few hundred years back or that's how it looked lies because the technology will shift to the background it becomes part of the magic so if you would go to the Future and observe the people there they are talking and then things happen and you who doesn't have that something and as a Caro culture person mimic the actions you speak and then you expect the same

6:17:14 things happen but nothing happens what is that magic they do who is granting their wishes and things happen but not for you what are you missing I agree with this completely and although I think he got pretty lucky in actually nailing the timing from that far out uh especially with the Transformer architecture and the follow on like analytical tools that come with it um I think croww is going to be roughly right about 2030 20 35 for like what I'll call transformative AI because I don't think it's going to be alive but I think it's going to be technologically transformative will it transform The Human Condition probably a little probably a little um will we squander it probably a little um will it be a huge transformation spiritually I doubt it is that the next big thing could be 50 years 100 years and just out loud uh Petri I saw your little green headphones blip up I haven't been on very much hold on my sweet I saw your little green headphones flip up and I haven't been on that much it's nice to see you man it's nice to see you little uh little chitchat get chitchat top of the morning to you I hope you're well the easier it is to get information the harder it is to understand it why is this so evident today you don't need to wait for the newspaper to be delivered on your door in the morning you can just open your phone and all the news in the world are coming to you and all the information sources even the original ones are there for you now no need to go to library

6:19:21 find the index card and then get what is available on the C based on somebody's decision all the information is on your fingertips now the information retrieval and consumption space has been divided the oldest Generations are using the most the newspapers the traditional mainstream Media Fox News and the kind and the youngest Generations are using YouTube and Tik toks and what is common for both of these they are usually short formats and it is selected by somebody and they tend to be a bit shallow one-sided and to really to understand the issue you need to have some background and context and that usually takes time and this format doesn't really work for that type of uh analysis but it's not about the medium you can get good information from YouTube with a long form podcasts and documentaries you can get up to dat the latest information directly from the source from the micr blocking Services uh from Twitter mustone and other places it's about the skills to analyze evaluate and understand which sources are reliable and drawing your own conclusion having an agency taking responsibility what do you take in and more importantly how do you react and what is it what you put out to the world after you have consumed analyzed if any all that information a lot of factors are against balanced

6:21:26 views and more considerate and rational responses is starting from our biology which rewards the the latest new thing quick reactions negativity draws our attention gets our attention and we gray for those things short is always in some sense better than long long stos hard means working so rewards immediately sounds nicer it's easier laziness Works taking all that time to understand and build your perspective based on the context doesn't happen immediately it takes effort traditional longer form high quality content the weekly magazines for example have erod it over the last few decades so they become more clickbait and the declining revenues have forced them to go along the lines of the other media and the good cont content is nowadays somewhere else you can find it in substacks and other places but you have to do the effort they're not bundled together anymore in The Economist Financial Times weekend edition or any of these your favorite wrapped formats which used to be the good labels and brands for information they are gone combine all this with the traditional as old as Humanity factors power money politics and the new formats of

6:23:31 manipulating people and this makes this almost complete but I have a life you may say and correctly so we don't have time to figure out everything we not supposed to do that nobody expects for you to have an opinion about everything the easiest way is just say I don't know but that's probably the hardest you're supposed to have an opinion you're supposed to pick a side why a lot of previously private matters have become public thanks to politics from your perspective the singular items may be isolated and they don't affect you too much so there's no point putting much time or effort to figure them out but from the perspective to whom it matters who have an agenda they are counting on people not paying too much attention to the details because they are getting the benefits out of it and you are getting the costs and all the downsides of it collectively at least and there's an name for this type of behavior regulatory capture when we extend this concept to International matters and security it becomes way more serious way more escalated and way more significant what was just uh some local lobbyists money games or local politicians power games is now the big arena for dictators even with the hundreds if not billions of people under

6:25:35 them and this can have and has and is having D consequences yes you can live your life enjoy the little things the big things in your personal sphere but just remember sometimes things are not exactly how they used to be and sometimes things can go really wrong because everybody's just doing their thing and hoping that somebody else is taking care all this difficult thing all those things which may not make you look cool nice or there's something controversial which needs to be sa or you have to put up some kind of resistance you have to for an opinion you have to stand for your values and for your rights for them to survive and exist in the future what used to be your private life may not be so in a very near future if your basic and fundamental security is no more and this is what an escalation is and how it happens if you're not paying attention and you're waiting for a little bit too long to form an opinion or focus on things around you ignorance Maybe please but can you afford it in the long term Petri is always with uh terrific word smithing yeah ignorance may be Bliss but my god do you over pay for it he who fails to increase knowledge

6:27:53 increases sorrow I think it truly depends on your design for people in the academic sphere I do think that's that act you know it applies to them a lot more for obvious reasons that there's a reason that they are in Academia and for the truly curious they have no choice but to expand their mind because that's their design and for them it might even be a bit of a curse why does the pumers who are still in power influential politician or otherwise active and have the means to shape and interact with the cause of the societies why they do not do anything about the security situation at this point this puzzles me because their parents and Grandparents were involved directly firsthand with the world wars all the from all the difficulties all the hardship so they heard all the stories from their parents they lived some of it themselves and still they decide do nothing when the situation is what it is at the moment and if you don't care about Humanity what about your children if you have some or grandchildren or even great grandchildren or your personal life if you still have some decades to live for how does the future look like does nothing matter it's not about blaming and shaming I'm just genuinely interested and curious whether there's any answer for this or not it's understandable for the younger

6:29:56 Generations the Tik Tok generation people in the 20s or younger and also a bit older they great grandparents or grandparents may be already dead so you haven't heard the stories and the situation is also different if if your parents are directly involved so you you live that uh still in some shape of form but for the younger Generations Things become more hazier and just something from the history books the link um in the first [ __ ] in this topic was about 8 years cycle so we tend to forget things and regard people who came before us as fools or something which doesn't uh influence or affect us because times are different and uh things are not as they used to be and we tend to think also that uh future is always better than the past and the Cycles are are not repeating and we are not falling for the same mistakes the other Generations did history is full of stories and episodes and times and even Nations where the progress has stopped in some of it uh permanently and for the whole Humanity it's also been close call a few times and we've been stagnated for decades and Cent iies in the past because the forces of resistance to chains or closing the society have been so strong that other forces of openness Innovation and progress have not succeeded yeah I mean I guess I feel like um the Boomers have in fact reacted along the lines they grew up with which

6:32:00 is maybe not so much World War II uh that's a little more distant for them and the early boomers are likely gone um but um they re reacted to Ukraine right away with support and they reacted to Israel right away with support um you know and I think it's a combination of you know those were the historical hot points as they were growing up and we've all gotten used to economically being able to sort of stabilize situations without sending sending our own kids there for the most part um so I think it's a combination of many things but I guess I'd ask which security situation um you think that they're outright neglecting um but the Boomers we well the United States at least and I think the Western Complex if you want to call it that um slept on a couple issues that they didn't grow up with and they didn't have perception for so they definitely slept on China um because China was never was never a geopolitical concern at that level uh and there was an economic benefit to the trade um and they definitely at this point I think we can say they definitely slept on trying to go the peaceful route with um letting not Islam spread but to me it is clear now that there has been a large campaign for a long time to so descent with the idea that our moment will come and I think that's what drives a lot of these um these these worldwide uh protests so I think that has been fermented and I also think they learned from the Russians that propaganda um uh ra politic is um is more powerful to an underfunded uh underfunded I don't want to call them a state CU they're disgusting but organization I mean that if you in Europe uh the war the world wars are still around us you still see the things at least in those places which the where the fighting was happening also if you

6:34:02 were a boomer your parents or maybe your grandparents were directly in the walls and maybe died there or you know relatives which uh were injured or or died and the trauma which came afterwards and the rebuilding and all those things are in a living memory and the horrors of the war are not something distant but very close to you and by the security situation I mean that we are actually now probably already over the peaceful times period and this can easily escalate into a World War situation or instabil situation for decades to come where there's nothing but uh bad times ahead of us and this is what's happening now because we sleep walking through the situation the Russia Iran China all those uh axes of evil uh which they start to be by association by colluding with each other are starting to step up because we have forgotten what it means to fight for the freedom and the values we have and we are not prepared to actually stand for those things we don't have the military capabilities in place we haven't started to even prepare for the uh horrible idea that uh others may not respect our borders they actually might see us weak and take what they want to take because of the in know motives the history repeating itself the appeasement what happened in the Europe in the 30s for example is a good example we didn't step up we didn't stop and those who are imperialists are not stopped by words they are only stopped by actions by greater force and that's what needed to happen and we've been letting this

6:36:06 situation to escalate for years and decades and uh it's not going to stop if we are not going to do something about it it doesn't go away but because of the situation is something which is not desirable people probably don't want to think about it and they hope going to go away it's not easy in in a political way either so the populists and the politicians prer to do whatever they prefer to do and it's not to do the right thing or the hard thing only by observing the last two years for example what has already happened who did what and what did not happen it is very evident to see that we are not ready to face the reality that we may be fighting in a war soon and we need to start to prepare for that situation and if you you look for example that how is the military preparedness at the moment how are we capable of coping with in the situation where we need to start to actually pick up the arms and fight for our rights do we have the ammunition do we have the stocks can we convert our production we already have a really good case with Ukraine they are running out of cells they actually already are out of cells and the EU promis to provide them with 1 million cells per year which is not much they they would that's the two months uh stock for them if they are in the maximum capacity at the moment Russia is using 20 30,000 cells per day so we already see now that uh the orders are not in place uh capabilities are not prepared for the times to come there's nothing to say it's all in the numbers

6:38:11 and in the action what has already happened oh uh you are absolutely correct and Europe is entirely screwed and chewing their own arm off um I took it from a US point of view Europe Europe should be very very nervous um you've got macron as like essentially a Hamas sympathizer at this point there seems no Unity to really get involved and at least support UK Ukraine from the back end um most of your energy comes from um Russian pipelines I believe and you're absolutely right to hear appeasement coming out of mcon it's crazy it's crazy as for the people I kind of feel like they've been lulled to sleep by this kind of performative socialism wearing capitalism's cloak and also they are trapped away from the ability to actually even Institute change because they are simultaneously fractioned within their own countries bureaucracies have sprung up that spraw their countries and into the EU it is hard to get things done and I think so I think Europe is in an extraordinary amount of trouble you have Russia obviously that's a clear issue you'll absolutely get dragged into whatever happens in the Middle East if it escalated your economy is anemic no one wants to do business there there socialism has essentially lulled the populists to sleep and I'll tell you your biggest threat in 2050 Africa is going to have the largest population on Earth in the lowest demographic

6:40:14 demographic age and the first people who leave who leave will be the wealthy and then will come the violent and the desperate and they will flood flood southern Europe in a way that I don't think a single country has ever seen an inflow of population and in the meantime they'll be talking all this garbage and all the stuff will happen under the hood and yeah so I took it entirely the wrong way if I were a European I would be very very nervous part of the reason I have really clear and immediate um responses in Ukraine and Iran and yeah I mean it is Iran so but in the Middle East right now um is you know my family was in work camps my family died in World War II my family like suffered from it and then in the wake of it my grandfather moved the family to the United States when my mom was like 13 or something I'm the firstborn American in a Dutch family um so yeah I mean the list goes on and on um energy foolishness Financial foolishness virtue signaling instead of actually looking out for your population this whole confused bureaucracy and monetary system that is the EU and will crack um so yeah I hate to say it but I I absolutely agree with with you and I can somwhat see in it mentions but I can't click through to it um absolutely the US will get dragged into it um but we have a lot of advantages so it it it won't be a good time for the us either uh but it'll be significantly better uh in the US than in Europe and you know people my my opinion is like all these Munitions we've sent off to Ukraine and we'll probably send off to

6:42:17 Israel and we'll keep shipping out the fact is our whole arsenal needs to be rebuilt anyway so what we're not sending the money we're sending them stuff that's been sitting on our shelves and is you know 5 to 7 to 10 year old technology so in the wake of that will be a massive spending boom um which will update our Arsenal so it there's there's certainly a gap there where the Munitions are out and the new production hasn't come online but I have very little doubt that it'll come online extraordinarily quickly um so yeah it's going to be a tough time for the it's going to be a tough time for the whole world uh I I think um but for right now my initial response was that the US reacted correctly to both Ukraine and and Israel as the two larger issues right now um and then as it goes forward we'll see but I think for now we're doing what we're supposed to be doing and again part of what we're doing is we're getting credit for large dollar amounts that are actually going to uh cause us to resupply our own uh defense industry and all the money we spend on defense it goes to American companies American workers American manufacturers American you know um commodity producers so uh it's a type of spending that isn't you know net devastating honestly I think China stays out of it I think China stays out of it because all of the you know tier level uh geopolitical players are will be depleting their own resources and distracted with each other and China can just fix its stuff grow handle its Dem demographic issues um and related there's a sentiment that I've seen on Twitter and heard and here that like we're going to give back Taiwan that is never going to [ __ ] happen there is no chance that with China the way it is right now we will give up the ability to Harry them in that theater in the eventuality um that we need it and in all these cases of course we do it under the opes of freedom but I mean geopolitical

6:44:21 positioning matters as well um so I don't think cuz once they get Taiwan they'll just run the table in the whole region um so um yeah that's my opinion on that here's a nice except from a book that's already translated I think into seven languages uh it's about uh Ukraine and also about history China has been radicalizing lately in the last uh few decades and it's becoming more aggressive in its uh acts and uh well if you look also internally there's a lot of turmoil and like the similar way like in Russia the best way to give diversions uh is to point it outwards and there are a lot of things which are leading to the direction that um I think we are once again um maybe seeing what we want to believe and not what actually is happening I don't disagree and I won't pretend to like know it absolutely what I would say is that China China has delivered an increasing quality of life for their population for a long time and now that is in Jeopardy and if they really get into the united into it with the United States that's even in more Jeopardy I see China as being like not neutral but neutral bending towards whatever makes it more of a pain in the ass for the UN United States especially and maybe even Russia part two China and Russia are not friends you know they make talk to each other they're not friends China has a big problem with India a little problem with Pakistan so on the backside of China's issues they have these other things and with their treatment of Muslim citizens I am sure they are also not all that interested in seeing a worldwide Jihad happen um and again I'm not saying I'm

6:46:27 right but you know they they can't turn and just mess with us because they have other historical problems they have to deal with and they have demographic problems and they if they go to war they probably have economic problems as well not so much because War doesn't pump up an economy but because in that case us will ABS absolutely strangle them um and China has a clear path to dominate the world economy with their Rare Earth minerals and their size and their control so again I think I'm not one to think that China is always thinking a thousand years ahead I mean they've had three Revolutions in 125 years um but I I wouldn't see Strat Al that happening I think G wants to be an old man still in charge of a country and not go to war and then finally the cold calculus is if China really turns and wants to be at war with the United States United States is absolutely going to pull out Europe and say okay you guys are on your own over here um and I mean yeah I just like I I really think if push came to shove two things are true I think the US would obsequious ly leave Europe to solve its own problems which may or may not drag Britain into it um and then I mean saying the quiet part out loud of all the countries on Earth that if they have their back against the wall and a lot to lose and the best ability to maybe survive it and do another nuclear event it is the United States I mean that doesn't make me happy to say but I think that's true when something is familiar to you it

6:48:30 doesn't mean that it common is your experience of gravity the way it is elsewhere when you're deeply within your context you lose the sight what is not obvious evident or common when you learn a new language it's not just about the grammar the syntax the words it's about the history it's about way of thinking it's about the context our expressions are immersed with our context history and everything we know have experienced and because we are part of it we are like fishing the water it becomes invisible when we are training llms they have a lot of that encoded implicit knowledge within all the multimodal text video and every other form of Expressions we have but we don't see it that way we just think that it's just words or something is too obvious and we can easily mistaken that for its own Agency for its own intelligence it finds new different perspectives on the things we have given to it it's new for us because it's different but it doesn't mean that it's something itself it has came up with there are a lot of things we have not discovered because we haven't connected the dots there's a lot of knowledge which is hidden forgotten but now it's coming back to the Limelight again and these are the treasures we are

6:50:36 now digging up and getting the benefits it's time to build so I did some building myself as well uh here's a tool where you can cross post to muston nost and blue sky at the same time it's pretty handy if you have an iPad but obviously works with iOS and also with Mac uh I will post the screen cap to demo at it earlier today it's still in the early days so if you want to test I can put a after the video the link for the test flight so B want stuff may break and will break but I've been using it for a while and uh seems to be pretty neat way of uh sharing articles and stuff you find on the internet I'll just want to post once and have it in a different places here's a quick demo uh about the tool I've been making so you find some interesting on the web and you want to say it to other people and this is done on iPad so you can just track and drop text you can uh change the content and make it as you like and if you want to sell a link you can copy paste it or you can drag and drop it and notice that uh that's the slide over view on iPad where the the tool is at the moment and you also notice that um you have different platforms that there's a blue sky master and nost and blue sky has a limit of 300 characters so you have to make it a bit smaller for that one must don't have 500 so you can put a little bit more stuff in there no doesn't have any at the moment or obviously there are some technical limitations so here you see I'm changing a little bit of the text for different uh platforms and while I'm doing that you can also see that there's a link preview

6:52:38 you can also add videos you can add uh photos as well and when you could to go and you have finally find the way your way of uh with the words then you can just push the blue button and off it goes to all the networks at the same time and you can check out that uh those posts actually went out there yeah you see that now they all posted and this is the link for the test flight all and every feedback is welcome and I really appreciate if you can also comment on the things that you like and whether you would actually be willing to pay for this thing uh because I would love to get some money out of that one as well