Creating The Future With Artificial General Intelligence (AGI) With Peter Voss Full Interview

[Music] [Applause] [Music] hey I'd like to welcome you to another episode of mission matters my name is Adam Torres and today I am coming to you from Brentwood California and I'm at Studio Place La and I am excited because Peter Voss is coming to town and we're going to have a whole lot of fun um first off Peter I want say welcome thank you for coming into town yes it's good to be here so we've been trying to get this together kind of working on this through uh shreni and some of your team for now what a solid I feel like it's been a year year and a half or so yeah it's been a few years yeah so um we got a whole lot to talk about today I want some updates on what's going on with uh io. a we're going to talk about which I'm really excited about AGI which I know you're known for coining which is artificial general intelligence and we're going to talk about where you're at with that and a whole lot of other things plus I mean I've been looking at what you've been doing and I I feel like um iigo you got a big Vision a whole lot to accomplish and I know there's going to be a lot of people involved so we'll get into that as well um but to get us kicked off we'll start this episode the way that we start them all and all our audience at home knows that's with our mission matters minute so Peter as you're aware at Mission matters our aim and our goal is to amplify stories for entrepreneurs Executives and experts we want to get their message out to the masses so that's what we do Peter what mission matters to you well that's very obvious it's AGI artificial general intelligence and that is to bring intelligence artificial intelligence to the masses that will improve people's flourishing and create radical abundance and that's really the mission that I'm on and have been for some time that's that's exciting and we're going to get way more into this and one of the things is we've heard about AI we've heard all these things in the news let's say for the last couple of years where it's mainstream um IO and io's been you've been working on these field over 15 years on these types of um I'll say not problems I'm going to say opportunities for Humanity um and we'll go into this but first i'm just curious because I followed your career I followed your journey since I since iOS come on my radar and I just want to know like when did you consider or think like that science and becoming a chief scientist or other wise when did you feel like that was going to be a part of your destiny like how how did that come up for you I I don't really know I can't pinpoint it specifically I think I've always been interested in scient scientific things that that's been balanced out by also sort of this interp entrepreneurial spirit that I inherently have I I don't only want to discover things I also want to have them work in the real world so they really these these two sides of the coin that have always fascinated me m now I found I mean I've done over 6,000 interviews I've interviewed a lot of people and I feel like a lot of times there's people when they're um really highly educated and I'm not putting anybody in a box right nobody wants that but sometimes when people are really Highly Educated they don't have that other piece to where they want to go out and step out on the limb and take that chance of becoming an entrepreneur not nothing against them right everybody has their own path but where did you get that piece from like that extra to make you want to go and see your vision of reality well actually I didn't finish high school so I got thrown in the deep end pretty quickly in terms of having to look after myself and you know find work when I was 16 years old um so it was kind of quite natural for me to figure out how to survive in in in the world and that kind of just led me to figuring out how to create things to create more things and there's also been this curiosity to learn things to figure things out and to invent things mhm yeah and and thinking about the the concept of I always tell people this I'm like for my long-term listeners they're like oh Adam what being an entrepreneur this that to me I was just trying to make some money I didn't know that's the fancy word we use now but it was just trying to provide value trying to figure out a way where I was going to fit in the world so thank you thank you for sharing that because that's I think inspiring as well to our audience um that are maybe in a similar I mean we're just coming off so when we're recording this just for everybody at home 2023 or excuse me 2024 we're already in another year all right P post pandemic we're all kind of in this thing where we're trying to figure out what's next many times and we're all growing in our own pace so I feel like that is very useful for our audience um I usually like to try to work in at least one Pay It Forward question and that's maybe for some of our younger entrepreneurs or people that are just getting started if you could go back to that that Peter that's kind of whatever age group I'm not saying it has to be 16 could be whatever you want to do and you give him some advice and what's on road ahead what kind of things would you tell them well I can think about sort of the biggest regret or a big regret I I have is that I didn't start my own company until I was 25 years old wow and I wish I had started earlier so if you incline towards the sort of entrepreneurial uh life and not everybody is then I think the sooner you can start you know even while you're at school if you do it part-time and get get your feet wet I think that's really useful there's nothing like actually doing it being responsible for for making things happen for you know getting customers getting a product out the door getting paid for it you know all all of that so yeah that's one one of the big regrets that I waited until I was 25 before I actually started my first company thank you thank you for sharing that I I feel that very that that helps me um so I'm sure it'll help our audience at home as well um so jumping around here a bit I I do want to Circle back to to the main event here today so AGI um first off for our audience that aren't Chief scientists or maybe uh scientifically inclined maybe talk to talk to us a little bit about just what that concept even means yes so artificial general intelligence is uh a term that uh three of us coined um in 2002 and it was really to get back to the original Dream of AI the the term artificial intelligence actually goes back 69 years it was coin 6 years ago and the original idea was to build machines that can think learn and reason the way humans do and there some very unique properties that human intelligence has and that was the original Dream and they thought they could crack this in a year or two rough 69 years ago I was just going to say yeah now of course it turned out to be much much harder so what happened in in the field of AI that over the the decades it changed into a field of narrow AI and there's actually a very important a subtle but very important difference and that narrow AI it's solving one particular problem at a time so um for example IBM's uh deep blue is a good example of the world chess champion in the '90s where you're solving the problem of playing a good game of chess but it's really the intelligence the external intelligence of the data scientist the programmer how to solve the problem how to build a machine that can solve that particular problem so in in in 2001 I got together with with some other people who also felt the time was right for us to get back to that original dream that original m mission of of of AI to build general intelligence where the machine itself is intelligent enough to figure out how to play a good game of of Chess without somebody specifically programming it to do that and that's what artificial general intelligence really is is about now at the time we didn't know that the the term would kind of BP arici intelligence and now of course the term has been widely used especially recently yeah so I I I like I want to stick in those early days a little longer because I don't I don't always get a Pioneer on the show from from this field so um was it fun like was it fun having those conversations and dreaming and ideating in the beginning like what was that like like even just in the early days well actually I I came to America in 1995 and to me it was like a kid in a candy yeah that's what I'm thinking when when I came to California I was introduced to some of my heroes you know authors of books that i' I read in in the field of AI and Longevity futurism and so on so it was really really very exciting uh time and you know then of course you get down to doing the actual work of it's hard so at least the ideating part you know we get a little bit of fun the dream big you know 20 plus years later I'm still as excited about that mission and uh and that brings us a little bit more to present day speaking of recent day and I just wanted to set the set the stage for how long you've been working on this um first off but nowadays I mean and the obviously I'm in media you know lots of lots of interviews and technology so nowadays AI in every headline and everything it reminds me of um it reminds me of kind of the original you add to any company and all of a sudden right it's supposed to be worth something so similarities I know there's some separation but that's the closest thing I can think of that I'm I'm hearing now at least in media specifically to be specific there so why does all this matter right now like why does it matter are we at a pivotal time like why does all this matter yes so I think especially with the release of chat GPT just over a year ago people are getting a real taste for what AG might be like a machine that can really do the things that humans normally could only do so I think it's giving us a taste of of this but at the same time um people are also realizing people working in the field are realizing that these large language models like cat GPT are actually not going to get us to this m to this Vision that we have of really truly generally intelligent machines machines with you know human like intelligence uh for example just a few weeks ago uh the chief AI scientist of um meta uh Yan Lun said uh large language models are an off-ramp to AGI wow they are distraction they're a dead end now you have to balance that against the hype on the one hand and the enormous amazing things that large language models can do you know like summarization and writing poems and you know what what amazing things that they can do but uh it's also become clear that it's fundamentally not the right uh right approach to to get us to AGI and I um I got I got to tell you my AI story so once my mom called me and said something about like how do you use that GTO or that I was like the car mom like what do you mean she's like the G something you know that this that I said uhoh we're getting close once my mom finds out something is happening I'm like we're definitely mainstream because we're getting close yeah in fact yesterday I was I was talking to somebody and and he said my parents are really amazed with it because they're using it they were using it for a to create a thank you letter to somebody and you know they were amazed how well it it it did that so yeah absolutely the technology can do some amazing things but it's not the real thing it is not the real thing and that's a big distinction I want you to let's get you're welcome to get as technical as you want to let's go a layer deeper on that CU I don't want to assume that everybody understands the concept of language models or why exactly it's not the right thing so maybe take us just a layer deeper there yeah so before I do that uh I think to to really be able to make the contrast I want to talk a little bit more about what AGI really is and what the implications of AGI are cuz the enormity of having truly human level AI um you know there's the general consent that that's going to be bigger than fire bigger than electricity you know internet how transformative it it it will be and and and can be so you know for example with real AGI uh for one you can have ai scientists so imagine one AI training itself up to be a cancer researcher yeah um you know at you PhD level cancer researcher you can now make a million copies of that yeah you have a million PhD level cancer researchers chipping away at the problem pursuing different Avenues and communicating with each other sharing knowledge much more effectively than humans can you know without egos getting in the way and so on yeah um and now you you can take that across really anything that requires scientific research you know whether that's better Battery Technology you know nanotechnology Nanobots that can go in our bodies and repair damage um you know any scientific field you know or or even governance I mean we don't seem to be very good at creating the right kind of governance for our for our society so being able to apply more brain power to problems of humanity so that's one area sort of research scientific uh approaches uh another area is so dramatically reducing the cost of goods and services by Automation and you know how that will create radical abundance I mean just by lowering tremendously lowering the cost of things uh so that's kind of the second area uh a third area I'd like to highlight is what I call a personal personal personal assistant so reason I put three personals there they're really three different important aspects of of the word uh the one is you own it it serves your agenda not some Mega corporations agenda and secondly it's hyper personalized to you so it really gets to know your history your goals you know who you interact with and and so on and the third personal is is the issue of privacy that you decide what you want to share with with whom so I have that you know vision of of applying AGI to giving everybody in the world this personal personal assistant it'll be you know like a confidant a helper U that can help you make better decisions a bit like a little angel on your shoulder you know that wow help help you avoid make make bad decisions and so on uh so I think that don't we all need one of those hold on don't give me a when when can I like get one of those I'm in all right sign up for it yeah in um so you know I think there are so many ways in which uh true AGI can can increase human flourishing tremendously um so I think it's it's really to understand how big a deal AGI is yeah to then say why we should work on it you know so getting back to the the question you asked is you know technically sort of why are large language models in offramp why do I agree with that so there there are a number of fundamental inherent problems with the technology um and in fact the the the name GPT already gives you a very good clue g means generative it makes up stuff mhm uh you know it's trained with 10 trillion pieces of information from the internet from all over the place good bad and ugly sure and so the system will create responses that are picked from that Knowledge from that huge knowledge but it doesn't know what's right or what's wrong so it just creates it generates it makes up stuff and that's why these system suffer from what people call hallucinations mhm you know you can't trust them basically so so that that's one big problem the second big problem is uh GP the pre P stands for pre-trained MH they are pre-trained uh with these 10 trillion pieces of information at a cost of hundreds of millions well I think the latest ones are like $200 million to train the system to do all the number crunching to take this massive amount of information and kind of condense it and build a a model that you can can use so that pre-training also means that they cannot learn in real time so if they hear something new like the conversation we having yeah uh if it's not something they already know they can't learn it they can't update that so it's a huge limitation and it's it's it's inherent there other limitations they don't have metacognition they can't think about their own thinking they don't know what they're saying yeah they don't know when they don't know things um and I I think the other thing that really worries people is the the vast amount of um electricity they need and comp computational power they need I mean we now have Sam Alman uh saying he needs to buy nuclear build new nuclear fusion nuclear power stations to do this you know uh $7 trillion worth I mean it's it's insane to talk about the amount of sort of the environmental impact and the amount of of power it needs whereas if you compare it to our brain it's 20 wats of power not 20 GA so you know large language models that a whole approach GP pre-trained and t uh talks about the technology Transformer technology which really locks in uh the need to pre-train and you know the whole cost expansion and so on so that's why it's an off-ramp you know and um why we need something else yeah I want to I want to spend some time um that we have here today um going further into io. and exactly where you're at with the company and the technology maybe talk to us a little bit more about that sure so um you know obviously the question then is if large language models aren't going to get us to AGI do we have an idea of what might and this is really what I've been working on for the last you know 20 20 plus years and so we've developed something that is called cognitive aih mhm um and its starting point is how people think and what's important in human intelligence and we can you know we can talk more about that but we've basically been building uh both prototypes and development prototypes but also commercial applications very successful commercial applications so we've been alternating between uh development on the one hand and commercialization so the commercialization gives you that um that reality check you know of in theory you can build something that sounds good but it also has to work in the real world so we've been alternating between uh developing the technology and then commercializing it um and you know that we we find that a very good model and we believe that this cognitive AI approach will actually get us to AGI now are there any um and not to put you on the spot but are there any like use cases or companies that you've worked with up to this point with where you're at oh yes absolutely um I mean we we've um automated hundreds of of call centers over over the years very successfully um the the one example one recent example that I can give is uh the 1 1800 Flowers group of companies including Harry and David and popcorn factory and so on it's about big Brands we all know and use in love yeah it's about 12 different companies so we have very successfully uh automated their their call center operation wow and uh so for example two months ago or just over two months ago when we had Valentine's Day uh we actually replaced 3,000 agents in their call center that normally they have to hire to handle this big spike because you know Valentine's Day everybody few days before you know calls in um and we've been able to do that uh with a 90% self-service rate you know where people can at any time they can ask for a human but our system is so good that it just does a job there's no waight time it has access to all of your information so you can I feel like some of the people listening right now they're like that that order from 1 1800 Flowers they're like wait a minute what was going on last year yeah it's amazing continue please yeah so you know we we very much on the commercial side um but you know there's still a lot of integration and you know we're not at human level intelligence there yet with a with a current commercial technology that's why we so so Keen to really focus on moving that technology forward and taking it all the way to to human level yeah now um what you're describing is amazing and I just have to think about like we're you know we're we're a small Media company I like to think we punch above our weight right but we're not trying to like change to this point so when I think about the amount of money or people or resources or collaborations that's needed to make this Vision or it like what does that look like I just can't even fathom like what does that type of thing look like yeah so the the amazing thing is with uh cognitive AI as opposed to uh generative AI cognitive AI doesn't need 10 trillion pieces of information to to get trained it's you know maybe a few million pieces of information so like a million times less MH uh which means it also doesn't cost 100 million or 200 to train the model so we talking about a much much lower cost to to produce it and not have the environmental impact all the negatives that go with it so so it's a viable alternative like it's a well yes uh the alternative isn't viable exact that's what I mean is viable so it's a viable alternative absolutely and um so you know we we currently on the development side we have a very small team there only 10 10 people on the development team right now so we are obviously wanting to uh increase it but we're not we're not talking about hundreds of millions of dollars you know we're talking about tens of millions of dollars not hundreds uh to actually get the technology um expanded improved you know from from where we are to get to human level what what what's a good type of um and you don't have to say any names but what what type of partner or like in this Venture would you be looking for like what do you think would be a good fit do they have to have like previous experience in that part of field whether it's investor or otherwise like what what do you feel like would be a great partner that's a very good question now with AGI being such a powerful technology um we are really looking for partners for Financial Partners who share our philosophy in terms of helping human flourishing MH because you can use it you can focus your development and your company development on how can we make the most money the uh that I'm not interested in this is a multi- trillion dollar opportunity so the money is not a not even a question but what is important is to develop the technology and to deploy it in a way that's really going to be beneficial to to humans you know for example that personal personal assistant when I say the first personal you own it it serves your agenda is very different from what all the big players are doing MH when they try to give you a so-called personal assistant they control it it serves their agenda you know and like you know Alexa and Siri and you know so it's having the the right kind of partner uh who who really wants to make AGI happen um that that's kind of really important for us and I I like that you bring this up and you didn't use this word but it's one that is often linked with around that conversation and it's Ai and ethics and ethical Ai and there's I've heard a lot of different terms um for Humanity right like what's going to happen next um interested in hearing a little bit more about just your concept of what what you feel like AI ethics like you know for the other leaders out there that may watch this even in the future like and and they're developing Technologies or other things like their responsibility like any thoughts on that yes absolutely in in fact um before I started working on AI um I actually took off five years to study philosophy wow psychology and AI uh to really understand um so many different aspects about thinking and living and values and and so on so in philosophy I studied for example epistemology theory of knowledge how do we know anything what is reality what's our relationship between our knowledge and reality but I also looked into um how do we know right from wrong ethics so I was fascinated with where that I actually spent six months uh trying to really understand Free Will and what free will isn't isn't because that's kind of key to ethics you know if we don't have control over ourselves or to what degree do we have control of ourselves to what degree can we be responsible for our our actions and then how do we know good from bad so I studied these various fields and of course also in the field of AI and cognitive psychology how do children learn Lear how does our intelligence differ from animal intelligence uh what do IQ tests measure you know so really to get a deep deep understanding of uh of of intelligence and you know to get back to your your your question is um it's unfortunate that in Hollywood you only seem to be able to make money if AI is the bad guy and destroys everything you know we have that kind of preconception of whoa AI is going to kill us all but really when you when when you analyze how we actually build AI not some theoretical model that you can build for yourself that can scare you to death but if you actually look at how we build AI we build AI to really help us to serve us that's actually how how it's built and the rationality of AI actually helps us make better decision and helps us avoid things that we would consider bad or even immoral um because rationality tells you in a way that um you know for example that a zero sum gain doesn't make sense you really want to engage in interactions where both parties benefit and so those are the kind of Ethics that come out of a rational approach I've I've actually written uh on sort of a rational approach to ethics and and how you know how that so that informed informed me as well in terms of how I see the future Ai and how AI can help us be more rational and be more moral amazing and and to take those 5 years and to do what you did I mean you're man of many surprises I'm I'm enjoying getting to know you better and getting to know more about like the heart behind the company and the vision um so thank you um Peter last question here um if I can if we could allow ourselves to dream um we're you're an entrepreneur I'm an entrepreneur many entrepreneurs at home that are watching this as well um you're your vision like the big the dream the what's next like of it for you like if it what like dream for a moment like what's next yeah well I think that's very clear I want AGI to happen as soon as possible and I'm looking for people who share that Vision to actually make it happen and so we can all benefit um from AI to have this radical abundance and increase human flourishing amazing thanks again so much for coming on the show and to the audience at home if this is your first time with Mission matters I hope you really enjoyed this episode um I look forward Peter to seeing AGI come to the masses and come to fruition and and to have you on the show as one of the original Founders who coined that term absolutely amazing and again to the audience at home if you haven't already done it yet hit that subscribe button this is a Daily Show each and every day we're bringing out new guests new thought leaders um new ideas and hopefully new inspiration to help you along in your journey again don't forget hit that subscribe button and we're going to put all of io. AI information in the show notes so that you can just click on the links head right on over and follow their Journey as well [Music] [Applause] [Music]

Channel: AigoAI

Views: 4

Keywords:

Id: yXiqBj78MsA

Length: 29min 56sec (1796 seconds)

Published: Tue Jul 16 2024