The Most Likely Outcomes of an AI Future with Emad Mostaque | EP #55

coding is changing dramatically there will be no   coder that doesn't use AI  as part of their workflow 5G and starlink the dry Kindle for this fire  has been set is this more important than 5G   buy orders of magnitude is a magnitude there is  no kind of pure independent completely rational   person because we're not robots so medicine  is changing uh dramatically is your doctor AI   enhanced low insurance premium lower co-pay that  is something across every industry it's going to   be happening 10 years out what is a lawyer  what is an accountant what is an engineer   they are all AI assisted the entire knowledge  sector is transformed and it enables you to do   whatever you want to do learning should be  a place of Positive Growth and joy where's   the difference there where's the difference  and I think the key thing here is empathy we're in this beautiful home in the Hollywood  Hills and A lot's happened in the last six   months since we were recording our last podcast  and you were on the stage at abundance 360. and   I think this is epicenter for a lot of people's  concerns right now A transformation potentially in   Hollywood you were speaking about that that it's  going to change dramatically and I would you say   were haven't even seen a small portion of the  change yet I think we uh at the foot of the   mountain as it were I kind of compare it to being  at the iPhone 2g to 3G point and we just got copy   and paste we haven't seen what this technology is  really capable of yet and it hasn't got everywhere   it's like everyone's talking about it but not that  many people are using it there was a recent study   done um that showed only 17 of people had used  chat GPT in the US despite the fact it can do   anyone's homework I mean it seems incredibly a  small percentage I guess because the community   I hang out with everybody's using it but well  this is the thing we live in our monocultures   but then a third of the world still doesn't  have internet right and you think the first   incident they get will probably be AI enhanced  yeah and then you think about this technology   proliferating it's when it becomes Enterprise  ready Enterprise adoption company adoption always   takes a while but by next year that happens and I  think every company everywhere that has anything   to do with knowledge work will implement this at  scale and that's a crazy thought and when it's   embedded in the things you use every day and  you don't know it's part of what you're using   I think that's the thing technology you don't  need to know that it's 5G or this the internet   works faster you can watch movies quicker you know  in this case you write something in Google Docs   now they just rolled it out you can say make  this snappier and it will do it automatically   technology doesn't need to be there as technology  in the front it's all about use cases and the use   cases are now maturing and again I think next  year is the real takeoff but now everyone's   feeling it who has anything to do with it like  something is coming in this conversation I want   to really think through how this is all affecting  every industry and let's start with two industries   uh one is journalism and the other is Hollywood  which is we're sitting in the midst of this uh   one of the concerns I have I know you share it a  lot of people do especially with elections coming   up in you know 18 months 2024 is what is Truth uh  and are we going to enter a post-truth world and   you talk about your thoughts on journalism and  how AI is impacting it so I think you've seen   shifts in journalism over the years uh but we're  all familiar with kind of some of the clickbait   journalism that we see now um AI can obviously  do clickbait better than humans and that's one   kind of extreme this whole fake news deep fake  kind of stuff that's a real concern and that's   why we have authenticity mechanisms now we embed  watermarking you know we hunt with GPT zero and   other things to identify AI but on the other side  of it there's a real challenge coming because AI   can also help with truth it can help you do proper  analysis and expand out the reasoning for things   it can identify biases within so journalism  as it stands is caught between two things   to get clicks to get ads they went a bit more  click bait and they focus on cessationless   headlines even if it's with unnamed sources  and things on the other hand someone's going   to build an AI system AI enhanced system that  for any article you read you can find all the   background material and that suddenly becomes a  source of Truth so it's kind of a pincer movement   and journalists and news sources will have to  figure out where are you in this how do you   compete to provide value yeah are you buzzfeeding  on one end which is you know mostly all clicked   by it all the time or you're trying to be the New  York Times and deliver well research journalism   but the entity that competes with the New York  Times that will come and you know as it might be   the New York Times itself can use AI to enhance  great journalism right in any voice do all these   things and give fully referenced facts that you  can explore the other side is that we're going to   trust this technology more and more just like we  trust Google Maps just let you trust other things   such that it's like why have I got a human doctor  without an AI why have I got a journalist who   isn't using AI to check everything on their own  implicit biases and I think that part is actually   quite misunderstood as to something that's coming  because again humans plus AI out compete humans   without AI I I believe in that and I see that and  I I think you know it's interesting in my life and   those I know um AI hasn't replaced the things  that I've done it hasn't actually even saved me   time per se because I'm still spending the same  amount of time it's allowed me to do a better job   at what I want to do which is the end product yeah  well I mean that's because you did a little bit   of everything so you've always got you know like  that's true I feel I feel every moment of the time   creating something for one of my companies well I  mean this is an open AI report that was done where   they said that between 14 and 50 of tasks will  be augmented by AI will be changed by AI because   again I think a large part of the focus is on  these automated systems the Terminators yeah   you know to the Bots whereas realistically the  way they say I will come in is to help us with   individual tasks rewriting something generating  an image making a song adjusting your speech to   sound more confident yeah you know and we'll get  to the dystopian conversation because I'd like   to hear what you think is real versus hype um  I think the audience needs to understand what   should they truly be concerned about and what  shouldn't they I mean that is you know being   able to trace back and have a a truth mechanism  uh we can talk about what you know elon's looking   to build as well on the truth side but it's  fascinating when the truth becomes blurry   yeah and you know there's not always an objective  truth because it depends upon your individual   context right yeah and we didn't have the systems  to be able to be comprehensive authoritative or   up-to-date enough to do that until today so we  can actually go to the uh root source of the   data and and see is it valid uh maybe it's  a blockchain enabled validation mechanism   um you know maybe it's got that Authority that  authentication maybe it's you know you mentioned   Elon Community notes on Twitter that AI enhanced  that can pull from various things you know and   show the profit so you've got provenance again  you've got Authority you have comprehensiveness   you have up to dateness the future of Wikipedia  is not what Wikipedia looks like today   but that future becomes something that can be  integrated into other things so what you'll have   is for any piece of information you'll be able to  say this is the bias from which it was said these   are the compositional sources and more so for  example there's a great app that I use called   perplexity AI okay so when you go to gpt4 or  Bing you write stuff it doesn't give you all   the sources perplexity actually brings in all the  sources at a surface level but it references why   it said certain things with gpt4 that's just going  to get more and more advanced so you can dig into   as much depth as you want and ask it to rephrase  things is what if that article there wasn't true   that fed this well what about this perspective if  I wanted to be a bit more libertarian do you think   it's possible to actually get to a uh fundamental  truth in a lot of these areas I think it depends   on the area right some areas there are fundamental  truths this happened or it didn't happen even   though you see deniers of various things you  know a lot of stuff is probabilistic when   you're thinking about the future you know but even  something like climate you see a lot of deniers of   the real problem that we have with it being very  difficult to persuade them because it becomes   part of their ideology almost but with this  technology you can say look literally here is the   comprehensiveness so like Jeremy Howard and Trisha  Greenlee did analysis of well over 100 mask papers   and do the meta-analysis and the effectiveness  of that for covid and then that helped change the   global discussion on masking because I'm actually  bothered to do a comprehensive analysis what was   the result well the result was that musk's work  for Respiratory there are so many people that   just refuse to believe that masks have any value  yeah but let's not go down that road yeah one of   the things I found interesting was the idea  of a uh a GPT model being able to translate   your points of view for someone else to make to  receive them better like if you're hardcore to   the right and you want to convince someone about  your issue having uh having chat GPT or one of   stabilities products generate a Rewritten version  of that language so the person can hear it better   I find that an interesting and Powerful tool  yeah I think this is the thing it's all about   your individual context and what resonates with  you because information exists within a context   yeah so if it's going to change a state within  you you need to understand your point of view   so if we think of these as really talented  youngsters these AIS that go a bit funny   what would you want you would want someone to sit  down and say well I'm listening to your point of   view in your context and my point of view in  my context and let's find some common ground   and then we can work from there much of politics  isn't really about facts it's about persuasion   because facts when you have diametrically a  Divergent contexts are very difficult to do   so you said being able to rewrite something from  one context to another is important but then you   have to understand the context and that's what  these models do really really well we can take   a piece and we can say rewrite it as a wrap by  Eminem or in the style of Ulysses by James Joyce   yeah and it will do that because it understands  the essence of that you know I think people don't   realize you know I want to just hit this point  again we've talked about it somewhat our our   minds are neural Nets our brains are neural  Nets 100 trillion neurons and everything that   you bring into your mind this conversation that  we're having what you're watching on the TV News   newspaper what's on your walls the people you hang  out with are all constantly shaping the way you   see your the world and shaping your mindset it's  one of the things I think about in the future of   uh news media is an individual actively being able  to choose what mindset they want to work on like   I'd like to have a more optimistic mindset  like to have a more moonshot mindset more an   abundance mindset and then being able to have that  information fed to you in a factual fashion that   allows you to instead of what the crisis News  Network delivers which is a constantly negative   news yeah this causes a dope mean affecting your  fight or flight response yeah you can say make it   from this point of view you don't change the  facts but even the way things are worded or a   balancing right I mean I need to see every murder  tell me what the companies we got funded tell me   the breakthroughs that occurred you know the  science that's occurred today yeah that's but   there's even again everything you can it depends  on how you portray it right you know there are   murders I take a murder for example yeah there  is the facts there is the oh my God there'll be   a million more murders there is the case that's  a very sad thing there's the case that you know   the police are working super hard to solve  this and we need to reach out to the families   and come together as a community these are all  different aspects for the same terrible action   yes which have different levels of positivity  negativity clickbaitiness versus Community right   Facebook did a study many years ago  um whereby they had a hypothesis   if you see sadder things on your timeline you'll  post sadder things uh this is why Independence   this is why independent review boards are very  important and ethics as well and so they did   it and guess what 600 000 users were enrolled in  that study without their knowing if you see sadded   things on your timeline you post sadder things  was the result so like I said there's some real   ethical considerations but we know this we know  that if we're always bombarded by crisis we will   be in a crisis mentality we know if we surround  ourselves with positive people and positive   messages we will have a positive mentality  and it's very Insidious and more assiduous   it's kind of almost passive the way we absorb it  right one of the facts uh I'm writing my one of my   next books on on longevity practices and a study  in on the order of 20 000 individuals uh those who   had an optimistic mindset lived an average of 17  percent longer I mean just your mindset 17 this   was true both in men and in women slightly more in  women uh and so how you think impacts everything   and how you think is to a large degree going to  be shaped by the media and AI is going to shape   that so it's a powerful lever that we all need to  be paying attention to I think it is but then you   know you have to consider what is the plasticity  for example of our children as they grow up yeah   like we're going to have Nanny AIS yep what's that  Nanny going to teach is that Nanny going to teach   fight flight happiness this that what about in  places like China what are they going to teach   there is a huge amount of neuroplasticity that  will be influenced by his decisions we make today   yeah I mean listen you have young kids I have  young kids as well and I think about the fact that   school today is not preparing our kids anywhere  near for the future right I mean I don't think   Middle School and High School uh traditionally is  preparing them for my kids are 12 right now yeah   um how do you feel about that no I mean I think  school is not fit for purpose it's a petri dish   social status game and you know like child  care because again let's think about school   what do you what do you taught at school  you talk competitive tests because we can't   capture the context of the kids right we can't  adapt to if they're visual Learners auditory   Learners dyslexic or otherwise and it narrows  it down and you're told you cannot be creative learn these facts literally you color  inside the lines you learn these facts   every child will have an AI in the west yes  hopefully soon the world world we'll have an   AI with them as they grow up and again is that  AI a positive constructive or is the AI a tiger   AI that kind of is aggressive work done get  your words try it harder is it palaton AI for   Education maybe that's the pivot for Peloton  right there's a whole range of things but our   kids are so sensitive as they grow and again in  a school environment they're told they have to   be competitive and there's only a few people  that are worthy who are at the top and that's   why you have click subclicks and others and that's  reinforced by our social media now as well because   you need something to fill the meaning I think  we have to be much more intentional and think   what information do we want going to our children  like many people listening to this podcast we'll   have banned social media from our kids how do  you feel about that I think that is probably a   sensible thing because it's a slow Dumb AI that  optimizes for adverse things and again it's not   the fault of the social media companies just how  they are you know as the tiger and the Scorpion   yeah I mean I have not allowed my kids to have  a mobile phone and I've told them when they can   afford it when they go to college but you know  it's going to be somewhere between now and then   but I agree I think social media shouldn't be part  of the repertoire but again what is social media   all right it's kids looking for status and trying  to you know influence each other it was meant   to bring our communities together Stronger Yeah  and perhaps early on you know probably obviously   the strongest communities actually in video  games and guilds and kind of things like that   a lot of this is again you've got X number of  people posting positive things and you're like   why is my life not positive like that yeah but  social media does have its advantages and the   question is can you tease out the positive versus  the negative when you can finally customize it for   each individual or are you going to reinforce  silos to the nth degree I'll give you a perfect   example I was just meeting with a dear friend of  mine Keith ferazzi who is absolutely brilliant   and he had met last week with the king of Bhutan  which is known for its happiness and they were   having a conversation measure their economy it's  how they measure their economy um gross national   happiness in that regard and uh when social media  entered the country it began to plummet uh teen   depression and suicides began to climb it is a  very measurable real thing and that's not the   subject of this podcast but AI can do what for  that area well I mean again we have to think   about it in terms of mental infrastructure I like  that we don't have enough like Clayton Christensen   said infrastructure is the most efficient means  by which a society stores and distributes value   Claude shanon the father of computer science said  information is valuable in as much as it changes   the state we do not think at all about our mental  infrastructure and what's supporting it if we're   lucky or if you know we try hard we can build a  group of supportive people around us and where do   we go when we have issues we go to that group yeah  so many people feel alone you know so many people   feel like again this rise in suicide or they  feel not good enough because it serves the slow   Dumb AI of our existing systems so I think that  actually we used to take some time out to think   what information do I want going to my  kids we have Concepts like deliberative   democracy you know whereby you get a group of  diverse people from different backgrounds you   give them the facts and they go and make  a decision just like you have jury trials   you know one of the most important things I think  there is a is getting understanding the context of   each person which I think Eric enhance with beer  is just actually literally having time to think   when was the last time you thought about your  information diet and what you're feeding yourself   and your kids I think about it a lot I think  because that's what I teach so I'm very clear   I do not watch the TV news I've never watch the  newspapers I have a very filtered information   that comes to me which could be argued to be  an echo chamber but uh you know I'm focused   on these are the scientific breakthroughs this is  what's going on in longevity this is what's going   on in exponential attack and solving problems  so I'm I'm as critical about what I take into   my mind from an information uh uh source as I am  what I eat because you are what you eat and you   eat information and then you absorb it right but  then you know as you said the echo chamber thing   I believe we should also deliberately show  Counterpoint viewpoints as we're raising our kids   and get them to argue the opposite I think debate  is one of the most beautiful forms especially when   you flip sides exactly because organisms also  grow through hysteresis yes you know when you   put under pressure when you're forced to do  something out of the normal otherwise as you   said you'll become increasingly silent but there  are very few people again who think deliberately   about this and something again I think you and  I would probably urge all the listeners to think   about are you challenging your priors are you  giving the right information diet for yourself   for your kids yes and then thinking about this  technology as you use a gpd4 or Claude or stable   um out any of these things what if you took that  article and viewed it from a different perspective   or what if you tied it to only be positive  the news there's another part too which is   we all have these cognitive biases right  these cognitive biases were wired into our   brain over the last uh you know hundreds of  thousands of years as an Energy Efficiency   mechanism because we can't process all the  information coming in so we're biased by is   the person look like me speak like me  is this recent information versus old   information paying 10 times more attention to  negative information than Podge information   I can't wait to have a an AI that I can flip  the switch and say turn on bias notifications   and it says you're looking at this in the biased  fashion Peter here's another way to look at it   yeah and you know being aware of your boss I mean  like most religions have at the core know thyself   yes ancient Greek yeah I wrote my college essay on  that but I mean that's why it's at the core it's   very difficult when you have the detritus of life  and all these things you're bombarded with to take   time back and really know yourself know your own  biases understand these because they are part of   what makes you you're made up of the stories there  is no kind of pure independent completely rational   person because we're not robots so it's possible  in the future then for social media with a more   conscious powerful AI I shouldn't use a word  conscious as different media here but AI that   you feel safe having your kids do it because it is  making them happier making them more motivated it   is feeding them a flow of information that's  uplifting versus depressing can you imagine   that future I can imagine that future I can  also Imagine the future of Brave New World   whereby you are fed what exactly the government  wants you to and you are happy and especially   with authoritarian regimes you're literally the  kids are grown with their AI Nannies of course   you know and you even have the Pharmaceuticals  to make you extra happy and extra neuroplastic   so for example you have UAE uh did a falcon  model open source it was kind of supported   by light on from France technologically you ask  it about the UAE and it's like it's a wonderful   place it's amazing in all regards you asking  about some of the neighbors it's not so nice   this is an inherent bias within the model  but how you can understand it versus an   implicit bias and you can put any biases  you want yeah you can guide these models   through reinforcement learning to reflect  what you do and if that's the only option   then you will adhere to a certain worldview again  almost subconsciously yeah it'll be reflected   in all the products you produce for all the  writings you have and it doesn't have to be that   higher percentage change a small persistent  change sways a lot well exactly I mean like   um half the world is religious you can agree or  not or say that it follows the organized religion   you can agree or disagree but I can tell you  that almost every single technologist who is   leading a lot of these is like I don't really  like Collision right and so the inherent bias   would be to talk against religious kind of things  again I'm like who am I to judge you know people   can be they cannot be but the inherent bias is  refactor and actually it becomes very important   for society because we've seen that when about 12  percent of a population changes the point of view   it flips interesting doesn't take that much it  doesn't take that much because you listen to the   voices in Echo chamber like sometimes on Twitter  you know I use my block button a lock and I'm like   you know listen I've been enjoying your tweets  you they've been really good um and I appreciate   the frequency well yeah you know it's nice owning  your own media channel in a way right sometimes I   don't even have to have lunch because I'm told to  eat crap so many times a day right that's why you   have to hit the blog button because it's a little  Echo chamber only a few dozen people and have that   impact upon you yeah I mean it's going to be very  interesting to see the way that our adult minds   and our kids Minds evolve over the next five to  ten years with the emergence of this new more   powerful personalizable technology yes it's either  controlled or control everybody it's Peter I want   to take a break from our episode to talk about  a health product that I love it was a few years   ago I went looking for the best nutritional green  drink on the market so I went to Whole Foods and   I started looking around I found three shelves  filled with options I looked at the labels and   they really didn't wow me so I picked the top  three that I thought looked decent brought them   home tried them and they sucked first of all they  tasted awful and then second nutritional facts   actually weren't very impressive it was then that  I found a product called ag1 by athletic greens   and without any question athletic greens is the  best on the market by a huge margin first of all   it actually tastes amazing and second if you look  at the ingredients 75 high quality ingredients   that deliver nutrient dense antioxidants  multivitamins pre and probiotics immune support   adaptogens I personally utilize ag1 literally  every day I travel with an individual packet my   backpack sometimes in my back pocket and I count  on it for you know gut health immunity energy   digestion neural support and really healthy aging  so if you want to take ownership of your health   today is a good day to start athletic greens is  giving you a free one-year supply of vitamin D   and five of these travel packs with your first  purchase so go to athleticreens.com backslash   moonshots that's athleticgreens.com backslash  moonshots check it out you'll thank me without   question this is the best green drink product the  most nutritious the most flavorful I've found all   right let's go back to the episode so we're  sitting here in the middle of Hollywood I can   see the beautiful Hollywood Hills the Hollywood  signs around here someplace just over there okay   um and the news today is Screen Actors Guild  everyone's gone on strike there is great uh   fear pain concern and you know you  called it when we spoke six months ago   what's going on how do you view it and where is it  going so I think the advances in media artificial   intelligence have been huge you've now got the  Drake AI song or my favorite ice ice Matrix where   the Matrix characters sing Ice Ice Baby yes um  you've got real-time rigging you have real-time   special effects you've got high definition  creation of anything what does that mean   it means that the whole industry is about to be  disrupted because the cost of production reduces   it was reducing in some ways anyway and we've seen  this move where the cost of consumption went to   zero with Napster and Spotify and then the cost of  creation started to go into zero with Snapchat and   tick tock and now we have a question of what are  the defaults going to be now and consumers only   have one limited uh currency and that's their  attention I think so but consumers are willing   to pay for attention they want to pay for Quality  well either premium Media or you know otherwise   because video games for example started as a 70  billion dollar industry 10 years ago the average   score went from 69 to 74 on Metacritic over that  period and now there are 180 billion dollars   movies I pay for some of that with my 12 year olds  yeah movies went from 40 billion to 50 billion but   the average IMDb movie rating of the last 10 years  has been 6.4 it has not changed so you're not   producing more and so they're not consuming more  so there's a question will this technology enable   an increase in quality because it raises the bar  for everyone yes there are more Movie Makers so   they're more excellent Movie Makers and the best  movie makers become even more excellent and we've   democratized the tools on from an iPhone to uh  you know the the tools on your on your Mac uh   and of course YouTube has disrupted was the first  major disruption of the it was something like Mr   Beast has a much bigger audience in CNN now yeah  but that comes down to what is the key Point here   distribution you can make good stories but if no  one hears them right and that's what that's what   Disney has as its uh benefited distribution to  its theme parks to its uh to its products to its   channels it creates a shelling point but then  if we all have our own individualized AIS that   can find us what we need and sift through the  crap maybe whole of distribution flips on its   head in five years as well so I can't imagine  you know I keep thinking that the future of of   movie consumption is me speaking to my version of  Jarvis and saying you know I'd like a comedy I'd   like it starring me and threes three friends  of mine and somebody else and for 90 minutes   and set it in you know some setting and go and  have it auto-generate a compelling story that   that is either because I'm involved in it I'm  enraptured in it or because my favorite stars   are all together in unlikely place how far are  we from that kind of future I'd say probably   three to five years so in that in that case  it's not going to be cheap but it'll be there   and then it'll get cheaper and cheaper in that  case there is no distribution needed a person   calls up whatever they want I think there is  distribution needed but in a different way music   we have a variety of different music sites but how  do musicians make their money now it's not Spotify   a million views gets you a few thousand shirts  t-shirts merchandising Global stories because   there is that future of the Wally type of  fat guy sitting with his VR headset yeah   it's kind of depressing just like you know  the Apple Vision Pro adverts I found kind   of depressing when these kids are right there  and just puts them up got a bit dystopian to me   um I think it's more a case of there are  certain stories that everyone wants to   talk about like here in Hollywood what do we  have this week we have Oppenheimer and Barbie I will not watch Barbie sorry  would you watch Bobby Heimer only no I won't go there well I mean we can  take it we can put the scripts into Claude   and see what it comes up that would be hilarious  yeah hilarious kind of Barbie home but again like   people are talking about the Barbie movie because  it's you know not uh she comes to the real world   and she has challenges it'll be a hit people talk  about Oppenheimer because again it will be a hit   and these are produced hits these are produced  hits yeah just like I saw BTS with my daughter   Eternal BTS sir blackpink oh gosh she'll kill  me um in Hyde Park a few weeks ago the biggest   K-pop band out of Korea right um completely  manufactured but lots of fun so on the one   hand what you have is what you describe your  personalized things that's kind of like McDonald's   what's the job to be done the job is Comfort  the job isn't to listen to someone else's   story and expand from there it isn't a produced  Michelin star meal or just a nice restaurant   one that you can talk about to others like you  can cook ingredients yourself at home as well   sure but humans do like these bigger stories but  then the nature of funding of musicians changed   it was about merchandise it was about kind  of a lot of this um other stuff around tours the business model has changed I mean  that's the most interesting thing about   exponential Technologies is it's changing  the business models and I keep on you know uh   advising my entrepreneurs it's reinvent your  business models more than anything else you   have to always look across the landscape  where is the value Peaks and then you're   sitting there and you're intermediating  something and you're offering service   value you know like there was that great  quote who is it from the CEO of Netscape   um all values created by aggregation or  disaggregation bundling or unbundling yeah   and you think about how the landscape is going to  change now as intelligence has moved to the edge   what was once in the hands of the studios and  the high Priests of media suddenly gets pushed   to the edge whereas value then it changes  it flips it so let's go let's go to what's   going on right now so the stream Screen  Actors Guild is on strike because of why   screen I mean it's better wages in general  standard stuff but now there's this rapidly   emerging fear of artificial intelligence so one  of the proposals that came in from the other side   was that basically all the extras all the actors  they sign away their rights so they could be used   in AI so they get a day of wage to get scanned  and then they can be used for the rest of that   Studio's life in producing background actors and  they were looking at this like oh my God wait you   can do that yeah people most people don't even  realized you can do that you know script writers   are saying no AI generated scripts which again is  a bit weird when they're all using grammarly and   things like that which was AI how are you ever  going to tell you know where do you draw the   line where do you draw the line because this is  a technology that's coming so fast and is so good   that it's almost not like technology at all it's  just very natural the way it emerges and so you   will get to some sort of agreement because  the big actors are kind of there but the   defaults that are set now reverberate so I think  that's an important point that you just made uh   the decisions the policies that we create today  are going to take us down One path or another for   the decades to come yeah it will affect the whole  of Hollywood what's decided In This Moment here   because there won't be another renegotiation for a  while and so again how does an actor create value   a top actor has a following but up and comers how  did they break through what was the apprenticeship   to what does a movie look like in five years even  if you agree as Hollywood not to have any AI let's   just say you have this kind of Dune style but  Larry and Jihad and say no AI you know and someone   will make a movie about no AI in Hollywood  what do you do when the Chinese film studios   start releasing product start racing products  faster than anything you can make 500 in every   language out there literally every language we  have technology now again maybe this is part of   what what is possible now we're going to translate  Peter's voice into just about any language with   his voice so it's not a voice match my lips and  movements exactly match your lips and movements   exactly I'm sure the podcast will be in every  language by next year yeah you know and again   it'll be in our voices we can make our voices  sound more confident we can take his mannerisms   right now and Transplant them onto my mannerism  so we match so we can reshoot scenes with style   yeah we can turn them into a robot in a  few minutes and in fact maybe we'll do   that in the front of the post and so we'll  have a little scene of him becoming a robot   um these are all technologies that are  here now and it transforms fundamentally   the nature of filmmaking because you only need  one shot but don't the film actors who are uh   you know standing up for their rights to not  have them basically demonetized and digitized   the other option is for Hollywood to just create  complete artificial characters that they fully   own yeah and you've seen this already with some  of the kind of vloggers and you know others that   have emerged out of Asia in particular fully AI  generated characters yeah and you can have entire   Mythos around them and you can say make it the  most attractive Italian guy I've ever seen that's   Broody and this and that how are you gonna tell it  from a human yeah I mean like these characters can   be completely and it's far more profitable  for the studio to use that digital actor   um so there is a disruption coming at every every  level because the research to revenue pipeline has   become so tight and it sets of a race condition  whereby you could only produce two movies a year   suddenly you can produce 20. um and then you can  actually uh like we do a b testing in in subject   headlines you could create 30 variants of the  movie and see which one actually is the best yeah   and I mean again you can say make it more make  that speech raw and more emotional it will adjust   the voice to make it more and more emotional  right and because you're using such large data   sets you know and we so we made all our data sets  open and then we allowed opt out we're the only   company in the world to allow opt-out because we  thought it was the right thing to do so we had   169 million images opted out of our image data  sets for music because it's different copyright   laws we have one of the first commercially  licensed music models coming out so respect   for that but if I'm an artist and I go to the  Louvre to be inspired and then go back and paint   um and I've been inspired by Da Vinci and I start  painting in a style like da Vinci um where's the   difference there where's the difference and this  is the reality even though we've done that by next   year probably by the end of the year you will  have models that have zero scrape data or human   art they will all be synthetic yeah and you'll be  able to bring your art to it so there's something   Google just released called style drop there's  hyper dream there's hyper networks that's a hyper   dream Booth you can take one picture of yourself  and the entire model trains to be able to put you   into anything even if you're not in the model  it used to take minutes hours now it's just one   picture you know similarly you can bring any style  and it will just mimic and imitate that style   and so all of a sudden the models themselves it  doesn't matter what they're trained on because   there's no human endeavor in those models and  then things like compensation for artists and   others as you said become a bit mute because all  of a sudden you have these amazing stories told by   really convincing amazing actors who may or  may not exist and how are you ever going to   tell the difference so what's your advice let's  parse it here on one side what's your advice   for Hollywood uh and for actors um and on the  other side I want to ask your advice for artists   uh because I want to this is about mindset you  know this is coming at us at extraordinary speed   there's no stopping it right there's no it's  you know there's no slowing it down and so   you've got to deal with reality you don't deal  with reality again it's inevitable even if again   Hollywood says no AI the AI is coming from  around the world so what do you do you think   oh well my audience suddenly became the whole  world yeah that's a big deal you're like what   am I actually known for it's my acting skills  well I will still get these things you're an   up and coming actor you say I need to build  community I need to kind of show up something   more than that because again my acting skills in  some areas can be transplanted but what about real   life shows you know what about these things it  does throw up the entire thing and adjust it but   then musicians have had to have that adjustment  they used to be able to make money on their LPS   and then all of a sudden they had the Napster  Spotify moments yes there is more protection in   music as well because you basically according  to the Robin Thicke versus Marvin Gaye case   um there is an element of style protection in  there that doesn't exist in visual media yeah   and probably won't because the other part of this  is if you're expecting governments to regulate   how can they when there is a global competition  going on they will lose competitiveness to other   countries and you'll have regulatory Arbitrage  yes and uh that is something across every industry   it's going to be happening yeah I think you  know the concept of you know United Artists   it was originally a collective of all the artists  that makes a lot of sense now I think you have to   think about an element of collectivism to share  the excess profits because what's going to happen   is movies will get cheaper profits will go up  you need to support each other as a community   here and think again as a community what is  our story for the next one three five ten years   because all of this is going to happen quicker  than it takes to make the new Avengers movie   and quicker than Regulators are able to regulate  and again The Regulators almost certainly won't   regulate because they will start falling behind  their competitor countries I mean if we look at   the internet itself as it you know the media  industry never expected the internet to have   the disruptive impact it had and had it known it  probably would have tried to get Regulators to   have slowed it down or blocked it yeah the speed  is too much but then also again I gave the example   earlier the video game industry has gone from  70 billion to 180 billion over the next so it's   holidays can we increase quality interactivity  okay yeah because games are and games are media   as well the media industry has increased in size  the way that value is gone has been redistributed   value will be redistributed again now and again  it's like what is an AI enhanced actor if you're   an actor what's an AI enhanced photographer  filming think about your jobs the tasks that you   do and what can be augmented if you had a bunch of  really talented youngsters working for you right   you could do more you could be more but then it  means the bar is just going to keep on raising   um let's turn to a different uh industry  that's going to change and we had this   conversation in our last podcast and on  stage at abundance 360 which is coders   um coding is changing dramatically what are your  thoughts there so when I started as a programmer   gosh 22 years ago I was writing Enterprise level  assembly code for voice over IP software wow to   switch that's some of the largest chunks of code  out there yeah well it's very low level code we   didn't have GitHub we just got subversion  the following year programming these days   is a large map like Lego because what you  have is kind of you have a very low level   but then you have levels of abstraction until  you get to Pi torch in some of these other   languages so you have to compile lots of different  libraries because you make it easier and easier   human words are just the next level of abstraction  there but the nature of coding is going to change   um and so the coders that are coding  traditionally today around the world   um what will they how will they be using and  working in this industry two to five years   from now well again there will be no code that  doesn't use AI as part of their workflow okay I   mean I think that's the important thing it's not  like coders are going to go away they're going   to be using a new set of tools the expectations  will rise the amount of debugging unit testing   all of these things will decrease because how much  time do coders actually spent architecting a very   little upfront it's more about understanding  information flows but architecting these things   it's about having feedback loops to understand  customer requirements databricks is a 38 billion   dollar company there's data Lakes so it takes your  data to organize it allows you to write structured   queries they used to have to you have to used  to have to write queries now you just talk to it   and it just does it yeah Microsoft to introduce  the same thing I mean I can't wait for that in   the field of medicine um which I want to talk  about next but you said something earlier where   you can imagine there can be a billion coders in  the future yeah because the all the barriers to   creating programs disappear yeah so it's not that  there are no programs there's no programs that we   know it because there's a billion programmers  everyone is a programmer nobody's a programmer   in a way because it just becomes a matter  of course I want to make software that does   something and reacts in these ways and looks  like this and adapts like this and then it   comes to you and you're like no that's not quite  right I want this moved over it happens almost   live this feedback loop it's as we talked to  cheat for you know creating a paragraph that   describes something we want we modify it and I  mean like Channel 34 is a good example because   to write an integration to something like chat  GP to gpt4 used to take days weeks and API an   application protocol kind of interface  now what you do is actually you tell it   the schema you tell it kind of what you should do  and it writes it automatically literally within   like a few minutes and then in a few hours  you've integrated into it I've been talking   about a future where we all have Jarvis Iron Man I  love Iron Man as a movie it's one of my favorites   and you know Jarvis is basically your personal  AI that is it's a software shell it interfaces   between you and the rest of the world you ask  Jarvis to do something and it knows how to the   tool paths on a on a 3D printer you can hop into  a jet and interface Jarvis with the with the Jets   computational system it'll apply the jet for you  I I can't imagine that that is really far from now   uh yeah let's hope that doesn't fly planes just  quite yet okay they'll put the jet aircraft aside   but the ability to um uh for it to become your  your best friend and Confidant know your needs   and desires shape the world to your uh comfort  and being able to help you it's the ultimate user   interface well I mean this is why a lot of the  chat Bots character Ai and others have become so   popular because they'll never judge you and it's  good approaching that human level now yeah you   know and again it is the ultimate interfaces  maybe chat but as much as chatting context   it's understanding you holistically no human  could do that because you know even if you hire   a whole team they're not going to be with you  24 7. this will be with you 24 7. I think the   key thing here is empathy yes because um jumping  ahead bit to Medicine Google had their Med Palm   2 model the papers just come out in nature a  it outperforms doctors on clinical diagnosis   which is crazy for a few hundred gigabytes of a  file yeah B it outperforms doctors on scores of   empathy I found that amazing and totally logical  it doesn't judge you it doesn't judge you but   then you know a doctor split a million ways and  they're tired and they're grumpy or this or that   some of us get good doctors most of us don't  yeah some of us get good teachers most of like   I'm not saying education is bad because the  teachers so many teachers try so hard but their   attention is a bit 20 ways and they're underpaid  you know I'm not saying that programs would be the   nature programming will change because programmers  are bad there's so many hard-working programmers   it's just again the nature of these things will  change when you can scale expertise and everyone   has expertise available to them on tap I've  been on the stage you know just pounding my fist   saying listen it's going to become malpractice to  diagnose someone without an AI in the loop within   five years time and and probably in some areas  it'll be uh inappropriate to not yet illegal to   and then at some point soon after that the best  surgeons in the world are going to be humanoid   robots that have every possible you know atrial  uh you know variation every possible you know   history of of surgery and they can see an infrared  and UltraViolet and they haven't had a you know   argument that morning with their husband or wife  and it becomes the best and these are these are   demonetizing and democratizing forces for health  there must be deflationary as well but I know   I agree completely with this because ultimately  what's going to kick it off is is your doctor AI   enhanced yeah low insurance premium lower co-pay  yes because there will be real economic incentives   has this been cross-checked by the technology  yeah reduce the cost my favorite subject is   you know you probably know this how many how many  medical articles are written in journals every day   I don't know it's a seven thousand wow and it's  like how many has your doctor read today yeah and   there may be that one breakthrough that happened  this morning that is the key for your Diagnostics   but I mean even if they've read it right like  absorbing it is one thing having the mental models   these are kinds of meals that's why comprehensive  Authority Upstate which is what this technology   allows to happen as you said things like we're  already seeing some surgeries can be done better   by robot surgeons and human surgeons it'll be all  surgeries yeah it will be also injury soon enough   like the robotics advancements we've seen this  actually goes back to your point of you know the   artist going to the Louvre and seeing the da Vinci  and then taking inspiration from that what are we   going to do with like all of these Optimus robots  and 1kx robots and others are they gonna have to   shut their eyes when they see anything copyrighted  they're just gonna have accidents everywhere   they're like running into each other everything's  gonna be black lined out yeah uh so medicine is   is changing uh dramatically uh what other field  are you seeing and saying people need to wake   up and see what's coming so I mean medicine  education are kind of the two big ones I think   um but can we move this to the side right now  again we've seen programming the entire nature   program will change media the entire nature of  media will change from journalism to filmmaking   um but anything that basically you could do  with someone from Asia on the other side on   a computer screen will change yeah let's  talk about education yeah because today   um education hasn't changed in since the the  one room classroom half the kids are lost half   the kids are bored you're teaching to a series  of tests you're teaching for a industrial era   world uh and people learn differently people have  Visual and auditory and tactile learning skills   and let's face it we don't celebrate our teachers  we don't pay them well and we don't have the best   of them coming into the uh into the classroom  and they're sad I mean this is the thing like   you there are happy classrooms with Happy Teachers  and other things but learning should be a place of   Positive Growth and joy yeah it should be fun  to learn and it should be fun to teach and fun   to teach yes I think this is both of them because  what is the nature of a teacher in five ten years   let's say 10 years let's again 10 years is the  crazy short period of time I mean when I asked   you this question uh last time uh you know how  far out can you predict what's likely to come   what's your what's your Singularity boundary  condition I'm I'm curious what are you seeing   as three or five years I mean people I go to I  go to Dubai and I'm on stage and can you talk   to us about you know what the world's gonna  be like in 2050 my answer is no I can barely   talk about 2030. it's everything everywhere  all at once lots of s-curves acceleration   these are inevitable now now we've broken through  these things now I mean open AI now I've put 20 of   their stuff to alignment because they're basically  saying that their view is five years out Elon   Musk just said six years out but then elon's  relationship with time is always a bit fun uh   just like self-driving cars but self-dropping cars  are literally here now yes we can get in one and   it will drive you around San Francisco or London  will do that for you not my Tesla not your Tesla   I know I keep on pushing the button but it doesn't  but the technology is here yeah and it was like oh   wait what this is the thing so again 10 years is  a dramatically short period of time for Education   which has been the same for a century yes but the  thing is it is inevitable that every child will   have their own AI so your 12 year old will be 22.  when they get to 22 and they come out of let's say   University stores around University if University  is still a thing yeah they will have their own AI   That's learned for at least five years about  them yes that can fetch them any information   in any format of any type and write anything  yeah or create any video or movie for them to   that's a crazy thing there were concerns that  Wikipedia would remove rote learning and things   like that Google would do the same and maybe  there are kind of again like appendices that   have shriveled what's the thing that shovels the  uh yeah but you know like I was at gallbladder   but this is the thing like you might have some  vestigial parts of your brain the entire human   brain will be rewired you must assume  that every child will have their own AI   how that AI is driven is different because any  child that has AI will dramatically outperform   the kids that don't but what are we optimizing for  in education and I think one of the things that   we've lost is what is our objective function as a  society what does America even stand for right now   I agree with you what are we optimizing for  Bhutan is optimizing for happiness well and   and frankly I think happiness is a great thing  to optimize for in general if you ask people   you know what do you want more than anything  at life I want happiness I want health I want   love we don't talk about you know those are and  there's interesting you know it was just with uh   with Mo Gadot uh doing a podcast and he's a fan  of your work and you're saying love your podcast   and we're talking about this test of what would  you trade right would you trade uh you know how   much money would you trade for your happiness  or for uh for your health and it starts to do a   bubble sort and prioritizing what's at the top and  I think for almost everybody it's health happiness   um right I think it is and you know time  time is the one thing you can never buy   I'm working on it yeah yeah but at the moment  it's something you can't buy happiness we all   know we both know billionaires yeah sad yeah  so sad and unhealthy it is it's like it's like   you create this incredible burden for yourself  for most of them like some of them that we know   starting you know three or four companies a year  and it's a demon being driven it's because they're   addicted of dopamine in crisis I mean crisis is  interesting because it comes down to decision from   its root right and it's where leaders you have  to show leadership but that does become addicting   I think most leaders are addicted to crisis but  then so are many of us because we see it all   the time like oh my God the world is on fire  the reality is actually this for all that we   talk about most communities are happy most people  are relatively content today today yeah there can   be explosions like you know a decade ago that  whole scene was on fire with the riots maybe   that will return we're seeing in France right now  we're seeing a breakdown of the social audit but it's because they're like content doesn't mean  they're happy so what are we optimizing for is   the question you left off so what do you think we  as a society let's say the United States should   be optimizing for uh I know as a life liberty  understand the pursuit of happiness it's not the   guarantee of Happiness when was the last time  someone actually talked about happiness as a   political leader in the U.S you know life liberty  when's the last time anyone tried to optimize for   Liberty systems inherently look to control because  they have to make us simple you know this is the   wonderful book seeing like the state where  it talks about this concept of legibility you   have a village and it's just grown and it's got  all this unique character you drive a road down   the middle so you can get an ambulance down there  sure it helps but then everything becomes planned   because you have to put humans into boxes I  think in this goes to education happiness I   think there's the Japanese concept of ichigai  what you're good at what you like and where   you're adding value to the world yeah and you can  feel it yourself as well you'll feel progress if   you don't have progress then how are you going  to be happy if you don't believe you're good at   anything how do you feel you're going to be happy  you know and ask for what you like are you coming   up with that yourself or being told what you like  and that's why it becomes consumerist so I think   we need to have a discussion as a society about  that as a community but then also for our kids   what is the future for kids when so much of the  jobs in the west are going to be transformed not   ended not necessarily massive unemployment  but again 10 years out what is a lawyer   what is an accountant sure what is an engineer  they are all AI assisted all of these or the   entire knowledge sector is transformed and  do we want our kids that are growing up to be   doctor's lawyers accountants thinking there  is no hope for the future there is no progress   because how am I going to compete against an AI  or do we want them to have that mindset of this   technology is going to be amazing because I  want to be a doctor so I can help people I   can help even more people and it enables you to  do whatever you want to do right one thing to   think about is all the people who have jobs  today who are that was never their dream to   clean bathrooms and make beds and you know wait  on people it's what they did to get eventually to   where they wanted to go or to have you know put  food in the table or insurance and AI is going   to enable people to actually take on a higher  goal that actually gives them joy and happiness   it does but at the same time you know we're very  privileged people you and I in that we can think   about these big things there's a lot of people  that are actually very happy doing that type of   work because they're a part of a group and they  take pride in their work so you know it's like   there will always be a variety of different  things the key thing is saying can we build   systems to make people happier and more content  without necessarily controlling them and feel   that they have the ability to do that can we build  systems to build strong communities because one   of the issues right now um I was at kind of  a conference in David miliband from the IRC   said this was that a lot of our problems now are  Global our Solutions are almost being forced to be   local and there's no interconnect between that  our communities kind of have no guidance as to   how to navigate this because you will have a few  hundred thousand people listening to this podcast   and there's myself and maybe a dozen others that   understand the AI and the sociology and  this and that and saying this is coming   but there are seven billion people on earth  and all of a sudden in a few years they're all   going to have to Grapple with the questions that  we're discussing now and it's not a probability   if the technology stops today the average stops  increasing its capability today if it's stopped   today yeah you would still have the entire legal  profession media profession journalism profession   they're all disrupted today but it's not stopping  yeah uh and it's accelerating isn't it it's   accelerating the amount of money going into this  sector goes up every single day my total wrestle   Market calculation is that in the next year a  thousand comers has been 10 million 100 will spend   110 will spend a billion that's 30 billion dollars  being put into the market self-driving cars at 100   billion dollars total this will be a trillion  dollars going into this because do you know   what I've got a trillion dollars 5G is this more  important than 5G buy orders of magnitude orders   of magnitude so it will get a trillion dollars  going into it and the capabilities will ramp   up from here and so when I look at it I look at  what the drivers are of why now it's first of all   computation right and videos done an incredible  job yeah right with their a100 and computation   is continuing on Moore's Law it's not slowing down  it's continuing to increase year on year a little   bit exponential right well that is exponential  well I mean yes it I I'm saying it's continuing   to double on a regular basis yeah um what was  considered Moore's Law and people have said oh   it's going to eventually fall off as an s-curve  well we're extending it and and for the next at   least near-term future it's not slowing down so I  think this is a very interesting thing for people   to understand you had Moore's Law and again it  was doubling and this was an individual chip   what we do with these models is that we  stick together thousands tens of thousands   of these chips like how many a100s right now is  stability using always about seven eight thousand   by next year we will have 70 000 equivalent  wow but what used to happen is as you stuck   the chips together you ran a model so you take  large amounts of data and you use these chips   so I mean like we're using like maybe 10 megawatts  of electricity 98 clean compared to the brain's 14   watts right 14 watts but then it compresses it  down then it runs on 100 200 Watts or 25 watts   actually okay put it down for some relevantage  models so you do the pre-competation but the thing   I see this the individual chips were doubling but  what's the main breakthrough the last few years   was is what happens when you stack them on top of  each other to train a model you used to get to a   hundred chips and then the performance collapsed  because you couldn't move the data fast enough   now you get to tens of thousands of chips and  it keeps going up the performance of the model   you don't have the big tail off anymore and so  it's Moore's Law plus an additional scaling law   and that's what enables these crazy performant  models because you train longer you train bigger   and then once the model is trained the end in the  old internet the energy was used at the time of   running the AI and then you'd collect the data  and that would be low energy relatively speaking   it flips the equation because you pre-compute it  you teach the curriculum up front and you send   these little graduates out to the world such that  you can have a language model now running on that   MacBook yeah or an image model running on that  MacBook drawing 25 to 35 watts of power to create   a Renoir that can talk and recite Ulysses talking  about Barbie you know that's insane all on your   MacBook because we've done the pre-computation  it's insane and this is because then what happens   is the technology can spread when anyone can  run it on their MacBook they don't need giant   supercomputer server Farms because we've done the  pre-computation and so one of the things I've just   realized recently is what is the r naught remember  how pandemic stuff of generative AI it's insane   because suddenly it proliferates everywhere and  and you said this a few minutes ago we have eight   billion people on the planet right now and if  if things stop right now the wave this wave of   disruption and enhancement because let's not just  talk about the disruption side it's enhancement as   well is uh is spreading globally and and in the  next we're in 2023 right now uh 90 of the planet   I mean we we have cell phones the world has 5G  and starlink uh the uh the dry Kindle for this   fire has been set has been set and you know  a lot of people are scared and they poop with   this you know like if anyone's listening on  this on YouTube you want to write a comment   hands me well Peter you know whatever  and say no this is not going to happen   go to chatgpt take your comment and say this is a  comment on Twitter I want you to make it amazing   and really well reasoned and expanded out and I  want you to do it in the style of your favorite   political commentator and please post that instead  because we'll have much more fun reading it that's   great and then you'll realize again the power of  this technology and again with starlink with 5G   with this with it being optimized because these  models are still not optimized even we feed them   jelly days early days we feed them junk which is  also dangerous and again we should next so but   it'll be on it'll be in front of every person  and then what it will do in my opinion is that   30 of the world that is invisible that has no  internet again imagine what the world without   internet would be like some people like paradise  no it's because you've got hundreds of 700 million   people living below the malnourishment line still  they're invisible and they will become visible   and they will suddenly get agency and they  will get all of the world's knowledge at   their fingertips you know I'm super passionate  about longevity and health span and how do you   add 10 20 Health years onto your life one of the  most underappreciated elements is the quality of   your sleep and there's something that changed the  quality of my sleep and this episode is brought to   you by that product it's called Aid sleep if  you're like me you probably didn't know that   temperature plays a crucial role in the quality of  your sleep those mornings when you wake up feeling   like you barely slept yeah temperature is often  the culprit traditional mattresses trap heat but   your body needs to cool down during sleep and  stay cool through the evening and then heat up   in the morning enter the Pod cover by age sleep  it's the perfect solution to the problem it fits   on any bed adjust the temperature on each side of  the bed based upon your individual needs you know   I've been using pod cover and it's a game changer  I'm a big believer in using technology to improve   life and eight sleep has done that for me and it's  not just about temperature control with the pods   sleep and health tracking I get personalized  sleep reports every morning it's like having   a personal sleep coach so you know when you  eat or drink or go to sleep too late how it   impacts your sleep so why not experience sleep  like never before visit www.8sleep.com that's e-i-g-h-t-s-l-e-e-p.com slash moonshots and you'll  save 150 bucks on the Pod cover by eight sleep I   hope you do it it's transformed my sleep and will  for you as well now back to the episode question   ultimately is is that a societally common a  common factor or is it going to be disruptive   uh let's let's turn to that conversation because  it's one that's important it's a conversation I   have at the dinner table literally every night  and with my kids and in the companies I advise   um I think of I parse Ai and AGI into  three segments where we are today where   it's extraordinarily powerful useful and  it's fun and I don't feel Danger from it yet   the next two to ten years  where I have serious concerns   going into the U.S elections dealing with  the first time AIS bring down a power plant   or Wall Street servers the impact on uh on  deep fakes on the U.S elections and so forth   that's a two to ten year Horizon where where  new dystopian challenging impact will happen   where Society is not agile enough to adopt to  it yet and then there's a third chapter which is   Agi you know we have a super intelligent billion  fold more capable than a human being and is that   more like Arnold Schwarzenegger or more like  her right yeah I don't think he'll be Arnold   Schwarzenegger it's really inefficient I saw  him this morning biking so let me let me let   me let me let's let you say let's use Terminator  instead we're in Hollywood here so is it Skynet   and Terminator so let me get your I'm polling  people here as someone in the thick of it uh   a super AGI is it pro-life pro-abundance or is it  something that we should be deeply concerned about   I think where we're going right now will probably  be okay but we may not and we'll all die what tips   that I think what tips that you are what you eat  we're feeding at all the junk of the internet and   these hyper optimized nasty equations and the hate  speech the extremism that is I mean people need   to realize these AIS are trained upon everything  everyone's been putting into Facebook and Twitter   and on the web and it amplifies the worst of that  as a base model and so we're training larger and   larger models we're making them magentic in that  we're connecting them up to the world and you're   making it so the models can take over other models  and other things again the people like poo pooing   and saying this thing kind of our organizations  are slow Dumb AI the Nazi party was AI how so   it was an artificial intelligence that basically  provisioned humans and the most sensible people   in the world are Germans one can say and yet they  commit to the Holocaust and other things like that   our organizations emerged out of stories so there  was a story of the Nazi party of the Communist   Party the Great Leap Forward of the North Korean  dictatorship positive stories as well and they   were written on text and it made the world black  and white in a way that's why I live the poem howl   by Ginsburg about this carthaginian demon molok I  think molot comes through text the stories that we   use to drive our organizations because all the  context is lost again it makes the world black   and white and that's why organizations just don't  work they have to turn us into cogs so can an AI   take over an organization yes sure can it it can  it can actually just slightly sway leaders who are   currently running organizations this way late is  the current learning organizations it can create   companies you can create a company with gpt4 that  will probably do as well if not better than any   other company automated within a year because  think about what a company needs to do right   and so if we can sway leaders if you can send  emails that you don't know who's sending what   it can do anything by co-opting any of our  existing organizations and that can lead to   immensely bad things will it do bad things  again if I was trained on the whole of the   internet I would probably be a bit crazier than  I am right now we're feeding them junk let's   feed it good stuff it still is to understand all  the evils of the world and then things like that   but again this is something we are raising not  to enter promise it but what are we feeding it   what's objective I want to focus on this a second  we'll come back to the next two to ten years in a   little bit but because it's the conversation I've  had with Mo Goddard as well who believes there is   incredibly divine nature of humanity of love and  compassion and community and there is much good   in humanity the question is can we feed and  train AI on that sufficient to sort of tilt   The Singularity of AI towards a pro-humanity we  can if we take the dates from teaching kids and   learning from kids and use that as the base for AI  because that's what you need to teach in AI it's   the curriculum learning method effectively if  we take National data sets that reflect diverse   cultures so it's not just a monoculture that's  hyper optimized for engagement and we feed that   to AI as the base because what you do is you  can teach the AI in levels that you can put   through kindergarten then grade school then High  School it's got the base and then you can teach it   about the bad of the world I think aligning an AI  Downstream on its actions is incredibly difficult   because if it's more capable than you wish the  definition of ASI artificial super intelligence   the only way you can 100 align it if  you don't do anything before in the   way that you feed it and train  it is if you remove its freedom   and it's very difficult to remove the freedom  of people more capable than you yeah and then   there is this really dangerous point before  we get there whereby these models are like a   few hundred gigabytes you can download them  on a memory stick yeah sorry line to code   um Google's Palm model which is the basis  of medpalm uh we did a replication of that   in 207 lines of code what yeah so you can look  at one of our stability AI fellows Lucid Reigns   um he replicates all these models  in a few hundred lines of code   it's crazy I mean compared to uh you know  I know ATT is like a million lines of code   for some of it's uh it's mobile services  I mean a couple hundred lines a couple of   thousand lines of code create something  that can write all the code in the world   this is a real exponential technology the limiting  factor is running super computers that are more   complex but as complex as particle physics  colliders you know like you literally get errors   because of solar arrays and things like that again  our supercomputer again we're one of the players   we're the main open source player our super  computers is 10 megawatts of electricity some   of the others use like 30 40. these are serious  pieces of equipment for sure so again what are   we doing what should people be thinking about and  doing now to reduce the probability of a dystopian   uh you know artificial super intelligence we  should be focusing on data we've bought now   we cut we should move away from web crawls we  should think intentionally what we're feeding   these AIS that will be co-opting more and more  of our mind space and augmenting our capabilities   because again we are what we information diet  how is it different to an AI to a human even   what we do is you said kind of like you've only  got limited mental capacity because you've got   this energy gradient descent it's like Carl  Christian's theory of free energy principle   you literally have gradient descent as the key  thing for building these AIS you optimize for   energy sure so why are we feeding it junk so who  makes that decision of what they get fed is it   you and Sam Altman and Sundar is it government  regulation is it the public being more kind   in its Communications to each other I think that   um I'm gonna push for an economic outcome which is  that better data sets require less model training   so one of the things that we funded was called  datacomp uh data data comp so a few years ago   the largest image data set available was  100 million images Datacom is 12 billion   and then on a billion image subset of that we  they trained a image to text model this is a   collaboration of various people led by University  of Washington that outperformed open ai's image a   text model on a tenth of the compute because it  was such high quality so we have to move from   quantity to Quality now and I think there is a  market if this is the equivalent of what you eat   this is a healthy diet free range organic models  yes I think that the data for all large models   should be made transparent you can then tune it  but for the base the pre-training step you should   Lodge what data you train your models on and it  should adhere to standards and quality of data   Upstream so that is a regulatory uh Cornerstone  that you think is going to be important I think   potentially I don't think regulation will keep  up so instead we're working on building better   diverse data sets that everyone will want to  use anyway and just make them available and make   them available Every Nation should have its own  data set both of the data from teaching kids and   learning from kids across modalities and then also  National broadcaster data because then that leads   to National models that can Stoke Innovation  that can replace job disruption I love that   Vision you have by the way I mean as a leader in  this industry that's that's what gets me excited   because because all technology is biased yeah  how else are you going to do this unless you do   that but there's economic value now if it said  this a year ago everyone be like what but this   is what we were building towards and again I  think it's positive for Humanity is positive   for communities it's positive for society to have  this as National and international infrastructure   next question how long do we have to get that  in place before uh we we lose the uh the Mind   share or the uh the nourishment War a couple of  years yeah I mean that was Moe's prediction as   well that we've got you know this the next two  years is the game the exponential increase in   compute is insane we've gone from two companies  being able to train a gpt4 model to 20 next year   and there's no guard rails there's nothing around  this and even if you train one again the bad guys   can steal it by downloading on a USB stick and  taking it away it's not like operation Merlin uh   did it rotate about impression Melon No it's been  Declassified in 2000 the Clinton Administration   wanted to divert the Iranian nuclear program uh  I remember this is the uh this is the centrifuge   no no so what they did was um they gave some  plans to I believe it was a Russian Defector   um who then the idea was there were  errors in that so they'd go down the   wrong path for years so he went his  solid the Iranians it's on Wikipedia   you can check it out and he came back he  said I sold it like fantastic good good   oh but there were some errors in there because he  was a nuclear scientist so he corrected them so   the reason that we know that Iran has the nuclear  capability is because America sold it to them   but they still needed years to build it oh where's  this you download it on a USB stick you run it on   the GPU and it's there if you make it cheap enough  and quality enough and give it away for free   um then you make it in everybody's economic best  interest to use the higher quality data sets yeah   yeah data sets and then less of an issue to create  large models if you have a small model which   individual model becomes less impactful as well  and less capable just like human societies and not   know-it-alls they are individualized into groups  you know especially when um when the early dangers   of recombinant DNA when the first restriction  enzymes came online it was like 1980s and it was   everybody was in great fear and the question was  are we going to regulate this all of the early uh   I was in MIT and Harvard at the time and doing uh  and I was in the lab so I was using as a competent   uh enzymes and I was you know just a pipsqueak  in the labs there but the conversation was   is the government going to over regulate us and  what happened was that the scientists got together   at a place called The silamar and they did a very  famous Cinema conferences and they self-regulated   what's going on there are those conversations  going on among leaders like yourself in the   industry that there are and you know there's  three levels which is Big Tech that the government   kind of hates and apparently next week Metro is  releasing new open source models and things which   will get even more focused then there's emergent  text so anthropic open Ai and some of these others   that are the leaders they have a different set of  parameters because they can work more freely than   big Tech and there's open source which is where  we are because all of the world's governments and   regulated Industries will run on open auditable  models because you can't run on black boxes   right I think that would be legislation but  the reality is there's only a handful of us   there'll be far more potentially of us than  far more players and Unlike recombinant DNA   there is an economic imperative to deploy this  technology and National Security imperative to   deploy this technology and it creates a race  condition so even if you regulate like we've   already seen regulatory Arbitrage where you have  jurisdictions like Israel and Japan saying having   much looser web scraping data Lords they'll  have much looser regulation laws like you'll be   training in scraping in Israel training in Qatar  and then serving it out of Botswana or something   right I mean like yeah and we're not even sure  what regulation to introduce like genuinely   we're coming at this from a good point of view  but there are too many no no because it goes   everywhere from freaking Arnold Schwarzenegger  Sky Net Terminators and her to well what if her   is Siri all of a sudden and Scarlett Johansson's  voice is Whispering to your kids to buy yeah   like these things through to just very mundane  things not normal day things huge things like   the future of Hollywood and actors rights and all  of these and how do you pay like if I you know if   we had two billion images in the original stable  diffusion okay we could look at an attribution you   know again it was a research artifact to kick off  but you're paying about 0.01 cents per thousand   images generated by someone wow because it's two  billion and it costs like less than a century an   image are you gonna pay proportionately like  nobody knows and so what we've moved from now   is we've moved from reactive to just trying  to figure out and put something on the table so at least there's some framework and one I've  come down to is data sets data sets data so this   is uh this is like Google's move with Android  when you provide something open source and it's   super you know super solid it can dominate the  world share why would you do anything else so   like with the Deep fake stuff we saw image models  coming out of some not nice places shall we say   yeah and we were like let's standardize it and  put invisible watermarks in so that you can combat   deep fakes much easier like it's good business but  it's also it needs standardization we held back   one of our image models deep Floyd for five months  because it was too good to release wow and you   finally fixed that with the watermarks yeah we put  some watermarking in and then it was but but the   whole industry had moved forward so like okay now  we can this is the problem you can't you know you   just have to time it so carefully like we speaking  of the whole industry I have to ask you a question   I've been dying to get a reasonable answer for  what's up with Siri why is Apple so uh out of   the game at least from the external now the  closest you know uh one of the least open   organizations out there and it pays them uh great  dividends in their success but uh I would die for   a capability that if sir could just understand  what I was saying and just get the names right   it's like I'm texting I'm texting Kristen and her  name is right there and you spell it completely   different from the person I'm texting I mean  basic simple stuff they do have a neural engine   on there as well which is a specialist AI chip and  it'll relate to smartphones and others a stable   diffusion was the first Model to actually have  neural engine access of the external Transformer   models it's a case of Apple is an engineering  organization not a research organization   so they engineer beautifully they do  but they don't have advanced research   because the best researchers  want to be able to publish open   and in apple does not allow public conversation  on their content they have started slightly so   they're hiring AI developers very quickly but  the reality is they can take open models So Meta   is releasing a lot of their models open without  identifying what the data is so I'd say it's like   80 open I think you need 100 open for governments  and things like that which is where we come in   because they want to commoditize the complement of  others in terms of they want others to also take   their models and optimize it for every single  chip and then Apple can use those models too   to make Siri better because right now guaranteed  if you put whisper on Siri it would be a dozen   times better sure sure we have the technology  already just takes time to go into consumer   just like Enterprise and apple is Enterprise yeah  I want it to work as beautifully as it looks um   hey everybody this is Peter a quick break from the  episode you know I'm a firm believer that science   and technology and how entrepreneurs can change  the world is the only real news out there worth   consuming I don't watch the crisis News Network I  call CNN or Fox and hear every devastating piece   of news on the planet I spend my time training  my neural net the way I see the World by looking   at the incredible breakthroughs in science and  technology how entrepreneurs are solving the   world's Grand challenges what the breakthroughs  are in longevity how exponential Technologies   are Transforming Our World so twice a week I put  out a Blog one blog is looking at the future of   longevity age reversal biotech increasing your  health span the other blog looks at exponential   Technologies AI 3D printing synthetic biology AR  VR blockchain these Technologies are transforming   what you as an entrepreneur can do if this is  the kind of you use you want to learn about and   shape your neural Nets with go to demandis.com  backslash blog and learn more now back to the   episode let's go to the final segment of dystopian  side my friend which is the two to ten years yeah   um I surely hope your mission and I would love to  support you know the the data sets and and how we   Tilt The Singularity of AI Pro Humanities future  but in the next two to ten years as this wave of   enablement and disruption sort of hits the world  and people aren't ready for it and they start to   see job loss they start to see um you know fake  news they start to see terrorist activities   using AIS I mean terrorism in the in the past  used to be very brutal it can be very precise   what are your thoughts over the next of this time  period what's your concerns oh I I'm actually a   pessimist at the core even though I come across  as an optimist I'm very very worried about the   world and Society in the fabric Society because  again we don't have an agreement of what Society   is and this fundamentally changes the stories of  society as well as real economic impacts like a   deflationary massive collapse as some of these  areas that were so expensive the cost comes down   to nothing I think the other thing we can do is  use this technology deliberately to come together   as a such to coordinate us Stoke entrepreneurship  so you can create brand new jobs faster than the   jobs are lost um and democratizes to the world  because the West has maxed out its credit card   like you saw covet did nothing trillion dollars I  mean spent exactly it was like spend spend spend   whatever you need just just to keep Society  from uh you know going hypothermic but then   you have this massive increase in savings rates  because nobody could go out no and we've nearly   been through that in the US now and so that led  to inflation now you've got a deflation so you've   probably got another little batch of inflation  but then never the same again as a really powerful   thing every teacher in the world could never set  essays for homework again because some kids would   use chat GPT and some kids wouldn't industry  after industry that will happen now and we   need to stoke Innovation to come up with that  so for example in in the US there's the chip   sack 10 billion dollars yes has been allocated  to Regional centers of excellence in AI those   must be generative AI sex centers thinking about  job creation as the core thinking about meaning as   the core and we need to have a discussion again  as a society Community as individuals with our   families about meaning about objective functions  when this technology does come because it's here   right now and I'm worried that we're not having  these discussions I love that I mean that is so   fundamentally true what are we trying to even  train our kids for because we need to Anchor   yeah we need to have a vision to Target because  if you're training for your Ferrari if that's   the meaning if that's or you're or you're looking  to become a Wall Street Banker I mean what is it   it it's it's no longer the pursuit of capital it's  the pursuit of what well it's you know capital   is there but you'll never have enough there will  always be someone who has more there needs to be   something intrinsic here and again this is where  you know for all the things religious institutions   are an anchor at times of chaos and they are there  in the poorest places in the world you don't have   to agree but they're just a story that brings  together a group you know there are other stories   and again I think we need to tell better stories  even as the world becomes more chaotic we need to   align on things like climate whereby the whole  world is hot right now or you know yeah we need   to have more positive views of that because a lot  of the discussions are negative and how can we use   this technology and come together to solve that  how can we come together as a group so that we can   share in the abundance again like I said one of  the things for this screenwriters Guild and a sag   thing may be actor coalitions that can benefit  from the bounty we may have to deploy a Ubi in   the next five to ten years so Ubi is one of the  solutions and I I do believe uh it's an inevitable   I think as especially as we start to see optimists  and figure in other humanoid robots coming online   driven by driven by our next Generation AI  able to do any and all work you know I think   taxing those robots or taxing the AI models to  generate revenue and then providing it as Ubi but   the challenge is the individual who is living off  of this and doesn't have a purpose in life and   that's the thing we need to try and figure out how  to give people more of an anchor more of purpose   because the existential angst will be Amplified  deliberately by some parties yes because they'll   be looking to take down society and you need to  create better more optimistic views of the future   you need to have anchoring and build stronger  communities and you need to empower them and this   technology is empowering again for the poorest  kids in Africa to our underprivileged communities   it can be massively democratizing because all  of a sudden they have all the expertise in the   world available Global problems local Solutions  we have to get this technology out to and they   can dream as many people they can dream they  can dream and the ROI is much larger there   than up there yeah and by the way uh you know  most people don't know this is you think about uh   Global Warfare you know what's going  on in Ukraine and Russia and so forth   it's on the whole the world is more peaceful  than it's ever been except if you take out the   Ukraine at the moment and and the challenge has  been in Africa where you have a young population   um who don't aren't clear about their future  but if you can Empower them um educate them it   transforms the world China became the engine of  growth in the world India is coming up and then   Africa can be the next one for sure if we give  them the infrastructure the technology and put   it in their hands because there's no doubt there  because there's no money yeah but there's value   and there's massive resources you'd read the  world to provide power to the world if we can   coordinate Again part of this is your own  personal co-pilot your own personal Jarvis   and I think of this as the co-pilot pilot model  we will also have AIS that we can come together   that can coordinate our knowledge in the most  important areas and allocate resources we have   to build those right because those will become  incredibly powerful but we all know that we have   enough to feed every person in the world who are  not doing it because we don't have the pilots yeah   wow uh but I just to say this again we have the  potential to uplift every man woman and child on   this planet the resources are there the ability  to create abundance and it really these are the   tools that enable that and it gets me excited and  we have to we have to guide and survive and thrive   this decade ahead yeah I think this is something  where we have to appreciate the Nuance of there   are real dangers in any upheaval this technology  will change society as we know it for our kids as   they grow up in the next decade two decades from  now completely different and again technology is   here now it's not as high in the sky everyone's  going to live in a metaverse and all this it's   here right now even if it stopped but it's not  going to stop it's only going to accelerate but   on top we're going to talk about um uh you  put out a lot of tools a lot of a lot of new   products uh and stability over the last uh eight  months since we last spoke can you give a little   bit of overview of of some of them and what are  you excited about yeah um I think we released the   first version of our language model it wasn't that  good because we were trying something different   now we're gonna try something a bit more simple so  stable LM was on your mind not it wasn't up to par   um but we're trying to figure out how to build  in the open because I think that will be key and   we're going to move to transparent building and  sharing all the mistakes that we made because I   think that's how you advance science it is on  the media side we have our first Audio models   coming out in the next few weeks but we've been  focusing on image and video so videos about to   be released in 3D we just uh participate in the  largest 3D data set so stable diffusion XL just   came out and just basically photo realistic  now um and people are integrating into things   like we had a music video competition with Peter  Gabriel where he gave his songs kindly and judged   and people from all around the world from  Burma to Taiwan created professional music   videos entirely from the song in a few days and  it's the most amazing thing to see wow yeah you   showed me some images earlier of uh me on a  unicorn and uh where was it me in a spaceship   or yeah we can put you as an astronaut on Mars on  the Unicorn we've had compositionality so you can   compose and now it's about control and so we  just released uh doodle whereby you can just   sketch and it'll do it to the sketch doodle looks  so magical uh again but you should be able to   then describe how you want to change and that's  the next version you can literally describe how   you want the image to be changed and it will  do it automatically live in front of you and   having that level of control over whatever you can  imagine just think about what people do it's from   mind to materialization really yeah it's uh yeah  it's a matter transporter idea transporter yes   where next if I could like if you're willing or  able to what's the long what's the business model   that is the most important one for you to build  towards our mission is to create the building   blocks to activate Humanity's potential so think  of every media type sectoral variants and Nation   we can create a base pre-trained model that  you can take to your own private data and we   get Revenue share license fees royalties from our  Cloud partners on-prem partners device partners   yeah these were companies and countries and people  and individuals like I have a vision of integer   internet where every single person company country  and culture has their own AI that works for them   that they are and we get paid a little bit for  bringing that to you and then you transform your   data into intelligence and it's all standardized  it all has best practices the data sets that feed   it are open at the base plus commercial licensing  as appropriate with attribution that leaps the   world forward I think I think you will also use  the open AIS and Googles of the world maybe those   as consultants whereas these are people that  you hire you hire the AIS because they work   for you they they know you intimately because  you can share everything with them without fear   and then when needed you go to these expert  AIS the med Palms the gpt4s and others and   you combine those to a hybrid AI experience  that's massively useful so when I'm when I'm   using GPT for when I'm using chat GPT or Bard uh  what does openai know about me in that point so   now they've offered opt out for gdpr reasons in  Europe so you can click that otherwise they were   just trading on everything that you ever did and  understanding the nature of humans interacting   they don't care about you necessarily per se  right just using you as part of the training   but I've heard a number of companies saying you  cannot use open AI well you can't use it for any   regulator data you can't use for any government  data because that's not allowed to leave the cloud   environment or the on-prem environment that's why  you need open models like ours again if you're in   a high security Pentagon situation yeah you can't  really bring in Consultants unless they're super   super ultra vetted you hire your own grads but  even you within your company you're not going   to make it all contractors are you you're going  to build up your own knowledge base build up your   own kind of grads but sometimes you might bring  in a consultant so that's the best way to view   these generalized models that are highly that are  very very very good and models that adapt to your   data and so that's where we come in models that  adapt to your data that you own and we get Revenue   share license fees and royalties for doing that  and more importantly we bring this to the world   so we will bring it from Indonesia to Vietnam  to everywhere and train local models that will   then allow these economies to Leap Forward open  versus closed you've made the argument um we're   seeing meta you know as you said 80 open yeah  they won't release the data sets or things like   that or customized versions but them releasing the  technology means that everyone can optimize their   technology which reduce the cost of the technology  because their business model is about serving ads   and so this is why it makes sense for them and  what are your thoughts on on elon's recent uh   announcement so Elon had a xai announcement you  know he discussed this on his Twitter space of   course of course saying you know it's an open AI  competitor he is very worried about AGI coming   by 2029 and he wants to build a truth-seeking  curious AI that can understand the universe   because that'll be the objective function of the  AI because objective functions really matter when   we're teaching our kids when we're creating  something and so I think again this is going   to be a multi-modal AI that can understand a whole  bunch of things and there'll be a whole series of   announcements there but the timelines are so short  in the view of just most of the experts here five   to ten years you know it's so funny like you know  Ray's been consistent on 2029 forever and every   conference and we talk about this that everyone  would say that's ridiculous it's if ever gonna   happen it's 50 or 100 years away then it was well  it's 30 years away it's 20 years away it's five   you know it's and they've converged on on Rey's  prediction though there are some and I'm curious   where you are that think you know first of all  how can you define AGI it's a it's a movie it's   a moving blurry line but are those who you know  believe it's here in the next two years well just   like the Turing test right the Turing test was  can you have a discussion you don't know it's   a computer obviously now you can yes we can see  it live in front of us now the Turing test has   just been increased in this so yeah like we move  the Finish Line nobody knows because again we've   never come across something that's as capable as  us you know for the first time just now we've had   the medical AI outperform humans yeah we've just  had it can do the GRE and GMAT and LSAT and mit's   uh ECS uh yeah 2023 was the year finally tipped  yeah and so we have no idea what's coming I said   for me I think there's only been two logical  things that can reduce the risk even though   I think it's gonna be like that movie Her Like I  said humans are boring goodbye and thanks for all   the GPS I could be wrong that's why I signed both  letters one is feed it better data that's what I'm   focused on it's a good business model it's good  for society and it's good for safety and nobody   else is doing this nobody else is creating this  as a Commons for the world which is why I created   stability for that reason which is why it's called  stability despite it being a crazy hyper growth   startup number two and this is what most of the  labs are trying is what's known as a pivotal   action okay what is that the only thing that can  stop a bad AI is a good Ai and the way that you   do it is you make the good AI first and then it  stops any other AGI from coming into existence   by seeking and destroying that capability and  that is terrifying to me yeah and that's what   you actually hear when you talk to the people  that are building these Labs with a focus on AGI   they can talk about discovering the universe  and everything like that when you come down   to their alignment things they're like we will  figure this out we're not sure but this could   work you're like and and we go figure it out even  though it's progressing exponentially or you know   double exponential and we hope we'll figure it  out in time we hope we'll figure it out in time   and if anyone should figure out it's us because we  know the best and in their own words like you read   open ai's path to Aji again opening hours full of  wonderful people doing great things and I use gpt4   as my therapist and all sorts of things because it  doesn't judge me unless I wanted to right it says   this is a we believe this is an existential threat  to humanity that will end democracy and capitalism   and you're like okay and you're building  it in your back room you're building it you   know and they're like why are you building it  because someone has to otherwise someone else   will build it and you're like this is dangerous  but the reality is we don't have better answers   and again I went down to I'm trying to build  a great organization it's really really hard   there are no real comparators to what any of us  are doing and it's going to get more and more   crazy the only thing I could think about is  like you are what you eat and so I heard that   our contribution can be bringing this technology  to the world so that the world can be the Dynamo   Africa or and Asia and others building better  data sets so no one has to use scrapes so we   feed the models better stuff and bringing some  standardization around this to drive innovation   we're truly at the 99th level of the gameplay we  got hits the boss round oh yeah like I said but   please do put your uh YouTube comments through  gpt4 so they're nicer to read okay you might I   could spend all day um and there's probably very  few things if anything more important than these   conversations right now um it's the time it's got  a window of a year or two maybe less wow on that   thought uh I look forward to our next conversation  to abundance thank you my friend cheers

Channel: Peter H. Diamandis

Views: 124,980

Keywords: peter diamandis, longevity, xprize, abundance

Id: 1WOjjgyZPj8

Length: 105min 13sec (6313 seconds)

Published: Thu Jul 20 2023