Artificial General Intelligence (AGI): What’s Next for AI? with Peter Voss
26:48 Share

One of our big customers is 1-800-Flowers. They use our system for customer support. The system has all of the knowledge at its fingertips, but it also knows what you said earlier in the conversation. It remembers what you said in previous conversations. For example, two months ago, Valentine's Day, we replaced 3,000 operators that they normally needed to hire just for a few days at that spike, which doesn't help anybody because how well can you train people?

if they're going to be working for less than a week. So it's a win for everybody. There's zero wait time and the system knows about your order, where it is. It can change your order and solve all sorts of issues that may crop up. Welcome to Authentic AI for Entrepreneurs, the podcast that shows you how to leverage the power of AI technology without wasting your time or selling your soul. Let's embrace making AI work for you.

Hey, hey, my human friend, and welcome to another episode of Authentic AI for Entrepreneurs. I'm your host, Kinsey, and I am really excited to share today's conversation with you. We are chatting with Peter Voss, and he is one of the pioneers of a term and concept called artificial intelligence.

Artificial General Intelligence. And basically, if you have been wondering what the future of AI models like ChatGPT could look like, then this is an episode for you. Artificial General Intelligence is really taking the AI tools that we know today, such as ChatGPT, large language models that have been trained by humans on a lot of data, artificial

artificial general intelligence or AGI is taking this kind of concept to the next level and honestly to just bringing it kind of back to what we all think AI is in the first place, right? Which is a technology that can think for itself that has lots of higher level learning and cognitive skills and

Peter will explain this concept way better in our interview, but it was a really, really interesting conversation. And honestly, too, if you have been feeling at all hesitant about what the future of AI could mean for our culture, our society, or even just our work and our jobs,

Peter has some really, really interesting takes on that as well that I think we all need to hear. So I hope you think that this conversation is as interesting as I did. If you do enjoy it, don't forget to share it with another friend who could also use to listen to it. And while you're at it, why not leave us five stars or a friendly review on whatever podcast platform you're listening to us on? Because it really does help us to continue spreading our authentic AI message and growing our authentic AI community. All right, let's go ahead and dive right in.

Hi, Peter. Thank you so much for coming on Authentic AI for Entrepreneurs. I'm so excited to bring you on the show. Well, thank you for having me. Yes. And I did already introduce you just a little bit in the intro, but I'd love to hear from you. Who are you? Tell us a little bit about yourself. Yes. So I'm CEO and Chief Scientist of iGo.ai. And I've really been working on artificial general intelligence for the last 20 years or so. And that's my big passion, I think, is

will be able to really help human flourishing and create radical abundance. So I'm very positive about this future. Oh, man. Yes, I am so excited to chat with you about this because we haven't really talked a lot about AGI yet on this podcast. But before we dive into this,

that, I was to ask people, how did you get started even dipping your toes into the world of AI? What did that look like for you? Yeah. So I started out as an electronics engineer, then working on hardware. And then I fell in love with software, basically, when we were able to start programming chips. And so my company turned from a hardware company into a software company. I developed a comprehensive accounting package for small businesses.

The company did really well. We went from the garage to 400 people and did an IPO. So that was really exciting. And when I exited that company, it was what big thing do I want to tackle? I have a lot of time on my hands to really study and so on. The obvious thing for me was how can we build smart software? How can we build software that can think and learn and reason and have common sense? Basically, ensure that has human level intelligence.

So I embarked on a five-year tour of studying all different aspects of intelligence from epistemology, theory of knowledge and philosophy. How do we know anything? How can we be certain of things? What is reality? What is free will and consciousness? And to really understand that.

human cognition. Also, I studied cognitive psychology in terms of how children learn, how our intelligence differs from animal intelligence, what IQ tests measure, and a lot of other things like that. And then, of course, I studied what the work that had been done in the field of AI. A combination of that was that I

came to the conclusion that the original idea of AI had been lost. The original idea of AI, and the term was coined 69 years ago, actually, and the original idea was to build thinking machines that can think, learn, and reason like humans. And in fact, they thought they could do this in about two or three years. They could crack this problem. Of course, it turned out to be much, much harder.

So what happened over the decades is that AI turned into narrow AI, where you are just solving one particular problem at a time. And there's a huge problem with that. And that is you're not actually building a system that has intelligence. You are using the programmer's intelligence or the data scientist's intelligence to write software that can solve that particular problem in biomechanics.

got together with a few other people who also had the same idea of returning back to the original dream of AI to build thinking machines to get away from narrow AI. And so in 2002, we coined the term artificial general intelligence, three of us, and we wrote a book on the topic. And that's what I've been dedicated to for the last 20 years, alternating between development, R&D, development,

And commercialization, as we develop the technology further, we've been able to commercialize it successfully. But then we also want to go back and bring it closer and closer to human level intelligence. Yeah. Oh, man. For the listeners who aren't really understanding exactly what AGI is.

means or artificial general intelligence means. Can you explain to us how that's different than the tools we've been using now? Yes. So artificial general intelligence really is to get to full human level intelligence so that you can have AI. For example, you can have an AI that trains itself. It can learn itself like a human can to become a cancer researcher, say. So you could now have an AI that's a PhD level cancer researcher and you can make a million copies of that.

You now have a million PhD-level cancer researchers chipping away at the problem, communicating with each other, sharing information, and so on, and having a higher degree of rationality than humans. We are not all that rational, actually. And so we will make tremendous progress in this. But then that same AI technology can also be reduced to dramatically bring down the cost of goods and services. Another example is where we could each have our own personal, what I call a personal-personal relationship.

personal assistant. And the reason I use personal three times, there's three different meanings of personal. The first personal is you have a personal assistant that you own. It serves your agenda and not some mega corporation's agenda. It's hyper-personalized to you knowing your history, your goals, your friends, and so on. And then the third personal is the issue of privacy that you control what it shares with whom. So those are the kinds of things that are possible with AGI that I

Not at all possible with ChatGPT and large language models. So as impressive as the current technology is in terms of writing poems or summarizing things, the big problem is, A, they hallucinate, they make up stuff. They are trained with 10 trillion pieces of information, good, bad, and ugly information.

And they don't know what's good. They don't know what's bad and ugly. So they don't know what's right or wrong. They're not aware of their own process of their own thinking. They don't have metacognition. And most importantly is they cannot learn anything in real time. Now, as we're having this conversation, we are learning from each other and we are

updating our model. We are updating what we know. You hear something and you think, does it make sense to understand this? Does it contradict something I know? What else does it make me think of? So we have this very active process of learning in real time, and large language models cannot do that at all. So they really, in fact, one of the luminaries in the field of AI, Jan Kuhn, he's the chief scientist at Meta Facebook, and he said large language models are an

off-ramp to AGI. They are a distraction, basically a waste of time to get to AGI. So the technology we have right now, as impressive as it is, being able to create short videos and so on, they're simply not reliable and they cannot learn in real time. They can't think properly. They don't know what they're doing. The one that we've been working on called cognitive AI, where the system can learn the way a child or the way we learn interactively. Yeah.

Oh, man. Yes. The next level of AI. I love it. OK, I will say just to play the devil's advocate here, the AI scaries. I can feel the flares coming up for people when it comes to this, because we hear about this AGI concept more so. And of course, it's what we've all grown up thinking like AI is and it.

very much more closely linked to that Terminator future or for my Disney fans, the smart house scary lady. But can you please speak to any kind of concerns people have? I can think of two, especially that my audience might have, which is the scary Terminator future.

Or also just like the fear of AI tools may be way more likely to take our jobs, which is definitely a concern in entrepreneurship and stuff. Can you speak to those? Yes, certainly. The one existential risk is basically that AI will take over. Now, unfortunately, certainly in the Western world, Japan is actually a bit different. The Japanese tend to love robots.

But in the Western world, it's Hollywood and the AI is always the bad guy and wants to kill everybody. And it's unfortunate because there's really no basis for believing that. And there's another problem that we have a whole industry of AI risk organizations that sort of feed off that fear. And they basically continue to perpetuate the idea that AI is going to kill us all.

unless we can fix it and give us millions of dollars and we'll work on fixing it. And so it's like a shakedown almost. It's unfortunate that you have this whole industry of creating fear that is really not grounded in reality when we're

We build AI, AGI, we build it specifically to help us, to do things for us. It's not going to have that inherent drive to protect and reproduce like our reptile brain. We, as we are still animals, we have that inherent survival and reproduction instinct drive that we have obviously for evolution to work. But in AI, there's no reason to build that in. AI are not going to have that inherent drive to reproduce.

dominate or do their own thing. Now, it's a

It's a complex argument and it's not, one can't just dismiss the risks and dangers, but a lot of arguments are completely unfounded. Now, as far as your second scenario is concerned, saying that it'll take our jobs, don't a lot of people want their jobs taken? Doesn't everybody want to win the lottery that they don't have to work anymore? I think that's really the perspective one needs to have is once we have AGI, it will bring down the cost of goods and services dramatically.

tremendously. It'll really create this radical abundance that work will be optional. We'll be able to do the things we want to do. And there'll be a lot of things that are much more interesting than the work that most people do. Most people, if they could afford it, they wouldn't actually do the job that they're doing. For example, to help uplift the rest of the world, once AGI becomes available, a lot of people would want to do that. Now, you may want to spend more time with family, raising a family or doing art,

creating things and so on. One really needs, I think, to see that as a positive that we will not have to work nearly as much. Now, of course, there'll have to be some kind of a universal income somewhere to do that. But with the abundance created by AGI, I don't think that'll be a problem.

Yeah. Oh, wow. So interesting. I love kind of these shifts in perspective that you're offering us. Oh, my gosh. I was chuckling as well when you were talking about how AI isn't inherently human. How it came up for me, how like humans can mirror. We mirror ourselves in these things.

human-like robots. Correct. It's so funny, exactly. Like, they aren't programmed. They aren't inherently human. Can you please share a little bit about what AGI looks

looks like or like what kind of applications and what you guys are working toward for the near future? Yes, certainly. So we don't have AGI yet. The definition of AGI is to have human level ability to think, learn and reason to learn like humans. So we don't have that yet. But the technology that we've been developing, which is not big data, it's not statistical, it's not generative AI.

It's cognitive AI. So the starting point is not how much data do I have? How much computing power do I have? All the large companies, Google and Amazon and Facebook and so on, they have a lot of data. They have a lot of computing power. So in a way, it's just as a hammer. We've got what nails can we find? So the approach we've taken is the starting point is what is important in intelligence? What is important in human intelligence?

And as I mentioned, it's the ability to learn interactively, to deeply understand what you're saying, what you're doing and what you're reading and so on. So we've been developing this core engine for the last 20 years and we are very successfully commercializing it.

proto-AGI. So we're not at the level of humans, but the architecture is the correct one. One of our big customers is 1-800-Flowers group of companies. That's Harry and David and Popcorn Factory. And it's about 12 companies. And for example, they use our system for customer support. I have a very good experience with that because the system has all of the knowledge at its fingertips, but it also knows what you said earlier in the conversation. It remembers what you said in previous conversations.

And for example, two months ago, Valentine's Day, we replaced 3,000 operators that they normally needed to hire just for a few days at that spike, which doesn't help anybody because how well can you train people?

If they're going to be working for like less than a week, you know, on that. So it's a win for everybody. There's zero wait time and the system knows exactly about your order, where it is. It can change your order and solve all sorts of issues that may crop up. So that's how our technology is currently deployed commercially. We can also use it for marketing.

say, a personal coach to help people manage diabetes as a personal assistant to do that or as a sales assistant. The core engine lends itself to be deployed in anything where you have a conversation. It's just that we're not at human level intelligent. So there's still a lot of effort that has to go in to teach the system how to handle different situations. And once we get to AGI, the system will be able to learn that by itself.

Do you want to create more Instagram Reels but get stuck on what to post? I hear you. That's why I'm excited to share our new free resource, the Reels Prompt Playbook for ChatGPT. This playbook includes 20 ready-to-use AI prompts for Reels hooks and captions.

and makes it super easy to create trendy content that grabs attention without spending endless hours scrolling Instagram for inspiration. Whether you want to share behind the scenes, offer quick tips, or tap into the latest trends, The Reels Prompt Playbook has you covered.

Download your free Reels prompt playbook at diywith.ai slash reels. And the best part? These ChatGPT prompts are designed to work for entrepreneurs in any industry. Plus, you can pair these prompts with any trending audio or any video that you find on your phone. Basically, they're super easy, super quick, and will keep your content super authentic while saving you loads of time and headache. Grab your free playbook now at

DIY with dot AI slash reels and start showing up on Instagram consistently today.

Yeah. Okay. This is making sense. And correct me if I'm wrong, but your system is using like a voice cloning or an automated voice software within its model? Yeah. So we do both chat and also phone. So we have voice and text. In fact, it's omni-channel. You could actually switch from text to voice and back if you want. Yes, we don't really use voice cloning, but we use generated voices because it gives you a lot of flexibility. But

My philosophy and my company's philosophy is we never want to fool anybody that it's not an AI. So on our chat, the chat gets introduced. This is your AI assistant or so. So we think it's a mistake if companies...

try to portray themselves as human when they are not. I like that a lot because I think that's where some rub comes in sometimes too with any kind of AI tool really is the transparency and people feeling like there's some shadiness here, but there's nothing wrong with it and let's just be transparent about it and it's okay and it's becoming normal. I think it's the lack of honesty on the one hand

But it's also from a practical point of view. If you know you're talking to an AI, you're not going to be saying, how are you feeling today? Did you hear about the ballgame last night? Or how's the weather where you are? You just want to get things done efficiently. And that's fine. If you want to talk to a human, you'd call up your friends. Yeah, the nature of the conversation is different. And actually, that leads me into something else I wanted to ask you, because we were talking about cognitive AI. And I'm wondering how

How much emotional intelligence is playing into the systems that you're creating and stuff too? It's a very good question because it comes back to that it's not human. It doesn't inherently have the visceral emotions of love and hatred, guilt and jealousy and things like that. There's no reason to have them. The system can recognize these emotions in a person like a good psychologist when you're talking to a psychologist.

You may get very aggravated, but it doesn't mean the psychologist should get aggravated. But they need to recognize it and know how to respond appropriately. I think an AGI will have a very good EQ, very good emotional intelligence, but it won't be feeling those emotions. They won't be from personal experience. They will be from learning on how to recognize them and how to deal with them.

Yeah, very interesting. I know I'm really intrigued with how it goes back to how possibly the way we think about robots taking over the world. We're mirroring our human selves in these robots that are being created. And I'm interested to see how we adapt culturally to nature.

new AI technology as mirroring our own emotions in these tools too and whatnot. Yes, it'll vary a lot. Some people just embrace the technology and welcome it. And there'll be others that are like kicking and screaming and saying, this is really useful. But I had an interesting conversation at a

conference where I gave a talk recently and somebody came up to me and said the call center automation he says but boomers want to talk to a person and I looked at the person and said do you prefer to talk to a person or do you just want an AI to do it he said oh no I want an AI well you're a boomer so this is

perception that boomers want to talk to a human, but that's largely not true. Boomers also have had smartphones now for many years, and they've got very used to texting and getting stuff done, you know, sort of in that different mode of communication. People really typically, when they deal with a company, typically just want to get stuff done. That's the important thing. You don't actually want to talk to a human that doesn't know what they're talking about, or you can't understand their accent.

and then you're in the wrong department, you have to be transferred, and you have another 10-minute wait time, and then you have to start from the beginning explaining everything again to listen to. Your business is important to us. Please listen carefully because our options may have changed. You don't need that in your life. So I think people do adapt to technology if it works well.

And on the other hand, with AGI, it will have that, as we've said, it'll have that EQ to help people feel more comfortable with the changes that are happening. A lot of your personality and who you are and your self-esteem is tied up with your job and the position you have in a job. And suddenly an AI can actually do this job better than you can, help you transition to find things in life that are meaningful to you and that you'd really want to do if you didn't have to do the job in the first place. And AI can help us with that.

personal assistant. It's like a little angel on our shoulder that can help us think things through and motivate us. Yeah, it all comes down to like the story you're telling yourself in the perspective, right? People are saying, I just want to talk to a human because of the frustrating experience that automates systems that aren't really great, but it's not

the human versus robot. It's exactly what you said. It's efficiency and the ability to get the thing done as quickly as possible. And yeah, I always compare. I talk a lot about training ChatGPT on your brand so that it can create content for you and think and sound like you. I always compare it to the high school intern that probably just got back from like smoking weed outside or something

thing and he's like and like he would not do a good job either you know versus the fully trained eager intern who understood everything about your business and stuff and it's just thinking about it in different ways but you know unfortunately most

companies, most big companies, all the utilities and banks and that, they still have awful systems. They have systems that really have 30-year-old technology. It's just a decision tree, a flow chart. Ask you this and then it's yes or no or say this or that. Anything that doesn't fit in the flow, then it gets lost and the people just zero, zero, get me to an operator. It's really a pity that companies don't pay more attention to having better systems because the technology exists

Once they become ubiquitous with AGI, where systems will inherently be smart, understand you properly and so on, people say, this is much easier than waiting online to talk to a human who probably doesn't know how to help me anyway. Okay. I would love to hear your thoughts on the future of AI. Like, what hopes do you have as far as like AGI in your company or just how do you see...

things playing out. Right. ChatGPT, I think, has given us a taste of AGI, just its language ability that it can write a poem and summarize things and it can do quite amazing things. It just doesn't do them reliably, robustly. It doesn't know what it's doing and it can't learn interactively. But it's given us a taste of AGI. And we believe with our technology, with cognitive AI, we are just a few years away from true human-level

AGI, full AGI. So I see this coming soon. At the moment, all the focus is still on big data, statistical system, generative AI. It has a lot of momentum, a lot of money going into it, a lot of people making money, losing money and so on. But there's also an increased realization now. Like I said, Jan de Koon,

And a few others actually have recently been quite vocal about saying large language models, statistical AI is not going to get us to AGI. They just haven't come around to the view that it actually requires cognitive AI. They haven't made that identification. So I see that will be with us soon. And if we can do anything in our company, which we're trying to do at the moment to grow our team so that we can accelerate the development.

I think it'll be tremendous in helping human flourishing and creating radical abundance and having much more intelligence available to help ourselves problems like all the diseases, like energy, building much better batteries, solving pollution problems, helping us with governance, then that personal assistant to help each individual optimize their lives. So I see a very positive future and I can't wait for it to happen. Okay.

Wonderful. Thank you so much for this conversation. Can you please share how anyone might be able to connect with you if they want to? Yeah, that's very easy. You can just find me easily on LinkedIn, Twitter or our website, peter.aigo.ai. Or as I say, you can find me on social media.

and be happy to engage people, anybody who wants to help make AGI happen sooner rather than later. And for us to get the benefits, those are the kind of people I'm looking for. Okay, great. Thank you so much. We'll link everything in the show notes as well. And I appreciate this conversation. It was really great. Great. Well, thanks. Thanks for the good questions. Thank you.

How awesome was Peter, you guys? I was really, really intrigued throughout that entire conversation. I'm so glad that he was able to explain AGI to us in a way that makes sense, in a way that really does kind of quell any hesitations or fears that, you know, might be popping up for you.

when it comes to all of this technology moving so fast in this world. But yeah, I really cannot wait to see what Peter and his team bring to the world of AI technology. This is all very exciting. This is all happening right now, guys. And just the fact that you listen to this podcast and

you understand just a little bit better what the future of AI could look like for us is a really great step for us all being just involved in the conversation, right? So don't forget to share this episode with another friend who could also use to listen to it and leave us five stars or a friendly review on whatever podcast platform you're listening to us on because it really does help us to continue spreading our authentic AI message.

All right. I hope you have a wonderful rest of your week and I'll catch you next time I catch you. Thank you so much for tuning in to Authentic AI for Entrepreneurs, my friend. If you enjoyed this episode, don't forget to subscribe on Apple Podcasts, Spotify, or wherever you listen to your shows.