Conversations on Applied AI - Stories from Experts in Artificial Intelligence

Peter Voss - The Third Wave of AI

July 13, 2021 Justin Grammens Season 1 Episode 22
Conversations on Applied AI - Stories from Experts in Artificial Intelligence
Peter Voss - The Third Wave of AI
Show Notes Transcript

This week I am thrilled to be joined but a serial entrepreneur, engineer, inventor, and pioneer in Artificial Intelligence, Peter Voss. Peter coined the term "AGI" (Artificial General Intelligence) in 2001 with fellow luminaries in this fascinating area of Artificial Intelligence.

Peter is the Founder/ CEO and Chief Scientist at AGI Innovations and IGO and published a book on Artificial General Intelligence in 2002. He is a technology innovator who has for the past 20 years been dedicated to advancing Artificial General Intelligence. His experience includes founding and growing a technology company from zero to a 400-person IPO. In 2008 Peter founded Smart Action Company, which offers the only call automation solution powered by an AGI engine. 

Peter is currently focused on commercializing the second generation of his AGI-based "Conversational AI" technology called "Aigo". It is implemented using a brain-like cognitive architecture – also known as "The Third Wave of AI".

Peter also has a keen interest in the inter-relationship between philosophy, psychology, ethics, futurism, and computer science. You can follow and read his articles on Medium.

If you are interested in learning about how AI is being applied across multiple industries, be sure to join us at a future AppliedAI Monthly meetup and help support us so we can make future Emerging Technologies North non-profit events! 

Resources and Topics Mentioned in this Episode

Your host,
Justin Grammens

Peter Voss  0:00  
Fundamentally, all of the chat bots that are out there, except for ours are chat bots without a brain, I was the only chat bot with a brain. And basically, if it doesn't have a brain if it doesn't have a cognitive engine, if it relies on deep learning machine learning big data, it's fundamentally down the wrong path. It's essentially a waste of time going in that direction, all the energies that are being put into current chatbot. Basically a waste of time. And it'll have you know, less and less of an improvement. You just can't brute force intelligence like that you have to have the right architecture or you know what DARPA calls the third wave of AI. So the challenges we facing is really just being a relatively small company is having enough resources to develop our brain our cognitive architecture.

AI Announcer  0:51  
Welcome to the conversations on applied AI podcast where Justin Gremlins and the team at emerging technologies North talk with experts in the fields of artificial intelligence and deep learning. In each episode, we cut through the hype and dive into how these technologies are being applied to real world problems today. We hope that you find this episode educational and applicable to your industry and connect with us to learn more about our organization at applied Enjoy.

Justin Grammens  1:22  
Welcome, everyone to the conversations on applied AI podcast. Today on the show we have Peter Voss Peter is the founder CEO and Chief Scientist at AGI innovations and Ico theater coined the term AGI back in 2001 and published a book on artificial intelligence in 2002. He's a serial AI entrepreneur, technology innovator who has for the past 20 years been dedicated to advancing artificial general intelligence. His experience includes founding and growing a technology company from zero to a 400 person IPO in 2008 Peter founded smart action company, which offers the only call automation solution powered by an AGI engine. Peter is currently focused on commercializing the second generation of his AGI based conversational AI technology called Aygo. It is implemented using a brain like cognitive architecture, also known as the Third Wave of AI, which I'm super excited to talk more about. Peter also has a keen interest in the interrelationship between philosophy, psychology, ethics, futurism and computer science. So thank you, Peter, so much for being on the show today. 

Peter Voss  2:19  
Yeah, thanks for having me. 

Justin Grammens  2:21  
Awesome. Well, I'm super excited to have someone with your deep experience in the program to talk about AGI. Could you maybe tell us a little bit more about your path of getting into artificial general intelligence? 

Peter Voss  2:30  
Yes, certainly. So I started off as electronics engineer, started my own company then fell in love with software. And my company turned into a software company. And that's a company that then grew very rapidly, I developed a eirp software system, just love programming, and became quite successful. There's an IPO. So that was awesome. It's really when I exited the company, I had enough time available to think what big problem that I really want to tackle next. And what struck me is that software isn't very intelligent, it's pretty dumb. And you know, I was very proud of my own software. But whatever the programmer doesn't think of, you know, it'll just crash or have an error or not have any common sense. So, really trying to figure out how we can build intelligent systems. And that's how my whole story on AGI really started. Got into this was early on, right? I mean, this was, could this have been 20 years ago or so he started thinking about some of the shortcomings. Yeah, actually, more than 20 years ago, I actually took off five years to study intelligence, because I figured that I really needed to understand what intelligence is many different aspects of it to be able to design an intelligence system. So I studied epistemology, your theory of knowledge, how do we know anything? what is reality? What is our relationship to reality? Also, what do IQ tests measure? How do children learn? How does our intelligence differ from animal intelligence, all of those different aspects of intelligence to really deeply understand that, and, of course, to study what other people had done in the field of AI, you know, over the decades, so it's a combination of that research that I really came up with a design for intelligence engine and 2001. As you said, I coined the term AGI together with two other people. And we wrote a book on a topic to really differentiate ourselves from what pretty much everybody in AI was doing then, and it's doing now and that is narrow AI. So we really wanted to get back to the original dream of artificial intelligence, you know, that when the term was coined? 60 odd years ago, the idea was to build a thinking machine, you know, a machine that can think and learn and reason the way humans do. And we felt in 2001 that the time was ripe to give us another shot, you know, the technology, hardware, software technology, or advanced sufficiently to have another go at that. But it still was pretty mature at the time, right? There wasn't. We're not talking about GPUs or really a lot of data at that time. I mean, what have you seen that has changed now over the past 20 years, surprisingly, I don't believe that hardware limitations are the biggest thing. I think it's having the right design and having obviously enough effort put into it to be able to solve the problem. Of course, as we get more processing power, you know, more memory, things become easier, it becomes easier to set up tests to run tests to train systems and all of that, but I don't really think that that's sort of the biggest thing. I think we could have made a lot more progress even 10 years ago, or 15 years ago, if more people had actually been working on AGI, you know, rather than just narrow AI. Yeah, well, how do you define narrow AI for listeners, maybe that aren't familiar with that term. So as I said, the original vision of AI was to build a thinking machine, a brain that can basically learn the way humans do. We are generalists. We can learn lots of different things, you know, we can learn to play chess, we can learn to do medical diagnosis, to, you know, play music, or whatever, all sorts of things. So that's our general intelligence. And of course, we are also told users, we can learn to use tools to extend our capabilities. Now, that was the original idea of AI. But then, that turns out to be really, really hard, especially 60 years ago, I mean, when I started writing software, we were working with machines that had 16k, of memory, you know, and we managed to write programs that could do useful things, but to now work with 16 gigabytes of memory, and you know, is obviously much, much easier. So it turns out to be really, really hard to build AGI or to build, you know, thinking machine. So what happened to the field of AI is, they said, Well, we can solve particular problems, you know, we can take like chess playing. And that's, of course, the famous one IBM, Deep Blue beat the world champion. But what I've found, what I discovered is that it's really the programmers intelligence on today, the data scientists intelligence, on understanding the problem and understanding how a computer can solve it. So it's the intelligence that's really turned into code to solve that one particular problem. So you now have the ingenuity of people that can build a chess playing machine. But it can't do anything else can't even play checkers. Right? It's so narrow, and hence the term, right? Yeah, it's hard to narrow. But actually, what is more important than you realize this a few years ago is that it is external intelligence. So it's really the programmers intelligence that you are turning into code to solve that particular problem. Whereas what you want is you want a machine that has the intelligence to figure out how to solve the problem. And that has the ability to learn how to solve lots of different things that AGI that artificial general intelligence, but there's some other requirements as well, that the current approaches of deep learning machine learning Big Data simply don't need, you need to be able to learn interactively one shot anyway, we can talk more about that, we definitely will definitely will, I do want to dive more deeply into that you got me thinking about just yeah, if you were to write a program, Peter and I were to write a program, and we don't pick your favorite language of choice, we'd each have our own little nuances in there of how we would program it ourselves. And I think that's maybe what you're saying is, is now the programmer, I mean, there's just, it's just logic, it's just sort of like dumb logic that a program is going to put in, but you'll do it slightly differently, I'll probably do it slightly different, you know, we might optimize it one way or the other. 

Justin Grammens  8:22  
But this idea that you would actually flip the whole model around, you know, when I start to talk to people about machine learning, there's no longer an if then else statements in here anymore, right? It's the whole pack of it, you're just feeding it a bunch of data, and the machine kind of decides on what the right answers are correct? 

Peter Voss  8:37  
Well, except one is to be careful when you start using words like decide, it doesn't actually decide. I mean, the big data approaches deep learning machine learning are purely statistical systems. They basically simply do pattern recognition. And based on the data, you feed them, and whatever feedback you give them, they basically create categories or in some cases, sequences that they can generate. But there's no thinking going on. And there's no learning going on in the sort of traditional sense. That's why also your systems are inscrutable they can't explain what they're doing. It's a completely blind process and and of course, opaque process as well. And AGI the whole idea behind that is to completely open that up. I mean, how would that be different? is you actually saying it would actually would make some decisions? Exactly. Yeah, absolutely. So what AGI is to have a thinking machine or an intelligent machine, you really need to start by saying What does intelligence require? The kind of intelligence we're looking for, you know, human, like intelligence are certain hard requirements that you have, and some of them are, you have to have memory. You know, people aren't classed as intelligent if they can't remember something, you know, they heard two sentences ago, if they were paying attention, you know, memory is a crucial part of intelligence. If you didn't have memory, you couldn't really do intelligent things, because you'd have To keep experimenting and learning, so memory, short term memory, long term memory, the ability to reason to make sense of what you're hearing to be able to use context, because things and events can have a very different interpretation, depending on what the context is, you know, are you at work at the moment? Or are you at home? Who are you talking to? What do they already know? What are you trying to achieve? What is they already know? You know, what knowledge Do you have, all of those things impact how you actually react to a particular stimulus. So being able to take context into account to be able to disambiguate, if you're not sure, you're not to be able to kind of probe or experiment to interact with the world, to be able to learn in real time, with limited resources, you know, not infinite resources, but with limited resources. So all of these are requirements to have an intelligence system that you could, for example, use as a personal assistant or an elder care assistant or call center system that could really handle you know, more complex things that aren't just scripted. So the need that intelligence, you know, to, to remember what somebody said two sentences ago, you know, what they said last year?

Justin Grammens  11:15  
Yeah. And that's, I think that's where a lot of your focus is, at least, you know, and I said during the intro around call centers, but maybe you're taking it up to the next level, I guess, with your current company.

Peter Voss  11:24  
Yes, absolutely. So from a commercial point of view, I mean, I'm personally, I've always been very interested in technology, but also in business, I love the business side of things, I don't just want to understand things or discover them. And you know, then I wouldn't care about whether that actually practical or use now, you know, it's very important to me to also see that what I innovate has a practical utility creates value. And that, of course, in the business side of things, and, of course, you need to be able to create value, generally, to be able to get funding to do your research and development as well, you know, so they all can come together. So yes, at the moment, we are focusing on working with large enterprise companies to help them in whatever whether it's external facing customer support sales, whether it's internal helping an HR or an IT or in their sales department, those are the kinds of things we're doing. But we're also working with companies where we're helping them create intelligent conversational AI for training, for example, VR, or AR Training Systems, we've had interest from car companies to have intelligent, hyper personalized interactions with infotainment, and that the car gets to know you basically, sure, medical diabetes management is something that a pretty obvious one, but in the longer term, I see as also being able to offer a personal assistant to the individual. In fact, we have a special name for that, that we registered as a personal personal assistant. And it really should be a PPP II, because of the three different meanings of personal, so personal, you own it, it's yours, it's your agenda. It's not owned by some mega corporation to serve their agenda. So personal ownership, the second one is personalization. So that it's customized to you, it's not a one size fits all, it's personalized, it gets to know your preferences, your history, and so on. And the third personal is that it's personal in, in the sense of that is private, so that you decide what your assistant shares with whom Sure. So that's the kind of vision that we have that will offer these personal assistants that can then be utilized at work or at home, you know, can communicate with each other. And they'll shield you and protect you from these other bots and things out there.

Justin Grammens  13:50  
Some of the other things around your home, you mean, some of the some of the Google homes and the Alexa's and all that type of stuff?

Peter Voss  13:55  
Yeah, your personal personal assistant can deal with them, you know, and also be a little angel on your shoulder to keep you out of trouble, you know, to say, Hey, you know, maybe you want to think about this first, before you jump in there. You know,

Justin Grammens  14:07  
I think that's a great idea. Yeah. I think if people had somebody watching over their shoulder, asking you, you sure you want to do that? I think people would make a lot better decisions. You mentioned, the conversational AI, you know, all the technology around that has been fueled been hard at work, you know, doing that feels like over the past decade or so. And it's getting better and better. But you're right, it is still pretty stupid. You know, and the fact that I can ask it something once, but it doesn't remember the context of stuff earlier on. Right? That's one area you're trying to tackle. Are there other areas that you're kind of, I guess, where you're finding the most interesting challenges where you think we need to go with this technology needs to get better at?

Peter Voss  14:43  
Yeah, so I'd like to make the point that, you know, when you say a lot of work is being done and chatbots that that is true. I believe Amazon employs 10,000 people on Alexa. I mean, the mind boggles, you know, but fundamentally all of the chatbots that are out there Except for ours are chatbots, without a brain, I was the only chat bot with a brain. And basically, if it doesn't have a brain, if it doesn't have a cognitive engine, if it relies on deep learning machine learning big data, it's fundamentally down the wrong path. It's essentially a waste of time going in that direction, all the energies that are being put into current chatbots are basically a waste of time. And companies are finding that out. You know, we talked to large corporations all the time, and they've tried to implement chatbots. And almost always, they are highly disappointed in what it can actually do. Because they using big data technology is basically the quantity of that and not the quality of data that they focus on. And what you're finding is you can train the system was 10 times as much data. And I mean, you get things like GPT, three now, which has our trillions of effects, I don't know what fantastic number, you can have 10 times as much data. And it'll have, you know, less and less of an improvement, you just can't brute force intelligence like that, you have to have the right architecture, or you know, what DARPA calls the third wave of AI. So the challenges we're facing is really just being a relatively small company is having enough resources to develop our brain, our cognitive architecture,

Justin Grammens  16:19  
he mentioned the third wave of AI, I mean, what's the first two,

Peter Voss  16:23  
so you adopt this presentation A few years ago, where they talk about the three waves, and the first wave is basically, fundamentally sort of logic approaches to AI, which were very prevalent, you know, 70s 80s 90s expert systems, but it's basically based mainly on formal logic, but also some statistical approaches. But it's sort of the logic mathematical approach to AI. That's the first wave. And then the second wave is basically hit us like a tsunami about nine years ago, is when people finally figured out how they could use build neural networks that were actually very useful and very competent. And in many areas where statistical approaches are appropriate. They are fantastic. I mean, it is a revolution. Speech Recognition has very, very significantly improved image recognition, and, of course, targeted advertising. And that's what's driving this, you know, it's worth trillions of dollars, to be able to target advertising more accurately. And this is really what what's driving it. And that's why deep learning machine learning is sort of the only game in town. So that's a second wave is machine learning, deep learning, statistical approach. The third wave is basically this cognitive architecture approach is saying, What do you need to provide intelligence and to build a system that inherently has all of the components required to create an intelligent system?

Justin Grammens  17:47  
Makes a lot of sense. So we're just in the dawn here of the third wave. Yeah. And you guys as a small company, and that's usually where it starts, right? It's it's startups trying new stuff, the big guys are more focused on maybe more established markets, anything like that. But there's this third wave is really pressing forward. It's exciting. super exciting. Yeah. What does somebody do? And so what I like to ask everybody, the sound that's on the programming, like, what's a day in the life for you?

Peter Voss  18:12  
So I'm still very much involved in programming. So I program every day, I still love it. But then obviously, I've run the company as well. So you know, it's just managing teams, hiring people, focusing on where the problems are customer meetings, potential customer meetings, marketing, you know, it's across the board, I'm really involved in all aspects of the business. But I'm also very hands on,

Justin Grammens  18:36  
that's good. What are you writing in? C sharp, C sharp?

Peter Voss  18:39  
Yeah, we actually started using that 2001, when it was sort of stolen and beta or two just released. And I love the language, I think it's just a very efficient language, you know, both in terms of programming efficiency, but also in terms of running efficiency. So they're very happy with it. And now that has become available under Linux, you know, that obviously helps a lot with commercial deployment.

Justin Grammens  19:02  
For sure. Cool. Well, I you know, I'll include notes here during the show. So I'll put a link off to your website, everything like that, if you guys have careers posted at all, or reach out to you directly

Peter Voss  19:11  
reach out to us directly. From time to time, we are looking for engineers, we have a fantastic team. And we're very much at the size where everybody is basically just so thrilled about working, what we're working on, we actually are building brains, we actually have an intelligent system, and that we're not just doing another dump chatbot you know, for somebody or another advertising optimization algorithm, or another Uber for dog food. So it's great to have this kind of environment and and those are the kinds of people we're looking for is people who are really excited about what we're doing, because that's kind of the team spirit and you know how we can really shine? Yeah,

Justin Grammens  19:52  
for sure. That's awesome. As I think about conversational AI and chatbots, and just AI in general. What do you think about taking over I guess a lot of our jobs, right, so now we don't need the call centers anymore or other areas you're talking about. Do you see us? Do you see humans sort of phasing out of the picture? Or I guess maybe what do you see our role in the future of work going forward when these cognitive AI systems are put into place?

Peter Voss  20:16  
So yes, eventually, it's sort of from the marketing point of view, you often hear companies say, we don't replace humans, you know, we augment them and so on. But to be quite honest, I mean, that's basically marketing talk, companies don't like to admit that what they really want to do is reduce the number of people in their call center. Now, what has happened, in fact, is that people in call centers haven't been reduced, because more and more people are using these channels. So what we've seen over the last 20 years is in fact that, yes, the simpler tasks have been automated, and that will continue. But there hasn't actually been a reduction in call centers. I don't see that continuing, though, I think for now that, you know, that is still happening, because automation is still so limited, so damn stall. But in the longer run, of course, as you get closer and closer to human level intelligence in AI, it will replace a lot of jobs, I think it's a bit disingenuous to believe otherwise, you know, you on the one hand, have intelligent systems that are more and more intelligent, and AI is have such a lot of advantages over humans. I mean, you train one AI, and you can make a million copies of it, you know, you train one neurosurgeon, or AI researcher, or whatever, you know, work 24, seven, you know, and so on, you know, just consistent performance. So yeah, that will happen. But then, you know, when people feel that you have to look at the future of AI, as a world with radical abundance that we will have, because it will make things that we want and need in our lives, so much cheaper, and more available. So if you ask people would you like to win the lottery and, you know, be able to sit on the beach, or pursue art or travel or whatever, who would say no to that, you know, now, whether people will actually cope with it, you know, a lot of lottery winners don't do so well. But you know, that's something we will need to learn basically, how to really use our time when we don't have to work anymore, when when we can work and can do the kinds of things we want to do. So I think that's a future to look forward to. And especially if you have your little angel on your shoulder, like can help you with psychological problems with relationship problems with information, whatever.

Justin Grammens  22:37  
Yeah, for sure. I mean, I think people think about work as I have to manually sit down and do a job here for eight hours a day. Like, that's sort of been ingrained, I think, in our industrial revolution, you know, other people go to work, and they do it. And I think what you're saying maybe here is that there's going to be a shift away from that, like, it's okay to be on the beach, it's relaxing, it's okay, there's just the whole future of how you do it's okay to be more creative, a little more like flowing, because we have these cognitive AI systems that are going to take care of a lot of the brute force stuff that we used to do in the past, you know,

Peter Voss  23:07  
I think we're humans will always be wanted, isn't human relationships, you know, and we will be able to spend a lot more time figuring out how to have meaningful relationships and how to help other people have relationships to build communities and things like that. That's hard. And we're not doing that. Well, at this stage. You know,

Justin Grammens  23:28  
when you mentioned communities, yeah, I mean, I have an applied AI group, and we meet here in the Twin Cities in Minnesota, every month. But the other thing that spun off from that is this podcast. And you know, as you were saying that I think in my head, you know, what an AI be able to do this? Maybe in the future, right? Maybe I don't need to have these one on one conversations, I could actually have a conversational AI do that. But I think that would you wouldn't create as tight of knit of a community, as I say, as I think we've had currently built with these personal relationships.

Peter Voss  23:57  
Yeah, of course, I mean, while machines can already and will be able to sort of fake emotions, they can potentially be very good psychologists that can help people to have the real emotion, your interaction with another person. Because you know, it's our biology that gives us the feelings we have in our gut and our heart, wherever it is, and our embodied biological bodies. And you know, machines are not going to have that there's no reason for them to have that. Sure.

Justin Grammens  24:25  
So for people that are looking with regards to like sharing information, people that are looking to get into this field, I mean, do you have any advice, I guess, you know, what are even some skill sets that you guys are hiring for? Like, you know, kind of rewind your mind back here. When you first got into it? What are some sort of tips that you would give to people? Well, actually, let

Peter Voss  24:41  
me say something unique about our company is that two thirds of our employees are actually not engineers are not programmers. They are what we call AI psychologists, the professional I invented, so they have a training and linguistics and cognitive psychology and they basically train it Got to train our brain, their teacher to come up with a curriculum, basically teach it. Now, we also have to do some programming, we'll say that only about a third of the company where we actually do the underlying programming. So in our company, people with linguistics trained in cognitive psychology, on the engineering side, it's basic programming skills, you know. And, of course, we also have have some tools that we need to develop and you know, web development, front end, and so on, in addition to the actual brain development, that would be kind of the normal skills. In fact, we don't particularly look for people that have an AI background, because that can often be counterproductive, they have too much to unlearn. Because, you know, we have such a different approach, but more general advice. And that's from my own experience, I actually don't have any academic training at all, I'm completely self taught. So that's my own kind of bias and what has worked for me, as to me, the biggest thing is to get stuck in and you know, the sooner you can actually work with this stuff, you know, whether it's privately or you can intern at a company or somehow get, you know, actually to do this stuff. I mean, it's, it's basically, that, to me is the most important often academic credentials. And what you learn, a lot of it is just isn't really going to be very useful, you know, it's getting your hands dirty on the actual, whatever technology you want to work on, I think what's most important, and I mean, the best programmers are those that love doing it, they do it as a hobby. And not just that they learned that, you know, university, of course, you can get much better structure and you know, better techniques and learning a professional is certainly an advantage.

Justin Grammens  26:37  
Yeah, I mean, make yourself the dumbest person in the room is what is what I've always sort of done whenever I've moved to different jobs is I've always tried to find people who can help mentor me along the way. What do you do outside your professional life? You have other fun stuff you'd like to do? Or is it conversational, and AI all the time?

Peter Voss  26:54  
Well, pretty much so you know, running a startup is pretty much a full time job. So seven days a week, I am very interested in futurism life extension, and so on. So I do spend a bit of time you know, philosophy as well as ethics. I've written quite a few articles on ethics. You know, one of the things that I really wanted to understand is, what consciousness is and what free will is and things like that. So I tried to have cited a few discussion groups over the years, so like to have philosophical discussions and hang out with people who also have an interest in futurism and radical life extension. And then my other hobby, which of course fits in beautifully with life extension is motorcycles. I love riding motorcycles. So that's obviously very good for my life extension regimen.

Justin Grammens  27:41  
Oh, yeah.

Peter Voss  27:45  
I still love riding my racing bike. So

Justin Grammens  27:48  
well, good. I was trying to remember the name of the guy who there's been a lot of research recently, and I'm just completely blanking on his name. But I mean, just one of those people that says we could live to be, you know, 200 years old, you know, Ray Kurzweil or, yeah, yeah. But then there was a guy over the gray. I'm losing his name, but yeah, yeah. Yeah. Yeah, I knew he that he's written some books recently, and he's out of, I think, is out of the UK.

Peter Voss  28:16  
Is he, he's very actively working on age reversal. So rather than just trying to slow down aging is to say, his approach is how can we repair our bodies, you know, so that they actually get rejuvenated. Because otherwise we're just slowing down aging, you know, you just become more and more decrepit. Basically, you're just stretching that out. What you really want to do is you want to be able to repair and rejuvenate. And he's pretty much a leader in that field. very worthwhile looking at what he's doing.

Justin Grammens  28:48  
I just looked him up. It's David Sinclair. I said the UK, but I think he's actually out of Australia. Another. I think the English the English accent threw me off. But you've probably heard heard and read a lot of his stuff. Yes, correct. Yeah. Well, cool. I like to say I'll include notes to all the stuff here as people listen to the podcast, I can take a look at any of it. Is there anything like that I missed Peter, anything else that you'd like to people to know about? You, your company? Anything else you want to share?

Peter Voss  29:10  
Yes, of course, have a look at our company website. And, you know, we are very actively looking to implement our systems with large enterprise customers. And we have quite a few at the moment. That's what I focus in. And you know, anybody who's interested, you could look at my essays and just search for Peter Voss. And yeah, anybody who has an interest in that, feel free to contact me as well. Peter at I got a I'm easy to find on Twitter, LinkedIn, and Facebook.

Justin Grammens  29:42  
Perfect. Peter. Well, I appreciate the time. And man, you shared a lot with us. Obviously, you're a leader in the field. And I know you'll continue to do some great things here. As we get into the third wave. How far do you think this wave is gonna go, you know, as we're winding down here, how far the third wave going to go until we hit the fourth wave you. Thank you. We have this this third wave here is gonna last for another couple decades.

Peter Voss  30:03  
No, I think the third wave is it basically. Because once you have the right architecture, and we finding this, you know, not having worked on this for, you know, 20 odd years, is we just need to improve what we're doing. I don't have any kind of sense that we fundamentally need to use a different approach, you know, like quantum computing or whatever. So I think the third wave, is it, it's just different people ask me how long before we get to human level intelligence? I mean, we're still a long way away from that. I don't measure it in time are measured in dollars. I mean, yes, it will take a certain amount of time, but I think we could have human level intelligence in less than 10 years, if enough effort was put into it. So yeah, that's sort of where I see that I think we are on the road to having human level intelligence. It's teaching the system, all of the common knowledge that we common sense knowledge that we have, that's really hard. You know, that's one of the big challenges that we just learned such a lot growing up in the real world, you know, our childhood and just the interaction to other people with the real world. We just learn a lot. And getting that knowledge, that practical, common sense knowledge into an AI, you know, it's a very difficult task, and that's basically part of what we're working on.

Justin Grammens  31:21  
Sure, but you still believe it's possible. Oh, absolutely. And your quote about you know, I read this actually off of your off your LinkedIn profile. It was a quote by Oren Harare. I think it says it's electric lightbulbs, and that come from the continuous improvement of the candle? Correct. So I think the thinking behind that is, is we got to do something completely new. I mean, it's probably feels like the dollar thing you're talking about, you know, Edison wasn't just improving the candle, he was taking it to the new level.

Peter Voss  31:46  
Correct. And, you know, as you mentioned, the big companies are not likely to have that kind of innovation. You know, it just doesn't happen that way, like big oil tankers, very difficult to turn them all over the top leadership and the big companies, ai companies, they are experts in deep learning, machine learning statistical approaches that say, expertise. That's what they're looking to improve. And, you know, it's new startups that basically change the paradigm, who would have thought that tiny little Amazon startup could ever hope to compete against Barnes and Noble, you know, or tiny little startup Facebook could ever hope to compete against Google, you know, or Google themselves. When women are startup, you have to really think very differently and have to have a different approach. And big companies just have too much baggage and too much invested and too much to harvest their existing investment. You know, that's a focus,

Justin Grammens  32:38  
for sure. Well, wise thoughts, Peter, I appreciate your time. Again, thank you so much for being on the show, and we look forward to talking to you again in the future. Thanks for having me.

AI Announcer  32:49  
You've listened to another episode of the conversations on applied AI podcast. We hope you are eager to learn more about applying artificial intelligence and deep learning within your organization. You can visit us at applied AI dot amen to keep up to date on our events and connect with our amazing community. Please don't hesitate to reach out to Justin at applied if you're interested in participating in a future episode. Thank you for listening

Transcribed by