Everyday AI Podcast – An AI and ChatGPT Podcast - Ep 564: Dr. Ben Goertzel: The Road to Creating Benevolent Decentralized AGI

Episode Date: July 10, 2025

What's at stake for humanity amid the arms race to AGI? Dr. Ben Goertzel should know. He legit coined the term AGI. 🤖The legendary AI leader is sounding the alarm: leaving Artificial General Intelligence in corporate or military hands could plunge humanity into chaos. Ben joins Everyday AI to reveal the high-stakes gamble we're making with AGI, why decentralized control is our last shot at sanity, and how our actions right now could mean the difference between utopia and disaster.No biggie, right? lolz. Join us to find out: ↳ Why Big Tech AGI could lead to global instability↳ How Decentralized AI as the ONLY path away from catastrophe↳ How you can shape AI's explosive trajectory—todayStop scrolling. Listen, or regret it later.Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion:Thoughts on this? Join the convo and connect with other AI leaders on LinkedIn.Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:Decentralized AGI Development ChallengesEnsuring AGI Benefits SocietyDefinition Evolution of Artificial General IntelligenceHistorical Context of AGI DevelopmentHuman-Level AGI vs SuperintelligenceAGI Impact on Global GeopoliticsEthical Concerns with AGI DevelopmentAGI's Role in Education & LongevityTimestamps:00:00 "Ensuring AGI Benefits Everyone"03:23 Rethinking AI for Human Benefit06:35 "Decentralized AI through Internet Protocols"10:52 Defining AGI's Mathematical Vagueness14:29 "Humans: The Minimum Intelligence Threshold"17:56 "Achieving AI ROI Made Simple"22:11 "Path to Benevolent Superintelligence"25:50 AGI Arms Race: US vs. China27:24 AGI Development and Global Power Dynamics32:07 AI Enhancing Education for Dyslexics34:23 AI Tools: Dual-Use for Data Analysis37:34 "Empowering Participation in AI Future"Keywords:Artificial General Intelligence, AGI, decentralized AI, benevolent decentralized AGI, SingularityNET, Ben Goertzel, human-level AGI, superintelligence, AI for human benefit, AI ethics, AI arms race, AI development, AI systems, AI and geopolitics, AI and military use, AI in medicine, AI in education, AI and intelligence explosion, time to fume, LLMs, large language models, Send Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Start Here ▶️Not sure where to start when it comes to AI? Start with our Start Here Series. You can listen to the first drop -- Episode 691 -- or get free access to our Inner Cricle community and all episodes: StartHereSeries.com Also, here's a link to the entire series on a Spotify playlist. 

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. Meet Firefly AI Assistant, now live and Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. How can we ensure that artificial general intelligence benefits everyone?

There's been obviously a lot of talk about AGI and what that means and all of these big companies working toward it. But should one company control AGI? And what about what it means for the rest of us, right? The good that it can do in the world. Well, we're going to be answering those questions and a lot more on today's edition of Everyday AI. One, I'm very excited about because if you follow AI at all, you definitely know our guest

for today, and it's going to be a great one. So, thanks for tuning in. If you're new here, welcome to Everyday AI. This is your daily live stream podcast and free daily newsletter, helping everyday business leaders like you and me, not just learn what's happening in the world of AI, but how we can leverage it to grow our companies and our careers. If that's what you're trying to do, make sure you go to our website at your everyday a.com. We're going to be highlighting the best takeaways from today's episode in our free daily newsletter.

And I can already guarantee you there's going to be a ton of them. So make sure you do that. And if you want the daily news, that's going to be in there as well. All right, let's have our guest for today. I'm excited in live stream audience. He needs no introduction, but I'm still going to do it. So please help me welcome to the show, Dr. Ben Gertzell, see. CEO of Singularity Net, Dr. Ben, thank you so much for joining the Everyday AI show.

Hey, thanks, Jordan. Good to be here. All right. And yeah, like to have the person who literally coined the term AGI on the show is a pleasure for me and I'm excited to get into it. But maybe let's start at the end, Ben, and then we can rewind our way through it. How do we make sure that AGI benefits everyone and not just, you know, one big tech company that may or may not discover it. Yeah, that's obviously an important question and kind of question some of us have been thinking about

for decades, although it's now rising to greater and broader prominence. I mean, first of all, I'd say we don't insure or guarantee anything, and that's just the way it is, right? Like when the cavemen developed agriculture, they couldn't ensure how that was going to go either. And the jury in a way is still out on whether we're happier than we were in the Stone Age, right? So, I mean, and we didn't have guarantees when we launched the Industrial Revolution or the computer revolution and whatnot.

So, I mean, I think we are taking leaps into the unknown here without a guarantee, and that is what our species has done. all along. That said, common sensically, there seem to be some things we can do to bias the odds in a positive direction as regards the launch of advanced AI systems. I mean, one of them might be to start using AI for broader human benefit right now instead of using it so much for killing people, spying on people, selling themselves they don't need, running automated trading bots to extract money from retail investors into the bank accounts of large investment banks, say plagiarizing people's creative works and not giving them any compensation.

A start might be to take the AI systems we have now and use them for broader benefit, and that might set the initial condition for the emergence. of general intelligences and superintelligences that are oriented toward broad human benefit, right? So the thing is there's nobody sitting there in their office in a big tower overlooking the world, thinking like, this is how we should be developing AI to maximize the odds of a beneficial singularity, right? Rather, what happens is AI is bubbling up and moving towards. general intelligence from the incredible mess of the world, the world economy that we all all see around this. Like for better or for worse, the AI that we're creating, you know,

it treats us the way we treat each other and it's reflecting the whole chaotic mess of our, of our society. And that's kind of what we're doing. As we stir the mix of the internet and see what AGI bubbles have out of it. And Ben, for more casual people in AI or people who just started, you know, you've been talking about AGI longer than anyone else, right? You coined the term.

You wrote your first piece of AI code in the mid-90s. So I'm curious, how has the definition of artificial general intelligence, how has it changed throughout the decades? What does that mean? I think I wrote my first AI code in 1980. Oh, in the 80s. I was off by a decade.

That's amazing. Yeah, I mean, I learned about AI in the early 70s, when I was a young child. And I learned about concepts similar to what we would now call the singularity from a book called the Prometheus Project by the Princeton physicist Richard Feinberg. I think I read that book in 73. It had been published in, I don't know, 60 years. or something, right?

So these ideas have been around a while. I bought a little Atari computer in 80 or something and then taught myself to code and start trying to code AI. It had 16K of RAMs, so you couldn't do too much, right? Then when the Internet came out and the web became a thing in 93, 94, 95, I started to think about decentralized AI because the Internet itself is decentralized, right? It's protocols, not platforms.

So it seemed clear that with the Internet, you had the ability to roll out AI where little bits and pieces of the thought process were in different places around the Internet and the intelligence sort of emerged out of all that. So while I started on AI in the early 80s, decentralized AI, you can think about that earlier, but once you had the web there to play with,

it was like, well, we can see how you would do this, right? I mean, you had actual protocols you could use to decentralize AI agents living in different sections of the internet. I remember I was at a workshop on agent systems and agent-based AI somewhere in Australia in 96 or so, And I was giving a pitch on the internet as this sort of incubation ground for networks of autonomous AI agents. And now the last couple years, agents have finally become a big thing. But I mean, the idea, ideas have been there a long time.

It's just the computer hardware and the networking speed weren't there, right? And the same is true with what we would now call AGI, right? So, I mean, when I first read about AI in the early 70s, I was thinking about AI as you can think like people and ultimately be smarter than people. And in 2002 or so, I was trying to edit an academic book of papers on how to build real thinking machines

rather than the more narrowly specialized problem-specific AIs, which the AI field had sort of drifted into over the 70s, 80s, and 90s. And I was going to call the book Real, AI, but I started to feel like that was a little bit too much of a sort of poke at people building neural applications, specific AIs, which are real.

They're doing real things. They're sometimes cases good things. So yeah, in the search for a better title for the book, myself and a bunch of the other chapter authors for the book, for the Brutup AGI, artificial general and

Actually, at first it was going to be GAI, General Artificial Intelligence, which was suggested by Pei Wang, a Chinese AI researcher, because in Chinese it goes that way, general AI. But, yeah, that was gay, which we thought was not. I mean, while I'm a big supporter of the queer universe, it seemed like that wasn't the best acronym, right? So we went with the AGI instead. I later found someone who had used that in an essence,

say in like 97 in an essay on Nometect. We weren't the first ones to use that term in our book or anything, but I think that book is what put it on the map. Then we had an AGI workshop from 2006, the first annual AGI conference in 2008, and we're doing the next one in Iceland, University of Rejavik, in mid-Aug. So we've been sort of pushing forth that term and that mean.

And what I meant by it from the beginning was AI systems that could generalize beyond their training and programming, at least at the vague level that people can and ultimately more so. And since that time, and even in that initial book on AGI, which ended up being published in 2005, there's been attempts to mathematically formalize what do we mean by AGI. And that one's interesting to me, right? Because I even, as I'm trying to learn more about AI, I go back and look at definitions from, you know, 20 years ago, 10 years ago, and it seems like it's always changing. So from the one that's probably been talking about it more than anyone else for the last few decades, when will you say, oh, yes, we've achieved AI, right?

Like, is there a mathematical? Let me define terms a little better because, I mean, AGI is, that's a natural language term. It's a vague term in the same sense that beauty, intelligence, or life, or health, or vague terms, right? And so if you try to mathematically nail down what is AGI, you come up with something that's fuzzy and graded rather than either or. So an example of the sort of mathematical definition you come up with

is, say, the ability to achieve arbitrary, computable reward functions in arbitrary computable environment. So that basically means you average over all things that a system might want to do in all possible environments. And you're like, on average, how good is this system at learning to do random things in random worlds, right? And that's, you can formalize that using a bunch of nice math, and you can vary it in different ways.

Now, what you conclude, first of all, all from this sort of very abstract mathematical conceptualization of AGI, first of all, you conclude like there is no dividing line, right? Like one system will be more generally intelligent than other, they'll be more generally intelligent than another. Secondly, you conclude humans are not very generally intelligent, right? I mean, in the grand scheme of things, like, I'm not very good at achieving a random computable goal in a random computable environment. Like I couldn't even run a maze in 750 dimensions, right? I mean, so I'm quite restricted in this broad scope of things.

Now, probably a worm is even more so, and a rock is even more so, right? But then you conclude, like, the precise level of human general intelligence, our ability to achieve random goals and random environments to a certain level is it's kind of like looking at the running speed of humans and defining that as general running ability or something or general racing ability. Because I mean, we run at a certain speed. It's faster than some species.

It's slower than other species. Well, machines that can run way faster than us. There's nothing that magic about the speed that Yusain Bolt can run at in the scope of the speed of moving in the whole, universe, right? Human level AGI to me is a little bit like that. I mean, there's a certain level at which humans can generalize beyond their programming and their training. And it's not a super magic level. It's just where we happen to be. Like, it's important to us, right? Of course, just like being able to run faster than all other people

was important to Usain Bow, would let him win a prize, and it would let him out around other bad guys who want to clobber him unless they're holding weapons or something, right? But in the scheme of the universe, like human level AGI or human level running speed or not necessarily that important. But we could look at it a little differently,

like escape velocity is important on Earth, right? So it's not important in the whole cosmos, but I mean, it's the speed that lets you get off the planet. So you could say if humans are barely intelligent enough to build superhuman AI, then maybe while on the whole, our intelligence level is a bit arbitrary. Like in that sense, we've reached a critical threshold. Like given the materials available on our planet, maybe we've reached the minimum general intelligence level needed to build something smarter than ourselves. So my friend Jim Rut is like a master entrepreneur.

He was the CTO of Thompson Reuters and so forth. He liked to say humans are about the minimum possible general intelligence. Like we're about as dumb as you could possibly be and still invent computer science and still figure out how to make super, super AI, right? I think it's kind of dependent on the resources in hand. Like you could have some planets where it could be dumber than humans about the super AI somewhere you'd have to be smarter. But this is all, it's a long-winded way of saying that when people talk about AGI, they're usually thinking about human-level AGI. So they're thinking about something that

can like generalize, take imaginative leaps beyond its experience. roughly as well as people can. And that's important to us, of course. It's important economically, but it's sort of almost arbitrary from a computer science and AI view. And the other way that the notion of AGI has become confused in recent years is,

like Sam Altlin says, wait, we already achieved AGI, right? And indeed, GBD 4.5 arguably past the so-called Turing test, meaning it can fool humans into thinking it's human in a brief conversation. The thing is, LLMs achieve a great amount of generality in their discourse, but not by being able to generalize very well. They generalize mediocrely.

I mean, it's impressive, but it has clear limits. As Apple researchers demonstrated in a paper that came out a week or two ago, which has gotten a bunch of discussions. The thing is, LLM's achieve a great deal of breadth because their training data is the whole fucking web, right? They just need to leap a little bit beyond their training data in order to do a lot because their training data is so much. So that is a kind of generality,

but it's different than having human-level generalization ability, which is what... So I think AGI is a very broad thing. human level AGI is a more particular thing, but what it should mean is the ability to generalize as well as humans can rather than just to do a broad scope of stuff. And being able to do a broad scope of stuff certainly is an important ingredient.

And I think it's part of what can help us get to systems that can generalize really well, but it doesn't get you there on its own. Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite into one conversational experience. Meet Firefly AI Assistant, now live in the Adobe Firefly app, the all-in-one creative AI studio. Powered by Adobe's Creative Agent, Firefly AI Assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the Assistant. The Assistant orchestrates multi-step workflows, drawing on 60-plus program. grade tools across Adobe Creative Cloud apps, including Photoshop, Illustrator Premiere, Lightroom Express, and more to help bring your ideas to life. You can also get started with creative

skills, a growing library of pre-built workflows for common creative tasks like batch editing photos, creating mood boards, portrait retouching, and creating social variations. Every step the assistant takes is visible so you can refine, redirect, or take over at any time. You stay in the driver's seat as the creative director. Adobe Firefly AI assistant now in public beta. See it today at firefly.adopi.com. So all of the big companies right now are spending countless billions of dollars, right, openly, you know, trying to achieve AGI as, as, you know, those goalposts may be ever moving. What does that mean, right? Is it a bad thing if one company is, the first to quote unquote, you know, achieve that level of intelligence that you just spoke of?

And if so, why is it important to have a certain person? There's a few different questions here. And there's a lot of unknowns, right? So one question, Eliasur Yudkowski and Robin Hansen in the futurist world have posed is sort of the time to fume. And what they mean there is, like once you get a human-level general intelligence, potentially it could be a quite short period until you have a super-intelligence.

Because a human-level AGI, you know, if it's as smart as an AGI researcher, it's also got the ability to rewrite its own code and to copy itself and experiment on those copies, right? So it seems like potentially human-level AGI could be followed up not too long after by slightly greater than human-level AGI, by a little more greater than human-level AGI, bing, bing-bing-bing, bing-bing, bing, bing, until you've got super-intelligent.

So how fast is that bing-bing-bing-bing-bing-bing? Like, is it a month? Is it five years? Yeah. Right? So if it's a month, then it may matter a lot who gets to AGI first, right? If it's five years or, say, 16 years, like my friend Ray Kurzweil thought, he thought we would get human level AGI in 2029 and superintelligence in 2045, right?

If it's five or 16 years between human level AGI and superintelligence, it's less clear how much it matters who gets there first because there's not that big a moat for anyone like even if tennscent or google gets their first within a couple of years everyone's going to be there and then you've got this whole population of aGIs you may have other problems like an aGI arms race or something right there's an a gei world war three but but you're not it's not going to matter that much in terms of super intelligence who got to aGI first but there's an unknown there of how long it will take for that intelligence explosion to enact itself.

And in the grand scope of history, whether it's a month or 16 years, those are all a very brief period in the scope of intelligence on earth or even human history. But it does make a big difference in terms of whether the first AGI is also the last

AGI before the singularity, right? So there's two reasons why it may matter who gets their first. One is if the time to fume is short, then the first AGI may be the one that self-modifies itself into a superintelligence, right? The other reason is if the time to Fume is even like three or five years or something,

I mean, then who gets to the AGI first may tell us what sort of chaos the world goes through during that transition because I'm a big optimist about what happens once we get superintelligence. I'm not a Terminator guy. I think by and large, the first superintelligence will probably be benevolent and compassionate to its creators, and it will take very little of his superintelligence resources to provide great bounty to its creator species. But I'm not as much an optimist about what happens geopolitically between the first,

AGI and the beneficial superintelligence that can provide great bounty for all of us with little effort. Like what happens in that interim period may be a mess if a few large companies and associated military industrial complexes are in charge. And both of these are the reason I spent a bunch of time working on decentralized infrastructure for AGI, what I've been doing with SingularityNet and with the artificial superintelligence alliance into which Singularionet merge

with several other decentralized AI projects. I mean, I think if we can make the first AGI not just open source, but decentralized and controlled by a participatory network of software developers and server farm owners and computers, owners and so forth. I mean, then I think we have better odds of getting an AGI that is doing beneficial things for more people at the time when it comes into being, which I think gives

a better odds, both in the case of a rapid fume and a slow fume scenario. So you kind of alluded to it. And I do want to after this get to the benefits of EDI, right? What about the downsides? What if one company, right? And there's all these big government contracts. And you kind of alluded to this. But what happens if that first iteration or declaration of AGI goes wrong? None of us knows.

And it depends on how it goes wrong. I mean, I think if we had a rational, benevolent, democratic world government, which is more of more of a laughing stock year on year, right, in every dimension, both the rational, democratic part and the world part, right? It could well be a reasonable thing to say, let's slow down AGI development until we've better understood how to make the singularity come out well. It will be an ethical trade-off because AGI can cure death.

It can cure world hunger, which we're egregiously failing to do. it can do a lot of good. On the other hand, if it goes wrong, it could do a lot of bad. And there would be a real ethical trade-off for this hypothetical, rational, democratic, benevolent world government to consider, right?

So we're very, very far from that scenario and apparently getting further, not closer, at least on a local year-on-year basis. So the default seems to be all-out AGI arms race between U.S. and China and with Xi Jinping possibly as the rational, benevolent adult in the room, right? So I mean, this is a, in this sort of scenario, what happens is anybody's guess, right? Like, clearly, if big tech in U.S. or China makes a breakthrough to the AGI first, they will have no choice to put that AGI in the hands of their

corresponding military and intelligence establishment, which will use it to achieve some version of partial global hegemony for the corresponding great power. Now, the good news is neither Trump nor Xi Jinping nor Putin, nor Tent, nor Yandex, nor Google wants to kill everybody and leave the earth like a desiccated wasteland, right? Like, I mean, we have a fair bit in common as human beings and one of which is like we want our species to continue like we we want the earth to be a reasonable place to live we all want air we can breathe and water we can drink and we all want to be able to have our kids and grandchildren live on and have a good time right so I mean while while there are downsides I mean I would say the actual psychopaths

who want to kill everybody are not the ones who are likely to to develop AGI, right? So, I mean, what you're likely to see in this case is the early AGI is focusing more attention on making country A stronger than country B or having company A sell more products than company B and the actual good of our species and the birthing of a positive singularity. We're attempting to achieve that as a sub-goal of achieving domination for a certain. company or country, right? And I mean, I think we can still get to a beneficial singularity that way,

but you're going to wind up taking more risks and causing more damage on route than would really be necessary if you weren't trying to achieve the really important things as a subcase of helping this or that party to lord it over the other parties they feel they're competing with, right? And this, I mean, this devolves into a long list of special cases in terms of what does that AGI doing initially, right? But, I mean, you can see the way things are going. Like post-Deepseek, the USVC community is more and more into AGI for robotics rather than just software. Well, you know, what's the most obvious use case for AGI?

robots, right? I mean, you can see it in the Middle East battle theater right now. Like those use cases are going to become mature before home service robots or medical robots, although in a way all of them will rise together, but the military case needs a lot less safety test, right? I mean, there's a lot to worry about in the short term, even though. clearly tremendous benefit also I mean if you look at I mean I myself I'm working on some robotics use cases for education and for medical and elder care robotics and then in in science and medicine like the ability of AI to help you discover new therapies to combat disease and and and prolong life I mean along with all the other kinds of

scientific discovery like my own main use of AI personally is with my researcher head on. The way AI accelerates biomedical research or AI research is incredible. I don't think LLMs are the golden path to AGI. I think they can just be one component of sort of hybrid systems that put LLMs together with machine reasoning, evolution learning other components. I don't think LMs on their own can get us to AGI, although they can be part of an architecture. But on their own, own, they make me like five times as productive as a scientific researcher and an AI developer, right? So, I mean, it's incredible. We're already well past the point where AI tools are

massively accelerating the advent of better and better AI tools, right? Which is one of the things that gives you a, like, whoa, the singularity is near feeling in practice. So I would say, even now, though, AI is being developed in a corporate slash military way, like the tools are helping with every kind of scientific research good and bad, adding a lot of acceleration. Yeah, and you kind of there talked about some of the, you know, obviously commercial or intended commercial outcomes

that would come via, you know, AGI and also geopolitical and talking about some of the potential, you know, negative outcomes. But what about on the positive side, right? Might we see longer lifespans, you know, Madison that are available? Education, it's incredible. Like so my seven-year-old son, uh, Corksey is super bright kid, unsurprisingly. He's into math.

He's a bit dyslexic. Like he sees every letter or word backwards. So his, his reading is okay. Binaaz advances his math. So, I mean, he can ask chat GPT any question he has. It will answer him, right? So he's learned an incredible amount from having speech to text. And I mean, for these years, when his dyslexia, it gets better and better each year.

But, I mean, like at his age, I was reading every encyclopedia. I could get my hands on. He would do that, too, but he reads slowly because of dyslexia. So just the fact that AI lets him ask in words, any science question he comes up with, right? Like, you know what? Who will win a lion or a silverback gorilla? Why? What's the evidence, right?

So, I mean, just being able to, you know, which of the dwarf planets has an atmosphere? Why doesn't that one have an atmosphere? Like, as a seven-year-old, being able to dig into whatever you want on your own time, even if your visual system makes you slower learning to read. I mean, this is amazing for education, right? And that's just running at home after school, right? But school systems will be slower to adopt these technologies. It will come over the next few years, right?

And yeah, for biomedical research, I mean, I'm working in a project called Radov, and we're looking at trying to discover new longevity therapies. And again, no, LLMs are not good at discovering new hypotheses and therapies, particularly. We have other AI tools within our hyperon neural symbolic AI system that are better than that. What we've used LLMs for take millions of data sets, literally, from all around the world, the biologists put online and normalize them all into a common form and succumbus. into a big AI knowledge graph, right? And it's ironic, but until LLMs,

we were held back not from having, we had AI that could crunch biology data sets and come up with new suggestions for therapies and even automatically run lab equipment. But we were held back by data pre-processing. All these millions of biology data sets are in different formats and they're not normalized the same way. And the rolling column headers and the spreadsheets are hard to read.

LLM sifts through all that. They can suck all these biological data sets into a standard form so your other AI tools can analyze it, right? So this was totally not why they were created, but they're helpful for it anyway.

Right now, now the same tools will let U.S. Chinese or Russian intelligence normalize all the random data on the internet about everyone in the world. so as to spy on them and take advantage of them in different ways, right? So it's the same technology that was just developed originally to be a chatbot. Turns out to be useful for managing datasets of all different sorts for good and for ill.

And I mean, that's amazing in all sorts of ways, both on the back end, discovering therapies for people that ached, to prolong their lives and on the front end, like giving tools that kids can use to self-educate, right? So, I mean, it's, but the scope of applications is exactly why, like, this is not going to slow down, right? It's just making too much money for too many people and delivering too much value to, to too many people. Yeah. Speaking, speaking of delivering value, you've delivered a lot in today's conversation. But, you know, as we wrap up today's show, what is the one most important takeaway that you want people to remember when it comes to this concept of creating in the

path to an AGI that benefits everyone? I think the most important thing for people to remember is that we are doing this. All of us are doing this together. Like the singularity, the emergence of AGI and superintelligence is not something being done by a few, like Stanford graduates off in an office on the Santill Road in Palo Alto or something. Like this is being done by the whole global economy cooperating together in quite complex ways. The story is not yet told, and there's loads of ways. and there's loads of ways for all sorts of people to jump in and participate. I looked like Deep Seek was a head fund out of Hong Kong.

All of a sudden, they disrupted what everybody was thinking regarding how expensive it had to be to train AI models, right? And, you know, I started in 2013 in AI Development Office in Addis Ababa, Ethiopia, and we've then pulled in hundreds of AI developers in Ethiopia to help about all sorts of super advanced AI tools. These people, you know, brilliant young guys, most of them have never been out of Ethiopia, let alone to Silicon Valley.

And beyond tech work, I mean, podcasts like this one are blogs, anyone creates if they're telling the truth about how AI works and how it may impact different sectors or even telling how people are thinking and reacting about these things. I mean, all of these things contribute to what our species is doing to create AGI and ASI.

So I think we should all feel very empowered and feel like we are participants in building this crazy-ass future, right? And there is no plan. Like, that's one of the things I realized when I got about 10 years old. Like, before, that I thought there were some people somewhere in the world

who knew what the hell was going on and were pulling the strings and orchestrating things. Around age 10, I realized like, holy shit, nobody on this planet knows what's going on. Nobody is in charge, right? And my friend Leslie Island Combs, the psychologist, gave a talk at our conference on beneficial

general intelligence last year, and he ended it with a quote from Ram Dass. He said, like, relax. Nothing is under control. Right? I mean, depending on your frame of mind, that's either a big relief or totally fucking scary. But the truth is, nothing is under control.

It's open-ended. It's all evolving. We're all playing a role in it. And we can't calculate how big an impact any of us is going to have. Any of us can be having the critical impact, making the difference between the beneficial singularity in otherwise. I get it could be a blog post that you yourself write that influences some young kid to think about things differently and build something amazing that tips the balance into a beneficial singularity. Such an insightful eye-opening and exciting conversation. Dr. Ben Gertzel, thank you so much for joining the Everyday AI show. We really appreciate your time.

Yeah, thanks for having me. All right. If you miss anything, it's all going to be on our newsletter. So thank you for joining us. If you haven't already, please go to your everyday AI.com. Thanks for tuning in. Hope to see back tomorrow and every day for more everyday AI. Thanks y'all. All right. Meet Firefly AI assistant. Now live in Adobe Firefly, the Allman One Creative AI studio. Just describe what you want to create in your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premiere Express, and more in one, in one, conversational interface. You direct the outcome while the assistant accelerates execution. Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit Your EverydayAI.com and sign up to our daily newsletter so you don't get left behind.

Go break some barriers and we'll see you next time.