WEBVTT

1
00:00:00.031 --> 00:00:23.234
The scary open secret in the AI industry right now is that it's possible that we'll end up essentially creating a new species that ends up ruling the world with a 70% chance that this goes horribly wrong like human distinction. That's one possibility. There's many more. It's quite chilling what you're saying. Yeah, it gets me down sometimes. I basically told my wife, like, let's not have any more kids. It's too uncertain. I don't think they'll ever drown the workforce.

2
00:00:24.550 --> 00:00:37.578
Everybody should be afraid that their jobs are going to be lost. And I know this because I went to OpenAI in 2022. What I did there was forecasting as to what the next couple of years might look like. And unfortunately, most of the world is kind of asleep at the wheel and doesn't really realize what's going on with AI. So I resigned.

3
00:00:37.730 --> 00:01:04.679
I read it somewhere that you lost $2 million for not signing an anti-disparagement clause, meaning you couldn't criticize the company. Yes, for reasons I'm happy to get into. But the main thing I've learned is when I go talk to people at Anthropic and OpenAI about forecasting, they're like, it's not going to take that long. You need to shorten them again. Get them back to 2027 or 2028. Because these powerful CEOs, Dario or Sam or Elon, are racing each other to be in control of the most powerful AIs and are literally afraid that if the other guy gets there first, he might become dictator.

4
00:01:04.679 --> 00:01:34.463
I mean, Anthropic is on track to be the entire economy by 2030, but none of these people should be trusted with that much power. So this is the most important thing happening in our lifetimes, probably in all of history, in fact, and it's very important that it go well. So I think that there's a lot we can do to like steer things in a better direction. There's loads of benefits that we could get from AI if we do it right. And if we do solve the problems, then things could be absolutely amazing for everyone. Well, this report here in 2021, it was remarkably accurate. And then you just published this one. Yeah. So this is our new scenario. So let's go through these slowly and one at a time.

5
00:01:34.463 --> 00:01:57.245
I would be incredibly happy if all my predictions turn out to be wrong. Guys, I've got a favor to ask before this episode begins. The algorithm, if you follow a show, will deliver you the best episodes from that show very prominently in your feed. So when we have our best episodes on this show, the most shared episodes, the most rated episodes, I would love you to know. And a simple way for you to know that is to hit that follow button.

6
00:01:57.245 --> 00:02:25.493
But also, it's the simple, easy, free thing that you can do to help us make the show better. And I would be hugely grateful if you could take a minute on the app you're listening to this on right now and hit that follow button. Thank you so, so, so much. Daniel Cocotelo, at the very heart of what you do, what is your mission and why? So what would you do if you thought that superintelligence was coming in a few years?

7
00:02:26.624 --> 00:02:54.670
I guess it depends what the consequences were. Well, let's talk about it. So superintelligence, AIs that are better than the best humans at everything, while also being faster and cheaper, also able to operate robots that can do everything in the physical world that humans can do, but better, faster, and cheaper. If that really is coming in a few years, then we need to prepare and we need to think about how to make it go well instead of poorly. So that's sort of my answer. It's like, I'm doing that to the best of my ability.

8
00:02:54.670 --> 00:03:18.818
So you believe it's coming in a few years? Yes. How could you be so sure? I spend a lot of time trying to forecast this sort of thing. My sort of median estimate, 50% chance, is currently in 2029. Maybe it'll slip to 2028. It's possible that it'll take significantly longer, like maybe 10 years or something like that. But, you know, for reasons I'm happy to get into, seems to me like it's probably happening by the end of the decade. Yeah.

9
00:03:18.818 --> 00:03:47.303
What's less important is the sense of how close we are. What's more important is the pace of the trends. Anthropic, this time last year, was making something like a billion dollars a year. And now they're making something like $60 billion a year. So that's 60x growth in one year, which is extremely impressive even for very small startups. But for a company of their size, it might be the fastest growth in history.

10
00:03:47.759 --> 00:04:00.500
We expect that rate of growth to slow down, but even if it slows down quite a lot, they're still on track to be the entire economy by 2030 or so.

11
00:04:00.702 --> 00:04:26.150
Why should the average person care? The high level thing is absolutely everything is going to change for the whole world and including therefore for them and their families. Could change for the better, could change for the worse, depending on the details of how it's done. So for example, everyone could die, you know? This is the classic loss of control scenario or one version of it. If we do build these super intelligences and we use them to automate all the jobs and

12
00:04:26.150 --> 00:04:55.394
we put them in the military and we have them giving advice to politicians and so forth, they will eventually have accumulated enough real-world power that they don't need humans anymore. And they're smarter than us, they're more strategic, etc. At that point, we sort of have to hope that they are virtuous, that they have the goals that we wanted them to have, the values that we wanted them to have, etc. And the sort of scary open secret in the AI industry right now is that right now that is kind of just a hope. It's not something that we can

13
00:04:55.394 --> 00:05:10.885
be at all confident in. And in fact, there's lots of evidence and arguments that we're not on track to achieve that. So there's lots of reasons, like current AIs, for example, will often lie to people or they will like, you tell them to do something and they go do something else and then pretend that they did it, right? So,

14
00:05:10.885 --> 00:05:40.872
It's an inherently difficult problem to make something that's super intelligent and also has the values and virtues that you want it to have. And it doesn't seem like we're on track to solve that problem. Also, it seems like the sort of problem that you could think you've solved when you haven't actually solved it, right? That's a big reason why this is scary. So for all those reasons, it's possible that we'll end up essentially creating a new species that ends up ruling the world instead of us. And then maybe we go the way of other extinct species in the past that were outcompeted by humans. That's one possibility.

15
00:05:40.872 --> 00:06:10.336
There's many more. Even if you're not worried about that and you think that the AIs will be totally controlled, there's the question of who controls the AIs, right? When there's a couple of corporations that have made these super intelligences and are using them to automate all the jobs, well, that's a lot of power, you know? That's a lot of money. It's a lot of political power. They'll have the best strategists, the best advisors. You know, they'll think faster. Militarily, the countries that has these AIs will be able to absolutely wipe the floor with all the other countries.

16
00:06:10.336 --> 00:06:39.817
The AIs themselves, it's kind of a single point of failure, like central control system where, you know, the CEO of Anthropic, Dario, he coined this phrase, the country of geniuses in the data center. That was his phrase to describe what they're trying to build, you know? I think that's a little bit misleading. I think it would be more accurate to describe it as army of geniuses in the data center because it's not like it's a bunch of diverse, different AIs, you know, living in their different parts of the data center. They're all copies. Yeah.

17
00:06:39.918 --> 00:07:03.053
of the same big model and they're owned by the company. And so they all follow the orders given by the company, right? People should be asking questions of like, who controls this army or these armies and what are they gonna be doing with them? I think that we could very easily end up in a sort of, a situation where some tiny group of people are essentially oligarchs or dictators. And ironically,

18
00:07:04.319 --> 00:07:25.058
Both of these risks, the loss of control and the constitution of power, are things that people in the industry have been thinking about for decades. Even before the AI industry existed, people thinking about AI were talking and writing about these things. And then part of the founding narrative, the founding myth of DeepMind and OpenAI and Anthropic is these problems are real and...

19
00:07:25.970 --> 00:07:51.670
So we need to get there first so that we can handle it responsibly. Those are, I think, the big two reasons. But then I can go on. There's lots more reasons as well. So one thing is, you know, World War III, geopolitical conflict. If AI does, in fact, get incredibly powerful, that's going to change the balance of power between nations. That's going to disrupt a lot of things. That puts us at increased risk of crisis more generally, right? Another one, what about those jobs? Yeah.

20
00:07:51.670 --> 00:08:09.170
you're going to lose your taxi job, but not just the taxi driver, everybody, pretty much. There might be a few exceptions, like people whose jobs for legal reasons are only allowed to be done by humans. But for the most part, everybody should be afraid that their jobs are going to be lost, even if we manage to avoid all the other problems, right?

21
00:08:09.170 --> 00:08:35.393
This narrative has started to emerge. And I've had several interviews on the show where I've interviewed people who are very, very scared and anxious about AI. And these are people that have worked in the industry for sometimes decades. The counter narrative coming over the hill is that this is doomerism, that these people are, for whatever reason, just trying to scare people and that they don't really understand what they're talking about. How do you respond to that sort of counter narrative? And you must have seen this emerging yourself, especially from people who stand to benefit, dare I say.

22
00:08:35.393 --> 00:08:54.631
Yeah, exactly. This counter-narrative is fairly recent, and it's been pushed by the people who stand to benefit from it. And it's not true. These concerns have been around for decades, since before the AI industry existed. They're actually pretty reasonable concerns. If you take the companies at their word and imagine that they are in fact going to build superintelligence...

23
00:08:54.631 --> 00:09:23.588
Well, it raises a lot of questions like who's going to control it? Will anybody control it? What about the jobs? You know, like these are just kind of obvious implications to be thinking about and worrying about. Who are you and what's your story? My name is Daniel Cocutello. I currently run the AI Futures Project, which is a small nonprofit that mostly focuses on forecasting the future of AI. Before that, I worked at OpenAI. AI forecasting. Yeah, so...

24
00:09:24.027 --> 00:09:42.893
Think about how industry analysts who work for hedge funds and stuff will make these forecasts of like, here's how many cars Tesla will be selling five years from now. Or like, here's what the price of electricity will be in two years, right? That's forecasting. I was doing that, but specifically focused on AI.

25
00:09:42.893 --> 00:10:01.270
The reason I was doing it is because it's incredibly important to see where this is all headed. Why did you go to OpenAI? What did you do there? What did you observe while you were there? And how did it change your perspective on the future of AI, but also, I guess, OpenAI as a company? And for anybody that doesn't know, OpenAI are the company that produced ChatGPT.

26
00:10:01.405 --> 00:10:22.246
Yeah, so I went to OpenAI in 2022. A large part of what I did there was more forecasting. AI 2027 is a scenario that you may have heard of. I did like smaller, you know, lower effort versions of them internally for just internal circulation of like, here's some guesses as to what the next couple of years might look like. I also worked on evaluations for dangerous capabilities. So...

27
00:10:22.246 --> 00:10:47.373
you know, trying to measure the AI's cyber abilities or persuasion abilities or situational awareness. And I also briefly was on a capabilities team doing reinforcement learning to create agents. AI is in fact getting a lot better. And I can say more about why, you know, scaling laws, deep neural nets, bigger, trying to learn more data, become more efficient, more competent at those things. I also...

28
00:10:48.706 --> 00:11:10.981
became a bit more disillusioned with the AI industry. So OpenAI and Thrapik and DeepMind all had these sort of founding narratives of like, yes, these risks are real, but we've thought about them and we're going to try to handle them responsibly. And that's why it's important for us to keep doing what we're doing. And I increasingly came to think that these were rationalizations that

29
00:11:11.099 --> 00:11:37.998
to justify what they were doing rather than sort of like deeply guiding their actual behavior. And that when push comes to shove, they'll follow their incentives rather than do what's actually good. So you're inside OpenAI at the time and you start to believe that they're following commercial incentives versus the, I guess, social or societal incentives that they founded themselves on. Sort of. I mean, I wouldn't actually describe it as commercial incentives. I think I would describe it as...

30
00:11:38.285 --> 00:11:53.219
power seeking incentives. So like it's true that the companies care a lot about making a lot of money but especially at the very top of these companies like the leaders they understand that this is about more than just money.

31
00:11:53.489 --> 00:12:16.237
You know, there are these emails that came up in the lawsuit between Musk and OpenAI. A bunch of emails were surfaced in that lawsuit, which you can go read. And in some of them, the founders of OpenAI were talking back in like 2017 about how the reason why we made OpenAI was because we were worried that Demis Hassabis at Google was going to become dictator with AGI.

32
00:12:16.237 --> 00:12:34.732
Even back then, they were, this is obviously about more than just money. Like these powerful CEOs are literally afraid that if the other guy gets there first, he might become dictator and they don't trust each other. And so that's why they are racing as hard as they can so that they're the ones who get there first, so to speak.

33
00:12:35.508 --> 00:13:04.043
Have you met Sam Altman? Yeah. And did that shape your opinion of his incentives or why he's doing what he's doing? Because there's a lot, you know, speculated about what his incentives are. I mean, his most recent narrative says for the good of humanity. I think that's what. Yeah. I mean, I think the main thing I've learned is don't pay attention to the narrative. You know, like what they say to one person is just different from what they can say to some other person at the same time. And what they say in public is a third thing entirely. I think you should judge people by their actions, not by their words.

34
00:13:04.904 --> 00:13:33.794
And why are you no longer at OpenAI? Largely for the reason that I mentioned. So I became gradually disillusioned with how the company was going to behave. For example, when I first joined in 2022, at least the people I talked to, my colleagues at the company, there was this general sense of like, of course, we wouldn't actually just build superintelligence as soon as possible. Once we started getting really close, like once we started getting to AIs that could maybe automate the AI research process, we would pause and figure out how to make it safe.

35
00:13:35.195 --> 00:13:53.656
that's because we're the good guys and that's obviously the safe thing you should do rather than just going full speed ahead. But we're worried about other people who might not pause, you know, our competitors, Google, for example. And so that's why we need to be in the lead so that we have that room to do the safe stuff, right? That was sort of like a thing that seemed like...

36
00:13:53.656 --> 00:14:06.481
maybe like the median position or something among the colleagues I talked to when I was there when I started, including people like Sam, you know, including the leadership. And then by the time I left, I was like, oh man, they're really not going to do that, are they? Like they've sort of,

37
00:14:07.021 --> 00:14:36.029
you know, partly because this has become more politicized and they've become bigger and been under more scrutiny. People have started asking, like, why are you doing this in the first place if it's so risky? And so they've pivoted their narrative to being more like, actually, it's not that risky, you know? And so, yeah, I mean, it seems like they're just going to keep going roughly as fast as they can and hope that they can figure it out on the way. How did your time at OpenAI come to an end? I resigned in 2024. I had a nice goodbye party. What were the reasons you gave for quitting OpenAI?

38
00:14:36.772 --> 00:15:06.758
I thought that we were rationalizing too much and that we needed to think more about what would actually be good for the world. I wanted more freedom to publish. So AdOpeneye, as it became a bigger company, it became more of a normal tech company with incentives and, you know, a PR department and things like that. And so it started becoming more difficult to publish the sort of research that I was doing. For example, those scenarios that I mentioned, I couldn't publish those, right? They're just for internal use.

39
00:15:06.758 --> 00:15:33.691
I thought that that was a shame because right now most of the world is kind of asleep at the wheel and doesn't really realize what's going on with AI and doesn't really realize what's coming in the pipeline a couple years from now. And the companies aren't really incentivized to tell people that much about it. I mean, they say some vague stuff in a sort of hypey way, but, you know, well, they didn't want me to publish the scenario, for example, laying out like here's how things might actually look.

40
00:15:33.843 --> 00:16:02.648
I'm just super curious as to what it's like being in a company like that when, you know, ChatGPT3 is released. You were there at that time, right? Mm-hmm. Which was a moment where I think the whole world stood up and realized that this technology was powerful. Yeah. And the conversation really began from a society level. The company starts growing super quickly. Yeah. Quicker than I think anybody could ever have imagined. And what was it like inside there? What did you see change over that period of time? Yeah.

41
00:16:02.783 --> 00:16:32.129
I remember one all-hands meeting where Ilya said something like... Ilya being... Ilya Sutskiver, who was head of research at that time. He said something like, okay, now the world is starting to pay attention. Each of you is going to be the most popular person at every party for the next year. Don't let it get to your head. Focus on the mission. Got to build AGI. The company grew a lot. It already wasn't really feeling like a nonprofit when I joined, but it definitely didn't feel like a nonprofit by the time I left. Lots of new people came in.

42
00:16:32.365 --> 00:16:56.480
Ironically, the amount of conversation about superintelligence and the implications of superintelligence arguably sort of went down over time due to this growth, right? So because the company would double and then double again and then double again, all these new people were coming in from other parts of the tech industry who hadn't really been thinking about these things and were attracted by the high salaries.

43
00:16:56.564 --> 00:17:26.483
You lost $2 million for not signing an anti-disparagement clause, which would mean you could speak, you couldn't criticize the company. Ah, yes. Well, so I got to keep the money. Oh, you got to keep the money. Yeah, so what happened was after I had left, said my goodbyes, et cetera, I got the exit paperwork, and it included this clause that said you basically have to agree not to criticize the company again, and also a clause saying you can't tell anyone about this. And so I thought that was kind of,

44
00:17:27.749 --> 00:17:55.188
rich coming from a nonprofit that's supposed to be, you know, for the benefit of all humanity. So I didn't sign it. And if you don't sign, you don't get to keep your equity. So your compensation, you know, what they pay you is a bunch of money and then also a bunch of stock, basically. But then they had this stuff in the contract that they get to yank back your stock if you don't sign this thing. And my wife and I, you know, we're...

45
00:17:55.188 --> 00:18:01.955
upset about this. We talked about it for like a month or two, consulted some lawyers, and then ultimately decided to just refuse to sign.

46
00:18:02.613 --> 00:18:31.368
which would mean you would have lost $2 million. That's right, which was like 80% of our net worth at the time. Fortunately, it didn't go the way we expected. It blew up, basically, on the internet. When people heard that we had done this and that we had said no, it became this huge scandal. Employees at the company started asking questions in Slack and asking leadership, like, wait, what? Why are you going to take away our equity? What is this? Because a lot of people hadn't really noticed that.

47
00:18:31.368 --> 00:18:58.756
this before. It had been whispered about, but it hadn't been sort of like a thing that most employees knew about. And so they backtracked and they said, never mind, never mind. We'll change the paperwork. You can keep the equity. It's fine. And Sam Altman came out and said he was embarrassed that he didn't realize this was good. Yeah, he had no idea, apparently. You don't believe him? No. I think you probably know. And if he didn't know, then people close to him probably did, such as his head lawyer. Why did you decide not to take the $2 million

48
00:18:59.127 --> 00:19:11.868
I mean, most people would have, I think. It's true. Most people would have and most people did. And, you know, money is nice, but like, it's not the only thing, you know, sometimes it's good to take a stand on principle.

49
00:19:12.357 --> 00:19:35.628
I keep mentioning superintelligence. Perhaps I should say more about like the sequence of events that the companies are planning to do. So right now they're focusing on automating coding. They're taking their AIs, they're making them bigger, they're training them for longer, and they're especially focusing the training on getting them to be good at autonomously writing and editing code. Because...

50
00:19:35.628 --> 00:20:03.809
That will help the companies go faster, right? If they can automate the code, then they can do their own work better and faster and accelerate progress. The next step, which they've already begun, is to look at the rest of the research process as well. Coming up with ideas, analyzing experiments, communicating those results, all the other parts of the research process. They're trying to figure out how to train AIs to be good at those as well so that they can have AIs do the entire thing autonomously.

51
00:20:04.231 --> 00:20:30.185
When you say do the entire thing, what do you mean do the entire thing? So like Anthropic and OpenAI in particular are trying to automate themselves. Like they're trying to make it the case that they don't really need human employees anymore. They just have a giant army of AIs that's churning away, doing all this autonomous research to make better AIs, to train the new AIs, put them in charge so they can make even better AIs and so forth.

52
00:20:30.185 --> 00:20:48.359
And of course, not all just happening internally, but also interfacing with the world, going out and talking to people, collecting the data, setting up the training environments, doing the business deals, and so forth. They're trying to automate all of that. The reason why they're doing this is because they're trying to get to a position where they have...

53
00:20:48.865 --> 00:21:14.870
AIs that are superhuman at everything, superintelligence, and they're trying to get there before their competitors do. Needless to say, this is incredibly dangerous, I would say, you know? And in addition to being dangerous, it's a power grab, right? Like, if they actually succeed at this, then they'll be sitting on top of this army of superhuman AIs that will give them immense leverage over all sorts of other actors in the economy. Yeah.

54
00:21:14.870 --> 00:21:34.141
insofar as they can work out something with the presidents and integrate it into the military or whatever, then that would give the U.S. immense hard power over all other countries, right? Obviously, nobody knows exactly when this is happening, but a very disquieting thing has happened over the last year to me, which is that when we published it in 2027,

55
00:21:35.288 --> 00:21:51.961
people were generally of the opinion that my timelines were too short and that like probably it would take more than 2027 until we got to the sort of events that I was just mentioning. You know, recursive self-improvement, AI is automating the whole research process, superintelligence.

56
00:21:51.961 --> 00:22:17.105
These types of milestones, they happen in 2027. Which is this research paper you published. That's right. It's a scenario forecast that sort of lays out, like, month by month, a possible future trajectory. There was sort of like, at the time that we started writing, it was my best guess as to what would actually happen. Obviously, there was a lot of uncertainty, but, you know, I thought it's valuable to make a concrete guess just to sort of see what it might look like.

57
00:22:17.105 --> 00:22:39.042
And at the time we were writing this, a lot of my friends in the AI industry and in nonprofits and so forth that work on AI, a lot of people were saying like, yeah, that stuff's going to happen, but like, it'll probably take a couple years longer than you think. And now it's more 50-50, especially when I go talk to people at Anthropica and OpenAI. They're often like...

58
00:22:40.223 --> 00:23:06.532
Yeah, no, 2027, that's basically what's going to happen. Just like you wrote. Why did you update your timelines? Oh, yeah, context for this is after writing it in 2027, I shifted my timelines to be a little bit more conservative. So at the time that we published, my 50% mark was in 2028, not in 2027. And then after we published, progress just seemed like it was going a bit slower. And so I updated to 2030.

59
00:23:07.240 --> 00:23:16.522
which is still could happen sooner, could happen later, 2030. But now when I talk to people in the companies, they're like, it's not going to take that long.

60
00:23:16.623 --> 00:23:44.990
They're like, oh, you need to shorten them again. Like, get them back to 2027 or 2028, you know? So that's a bit disquieting. Again, don't know how long it's going to take, but this is the stated plans of the AI companies is to do this incredibly dangerous thing, and they think that they're just a few years away. So you wrote this report here, What 2026 Looks Like, and you wrote this in 2021, and it was remarkably accurate. Helped make a name for yourself amongst...

61
00:23:44.990 --> 00:23:57.359
amongst everybody in AI. And which one was it that J.D. Vance, the vice president, read? I think it was this one, wasn't it? Yeah, this one. And then so then you published this one, AI 2027, and this was published, I believe, in 2025? Yeah.

62
00:23:57.545 --> 00:24:24.967
Yes, that's right. April. What were you forecasting in here? What are the key things that you said in here for people that haven't read it? The high level version of it is they automate decoding, then they automate the rest of the research process, then the pace of progress accelerates dramatically, they get to super intelligence, they're working with the government, specifically the president, the executive branch naturally wants to control this technology and otherwise wants to use it to beat China and integrate it into the military and so forth. By this point, it's sort of

63
00:24:25.169 --> 00:24:45.723
doing basically all the work itself. I mean, it's super intelligent. So it's coming up with all these great ideas for how to integrate itself into everything and all these new technologies that's invented and so forth. And because of the race dynamics and because of the profit motive, they end up deploying it everywhere. And it builds robot factories, they build more robots, they build more robot factories, et cetera. Transforms the world entirely.

64
00:24:45.723 --> 00:25:03.644
And then at some point it has enough power, it meaning the AIs, have enough power that they don't have to pretend to be aligned anymore. Right? Then they stop listening to orders. That's the race ending of AI-2027.

65
00:25:04.049 --> 00:25:21.194
We also wrote a sort of different branch, which is the slowdown ending, which was intended to sort of illustrate the concentration of power issues that I mentioned previously. So what if, hypothetically, the alignment issues get sorted out sufficiently quickly? Like, what if it turns out that, like,

66
00:25:21.498 --> 00:25:46.456
It's not too hard. With two months of slowdown, we can figure out how to make the AIs robustly do what we want and have the values that we want them to have. So that's one possible branch. And in that branch, it looks pretty similar. You know, they take the jobs, beat China, etc. But instead of the AIs ultimately killing everyone, they create this sort of amazing utopia. But the amazing utopia is...

67
00:25:46.456 --> 00:26:12.832
whatever the people who controlled the AIs wanted it to be, right? And so that would be a very small group of people, like the president, some CEOs, etc. There should be a button just down below here. And if it says subscribe, you're already subscribed. If it says subscriber, that means you're not yet. And if you're not subscribed, please could you do us a favor and hit that button? It helps the show more than you know. And according to the algorithm, you're someone that watches our show, but you haven't yet hit that button. Thank you so much. Thank you.

68
00:26:13.253 --> 00:26:31.597
Is there any possibility, do you think, that we never get to this thing called AGI? And how do we distinguish AGI from this term superintelligence? What's the difference? Yeah, so the difference is that AGI is a more vague and weak term. Superintelligence is a bit more precisely defined. It's better than the best humans at everything, faster and cheaper.

69
00:26:31.597 --> 00:26:57.415
AGI is more like, it stands for artificial general intelligence, which means AI can do things in general rather than like some specific task. And so arguably we've already achieved AGI, right? If you use cloud code or something like that, it's like it can do a lot of stuff. It's almost kind of like a little employee that you can like have go do stuff. So it is quite general. It's not maximally general, though. It can't do everything. Whereas super intelligence by definition can do all the things that a human can do, but better. Yeah.

70
00:26:57.601 --> 00:27:15.067
And how does this sort of overlap with robotics? Because obviously we're seeing this huge robotics boom at the moment. There are some real world things that humans can still do because these AIs are still stuck on my computer. The way that people talk about this is that they basically just say, we've achieved super intelligence for cognitive tasks. Then you can talk about like...

71
00:27:15.657 --> 00:27:45.442
full superintelligence that can do the physical stuff. And are we going to get there? Are we going to get there with both? I think so. I mean, again, this is not something that we can be certain about. You asked, like, is it possible we'll never get there? Yes, it's possible we'll never get there. I don't think it's likely, though. I think that there's nothing sort of, like, magical about the human brain. It's, you know, it's just a bunch of neurons. It is possible for a digital system to do similar functions in the same way that, like, you know, a plane can fly.

72
00:27:45.442 --> 00:28:11.547
just like a bird, not in the same way as a bird, necessarily. Like it doesn't have, it's not flying in the same way that a bird flies, but it flies, you know? So it does seem like, yeah, like, seems possible. You've written all these, you know, these research reports. You're working on another one that will be released likely on the 9th of July. You have worked inside OpenAI. You then quit OpenAI because you were concerned about what was going on there and about the future of the industry. You know more than I do.

73
00:28:12.458 --> 00:28:39.914
Are you optimistic about the future or pessimistic? Are we heading to a bad place if things don't change based on everything that you know? I think we are headed to a bad place if things don't change. I'm not confident in that. I would say something like 70%. It's very, very hard to predict, of course. But yeah, it seems like the current default path is heading towards a very, very scary place. How do you contend with that personally and emotionally? It's rough. I mean, I think it's the sort of thing that like,

74
00:28:41.888 --> 00:29:11.707
gets me down on a regular basis, but also I've been dealing with this for so many years now that I've sort of gotten used to it, if that makes sense. Yeah, yeah, I'll put it this way. I would be incredibly happy if all my predictions turn out to be wrong and AI hits the wall, for example. It gets you down on a regular basis. I used to be known as a pretty chipper and optimistic person, but in 2020...

75
00:29:11.707 --> 00:29:36.142
my AI timelines predictions started collapsing due to GPT-3 and the scaling laws papers and the BioAnchor's report, which I can talk about if you're interested. But basically, some events happened in 2020 that convinced me that actually this stuff was quite plausibly coming by the end of the decade. And humanity is very obviously not ready for this, you know, in a whole bunch of different ways. And so that's obviously very scary.

76
00:29:36.226 --> 00:29:49.608
And that's an extremely scary world because of all the things you've said. But again, because of this recursive self-improvement where AIs can train themselves. And at such point, we're starting to lose hold of what's going on here. I mean, the AIs are already training themselves, to be clear. It's more like

77
00:29:50.013 --> 00:30:14.600
Closing the entire research loop. Okay, doing everything. Yeah, like right now, a lot of the training data is generated by AIs. A lot of the reinforcement, like the grading that happens, doling out of positive and negative reinforcement, is itself done by AIs. Can you explain that in layman's terms? Yeah, so an important thing for everybody to understand is that modern AI systems are not software in the normal sense. I mean, they are technically software, but...

78
00:30:14.600 --> 00:30:33.044
They're not lines of code, you know? It's not like some engineers at Anthropic went and wrote lines of code that basically says like, you know, when the user asks for this type of thing, then go do this type of thing for this many steps or whatever. There's nothing like that. Instead, it's a neural net, you know? What's that? Yeah.

79
00:30:33.179 --> 00:30:59.352
Well, think about how the brain is a bunch of neurons connected to each other that are firing signals back and forth. The brain learns over time. The types of patterns of firing that caused success, that caused a dopamine rush or various other types of feedback, get reinforced and fire more often. And the types of patterns that caused failure, like touching a hot stove, get anti-reinforced, get, you know,

80
00:30:59.352 --> 00:31:21.222
destroyed so that they fire less often. And as a result of all of that, you, over the course of years, learn to act in the world. And you learn all sorts of skills and you learn world models. You learn like beliefs about the world and you can sort of like mentally stimulate how it's going and stuff like that. So artificial neural nets are like that, except artificial. So it starts off as a giant...

81
00:31:22.673 --> 00:31:36.157
tangled spaghetti mess of randomly generated artificial connections called parameters. These days, they might be something like 10 trillion parameters in the biggest AIs.

82
00:31:36.983 --> 00:31:54.247
So it starts off randomly generated. So it's, of course, completely useless. Like if you give it some input, it'll just produce gibberish as an output. But then they train it. And they start with pre-training, which is where you give it a bunch of internet text and you show it the first piece of text

83
00:31:54.247 --> 00:32:15.307
And you put that in as the input, and then it gives a gibberish output. And then you positively or negatively enforce it based on how accurate that output was at predicting the next piece of text. So it's basically playing this game of like predict the next word. Isn't that how it happens with babies? I think I had a neuroscientist tell me that babies have more neural connections than...

84
00:32:15.307 --> 00:32:32.975
than adults and yeah it says toddlers have twice as many neural connections as adults and they I guess they whittle down through reinforcement yep we have more pathways when we're younger and just like the process of training in AI we're trained down to like remove the ones that aren't useful and build up on the ones that are

85
00:32:32.975 --> 00:32:59.570
Yeah, it's both pruning and strengthening. And it seems like in humans, it's actually more pruning than strengthening, but it's both. And in AI, it's the same thing. It's both. So the first portion of training is where they train the AI to predict text, which is kind of like training it to read. And a similar thing does happen in humans. So basically, the random tangle gradually takes shape and gradually sort of coalesces into more useful circuitry that...

86
00:32:59.570 --> 00:33:24.140
has stored lots of facts about the world and has stored lots of skills for how to, you know, process information and transform it and then produce predictions. That's just the first step. After they do the pre-training, then they try to teach it more useful skills besides just predicting text. And so, you know, by the end of the process, they've thrown lots of coding problems at it. And they've said like, here's a coding problem.

87
00:33:24.140 --> 00:33:49.081
go here's a coding problem here's an environment you have access to this virtual computer here's like the code base you're working with you can write code you can edit the code you can run the code you can read it you can use the internet go go go and it does that for a while and then based on how successful it is reinforcement happens and they have thousands maybe millions of examples of coding problems like that that they train it on and that's why they're so good at coding now

88
00:33:49.570 --> 00:34:16.267
So what does superintelligence look like in this regard? Is it just more of these connections? And how would they get more connections? Can you explain that to me? So there's different AI models, right? So there's like GPT-3 and GPT-4, GPT-4.5 and GPT-5 and GPT-5.5 and 5.6, right? Sometimes they're just the same previous model, but with extra training. Sometimes they're a whole new model that's been trained from scratch, including starting the whole pre-training process again. Yeah.

89
00:34:16.267 --> 00:34:34.643
Over the last couple of years, they've done several new rounds of starting over from scratch. And typically when they start over from scratch, they make the whole thing bigger. The artificial brain much bigger. Right now they're at something like 10 trillion parameters. Back in 2020, it was more like 175 billion.

90
00:34:34.998 --> 00:35:01.087
So we've grown like two orders of magnitude in six years. Two orders of magnitude? Yeah, like two 10xs, so 100x, right? So that process is continuing. They're also improving the algorithms themselves. So they're not literally just the same type of AI but bigger. They've also come up with all sorts of ideas for how to change the structure of the connections and the neurons and so forth and change the reinforcement structure.

91
00:35:01.390 --> 00:35:28.863
algorithms that they're using and to change the training data that they're training on. All sorts of tweaks that have made this whole thing more efficient. We're literally building a brain. Basically, yeah. As they make more brains, they're getting better at making, they're making them bigger and making them more efficient and so forth. And it's literally modeled on the brain, like the way it works, right? It's certainly heavily inspired by the brain, but I shouldn't overstate the analogy. Like there's lots of differences too. So for example, the transformer architecture. Which is?

92
00:35:28.863 --> 00:35:54.327
which is the architecture that they use for these LLMs, is not really recurrent. So the information sort of flows one way, rather than allowing all these sort of little loops on the inside. Also, the backpropagation algorithm is different from the sort of learning that naturally happens in human brains. So there are some differences, but yes, like, broadly speaking, we are sort of making artificial brains. It's kind of like for brains what, like, a plane is for a bird.

93
00:35:54.479 --> 00:36:16.366
That's a really good analogy. That analogy helped me think through a bunch of questions people often ask about AI when they said, can it be creative? But actually that analogy kind of helps me understand that actually that maybe that's not the question. It's can it produce something that you would consider to be creative? Because creativity, people think of it as like a process. But actually it's judged based on the output, isn't it?

94
00:36:16.366 --> 00:36:37.223
I mean, you can get philosophical about like, is it truly creativity that they have? But you can also be like, well, I mean, just look at all the stuff they're accomplishing, you know, and it seems like they're going to be accomplishing a lot more in the near future. Yeah, I do. I asked the question about how this weighs on you personally, because I can sense that you're actually personally bothered. I mean, I think the situation is crazy. Like,

95
00:36:39.890 --> 00:37:09.877
First of all, it's very exciting. Like AI is really fascinating and interesting stuff. I've been following the field for more than a decade now. I've been part of it for some years. And it's really cool, really interesting. And it's really fun to think about what's going on inside these artificial brains and why they are the way that they are. And it's really cool to see all the applications of this technology out in the world. But it really seems like we're on a pretty scary path. And the more you think about it, the more worried you get. And, you know, in stories...

96
00:37:09.877 --> 00:37:33.080
it always ends well, but this is real life. And I think we have to sort of stare reality in the face and tell it and realize that like, it might not actually end well, you know? Were there any recent, dare I say, I was going to say eureka moments, but paradigm shifting moments where even your own sort of mental model of what's going on here and how this is going to look were changed for better or for worse?

97
00:37:33.265 --> 00:37:59.894
For better or for worse, and probably for worse, things are kind of on track for AI-2027. There are a few things that have been different. Not exactly like paradigm shift differences, but like there have been some differences from what we expected at the time we wrote this. So the government has actually gotten involved faster than we expected and has been more aggressive than we expected. So the export controls on Mythos being the biggest example and also threatening Anthropic with being destroyed by the Defense Protection Act. Yeah.

98
00:37:59.894 --> 00:38:26.067
Another thing that's been surprising to us is that Anthropic in particular has gone from second place to first place in the race, basically. Why do you think that happened? Because it seemed like ChatGPT were out front and clear as it relates to OpenAI were out front and clear, but suddenly Anthropic have lapped them. Yeah, I mean, I guess they have probably higher talent density and better strategy, but

99
00:38:26.152 --> 00:38:46.300
But not by a lot, but enough to make a difference. Why do you think they have more talent? Well, they don't have more compute. Like, what are the inputs, right? Like, they're in the lead now. They used to be behind. What are the possible explanations for this? Well, it could have been that they had more resources, like more compute, more money. But that's not true. They have less resources and less money, right? Right.

100
00:38:47.431 --> 00:38:57.843
So then I guess Talentsin is the next best alternative. You could maybe say strategy. Some combination of those things, yeah. Something that wasn't just like the amount of resources they had.

101
00:38:58.602 --> 00:39:24.961
Just like Jon Jones, where marginal improvements in your cognitive performance can have a massive impact. Sometimes I podcast for 10 hours a day. Over the last couple of weeks, I've been in filming for a TV show, and then I have like one or two days off to get all of my work done, which means there's lots of cognitive load. And so I turn to ketones because I find myself more articulate, able to think more clearly, able to work out better when I'm fueled by ketones. And so the reason I became a co-owner of this company, and the reason why they now are a sponsor of this podcast is,

102
00:39:24.961 --> 00:39:54.273
is because I remember one of my team members called Christiana, she tried it once and came up to my desk and she goes, this is the best product ever made. And I think in part that's because she really cares about those cognitive benefits, as I do, as John Jones does, and as I think most of my listeners probably will. So if you haven't tried these yet, all you have to do is go to ketone.com slash Stephen, and you'll also get 30% off your first subscription order. You'll get exclusive Ketone IQ merch, and of course, cognitive benefits that might just change your life.

103
00:39:54.273 --> 00:40:19.568
Much of the reason most people haven't posted content or built their personal brand is because it's hard and it's time-consuming and we're all very, very busy and if you've never posted something before, there's so many factors in your psychology that stop you wanting to post. What people will think of you? Am I doing this right? Is the thing I'm saying absolutely stupid? All of these result in paralysis, which means you don't post and your feed goes bare and

104
00:40:19.568 --> 00:40:40.645
I'm an investor in a company called Stan Store, which you've probably heard me talk about. And what they've been building is this new tool called Stanley that uses AI, looks at your feed, looks at your tone of voice, looks at your history, looks at your best performing posts and tells you what you should post, makes those posts for you. You can also just use it for inspiration. And sometimes what we need when we're thinking about doing a post for our social media channels

105
00:40:40.645 --> 00:40:54.736
is inspiration. Building an audience has fundamentally changed my life and I think it could change yours too. So I'm inviting you to give this new tool a shot and let me know what you think. All you have to do is search coach.stan.store now to get started.

106
00:40:55.900 --> 00:41:21.922
A friend of mine who knows some of these people sat me down once upon a time in London. He's actually said this a few times to me, but I remember one particular conversation where he says that some of these AI CEOs predict the probability of extinction at being, I think he said 7%. I don't know why I have that number in my head, but I remember it being less than 10%. And the point he was making to me was that even if it was 1%, like if there was 100 buttons on this table now...

107
00:41:21.922 --> 00:41:46.002
And one of them would end the world. Would I dare press any of them? You know, I wouldn't press any of them. But he made the case to me that these AICOs are very smart and they understand super intelligence and that they think actually if there was 100 buttons on this table right now, maybe 10 of them could end the world. I've heard you say, I think it was on the Daily Show, the interview you did, you said that you think there's a 70% chance of human extinction due to AI. Yeah.

108
00:41:46.272 --> 00:41:54.423
I wouldn't say human extinction exactly. I would say something like 70% chance that this goes horribly wrong like human extinction. That's just one of several possibilities. But yeah, basically.

109
00:41:55.807 --> 00:42:18.622
Like, for example, possibly the AIs take over and then don't actually kill everyone. You know, maybe they do something else. Like just because they've taken over doesn't mean they're definitely going to kill us. Right. They might, but they could do something else. So that's why I don't usually say like 70% chance of like actual human extinction, but 70% chance of like something like AIs taking over, some sort of very big catastrophe like that.

110
00:42:18.622 --> 00:42:48.592
They could lead to human extinction. I guess I've got two points there, which is you've been around these CEOs. I mean, you've worked for Sam Altman at OpenAI before you quit. Do you think that they think there's a chance of human extinction? Yes. But I think that an important thing to understand is that people sort of believe what they need to believe in order to think that they're good people and that they need to keep doing what they're doing. This is what rationalization is. And so I think that the tech CEOs have genuinely convinced themselves

111
00:42:48.592 --> 00:43:15.052
probably things are going to be fine. And that the way to make things fine is for them to keep doing what they're doing. And like, they need to like, make sure that like, you know, Sam needs to make, Sam's probably thinking like, can't let Dario or Elon get there first. You know, I know Dario is thinking Sam can't get there first. Elon's thinking that like, you know, they, they, they've all probably convinced themselves that like, oh yeah, like maybe it'll go horribly wrong, but like, probably it's going to be fine. And probably, you know, I should be the one in charge.

112
00:43:15.187 --> 00:43:37.580
It appears to me that Anthropic are the only ones that are talking about the potential chance of extinction or a catastrophic event or the real downside still. They seem to be the only ones that are still publishing on it. And now they're actually becoming the enemy in many respects of the tech industry in San Francisco. I'm watching a lot of interviews and everyone's attacking Dario because he's saying, listen, things could go bad. They're calling him a Duma.

113
00:43:37.580 --> 00:43:56.581
and questioning his incentives, even with Mythos, which is a Claude model that they started to warn the world about. Again, he is attacked immediately for saying that. My question is, do you see him as being slightly different from Sam in this regard? Yeah, I mean, it seems like Anthropic and Stereo have been more willing to

114
00:43:57.712 --> 00:44:26.517
say and do things that are costly to their bottom line, at least in the last year or so. That's an example of it. Like, I don't think that really wins them favors in the administration or among their investors, to say that type of thing. And, you know, a better example is just the whole fight between the Department of War and Anthropic was an example of them doing something that, like, cost them a lot of money and, even more importantly, cost them a lot of power for something. Like, they could have just signed the contract, you know.

115
00:44:26.517 --> 00:44:53.213
That said, I really don't want to be in a situation where we're like, which CEO is the least bad CEO? Let's support that one. You know, like none of these people should be trusted with that much power, basically. Nobody should. Nobody should. Regardless. Regardless. Yeah. So on this point of the buttons, you do believe that they think there's a credible chance of extinction. Yeah, but they've convinced themselves that like, it's probably fine. And also it'll be even worse if I'm not doing it, you know?

116
00:44:53.213 --> 00:45:11.624
Like that's what they'll say inside the companies too. Like some people will be like, okay, well, if we stop, what about the other guys? Like they're not going to stop, you know? Yeah, this has always been why I've had this outstanding question, which is how does this not go bad when human incentives seem to rule the day when you look at history and all of the human incentives are saying, well, you're damned if you do.

117
00:45:11.624 --> 00:45:37.730
I.e. you're damned if you carry on developing these bigger and bigger and bigger AI brains. But you're also then damned if you don't from a geographical perspective, because the United States will lose to that country, or this company will lose to that company. So when you just look at human incentives and go, how does, how does, if just purely incentives and disincentives, how does this end? Well, it carries on going. Seems like it. I mean, there is a caveat to that, which is a hopeful caveat, which is that, first of all, if the world wakes up to all of this.

118
00:45:37.730 --> 00:46:00.511
then there can be a more serious conversation about regulation and international treaties and things like that. And that can change the incentives, right? So the government could come in and say, like, actually, here's some rules that you all have to follow. And because they're rules that you all have to follow, then you're not incentivized to, like, break them anymore because you get punished if you break them. And...

119
00:46:00.511 --> 00:46:22.178
Everyone else is also following them too. And so, you know, it's fine. So there is that sort of like ray of hope that like we can change the incentives if the government, especially the U.S. government, but then later other countries act to change the incentives. But that's not going to happen until people sort of wake up to all of this. The second thing is that even individually, at some point...

120
00:46:23.461 --> 00:46:45.787
you know, Dario or Sam or Elon might realize that, like, actually it's, like, not even in their own interest to keep racing unilaterally. And the problem with that is it's only if it gets extremely obvious and extremely dire. So, like, in AI 2027, in that scenario, there's this choice point that I mentioned. And in one case, the AIs are misaligned. In the other case, the AIs are aligned.

121
00:46:46.613 --> 00:47:03.370
At that choice point, we have like one branch that depicts the misalignment ending and one branch that depicts like they slow down a bit and solve the alignment issues. The instigator for that choice point is they see some evidence that their AI might be misaligned and plotting against them. Right. So if you actually see that evidence.

122
00:47:03.539 --> 00:47:30.826
then it's like, oh gosh, maybe we shouldn't put it in charge of everything and let it rip, you know, because that's evidence is staring us right in the face that it's untrustworthy, you know. But if they don't see that sort of very clear evidence, then I think they're going to convince themselves that they need to keep going, you know. But maybe they will see very clear evidence like that, in which case, even if we don't have regulation, they might just sort of voluntarily stop. So that's the second ray of hope.

123
00:47:31.197 --> 00:47:42.115
Like overall, I don't think that we're like definitely doomed, you know, like I said, 70%, but like I could see it working out pretty well as well. Hmm. What about jobs?

124
00:47:42.284 --> 00:48:11.005
Yeah. So I think I'm excited to at some point get into the new thing, which is the more optimistic, positive vision. And that will have a lot to say about this. Because in the prediction, you know, in AI-2027, by the time everyone loses their jobs, there are worse things happening. Or like, it's kind of like too late by that point. But yes, like once, if, I mean, just think about it. If the companies do manage to build superintelligence, then by definition,

125
00:48:11.005 --> 00:48:34.529
they're going to be able to take almost all the jobs or all the jobs, right? Because it's better, faster, cheaper than the best humans at everything. And that's, again, the timeline is by the end of sort of 2030, you reckon you think superintelligence might arrive. I'm trying to think about when we can start to see job displacement in the economy. We're already starting to see a little bit of it now, but not very much. Why? Because the AI's aren't good enough yet. Like, they're impressive, but they're not like...

126
00:48:35.592 --> 00:48:51.708
They're not just a drop-in replacement for a human worker in almost any field. And do you think that'll be sudden? I think it'll be sudden because of the intelligence explosion dynamics or recursive self-improvement dynamics. So you could imagine a different world where...

127
00:48:52.248 --> 00:49:20.851
It's gradual. And this is maybe how it is in a lot of science fiction, is the AIs gradually get better at a bunch of things, and they gradually automate this one industry, like pharma, and then they automate steering drones, then they automate driving cars or something like that. But what's different about the real world is that the companies have converged on this strategy of automating themselves first, automating the AI research process,

128
00:49:21.813 --> 00:49:26.032
And so if they're allowed to continue with the strategy,

129
00:49:26.943 --> 00:49:54.263
We're not going to see the robo-taxis and the plumber robots and the lawyer AIs. We're not going to see that sort of broad diffusion of AI into the economy happening first, because that's not what they're focusing on first. They're focusing on automating themselves, automating their own research, so that they can do everything that they're doing faster. And they want that to sort of get going and get to...

130
00:49:54.263 --> 00:50:21.280
very high levels of intelligence, very high levels of general intelligence, and then deploy more out into the economy, right? So by the time it's actually coming for, like, all these different jobs, they will have had fully autonomous AI research happening for months, maybe years, you know? And that means that, like, the AIs will be vastly superhuman at AI research and probably also vastly superhuman at lots of other things just as a side effect, you know? Yeah.

131
00:50:22.124 --> 00:50:31.608
If you're wondering what this looks like, well, we wrote about what it looks like. It's sort of like this wave smashing through the economy after they do the intelligence explosion internally.

132
00:50:31.962 --> 00:51:00.363
What I'm hearing there is that because the AI will be able to improve itself and train itself, it'll be getting better at everything at once, and then it will be released at kind of once. Is that accurate? Yes, but it's not even exactly that, because even if it's mostly just getting better at the things that it's doing, like research, that'll have some spillover effects to other skills as well. And then when it turns to focusing on those other skills, it'll be able to do them very fast. What jobs remain in such a scenario, do you think?

133
00:51:00.818 --> 00:51:22.790
I think that's actually a political question, not a technical question. Because? Because on a technical level, all the jobs can be done by the AIs if they've reached that level. And so it's a question of what jobs are allowed for them to do. And what kind of jobs wouldn't be allowed, do you think? That depends on who's in charge. So there'd be some sort of political conversation about what we're going to allow and disallow.

134
00:51:22.925 --> 00:51:39.007
I mean, in this scenario, the humans are still controlling them, the AIs. Depends on what you mean by control, right? So there's like, do the AIs actually have the goals and values that you want them to have? And are they going to robustly do that and behave as intended into the future? And then there's like, are they obeying your orders for now? Yeah.

135
00:51:39.007 --> 00:52:05.804
Are they obeying the orders is really what I'm saying. Yeah. So like even in AI-2027, in the scenario where the AIs take over and kill everyone, there's a period of like several years where they're still obeying orders. And they're, you know, taking some jobs, but not other jobs. And they're helping to make better weapons that the US government can use to like do its arms race with China and so forth. And that's why they're able to get so much power so quickly is because

136
00:52:05.804 --> 00:52:22.072
The government and the corporations and so forth trust them and is deliberately deploying them into all of these positions because it thinks that things are fine. But because these things are neural nets, you can't just look inside and see what it's really thinking. You can't really tell.

137
00:52:22.257 --> 00:52:51.721
I think this is a really important point, because unlike software, where we can look at the code and see what's going on theoretically, with AI, you're saying that we don't know why it's making the decisions that it's making, because we can't get inside. One note of optimism is that it doesn't necessarily have to be that way. Like, there's a subfield of machine learning called mechanistic interpretability, and a broader subfield called interpretability more generally, that's trying to solve that problem, and trying to take these trained artificial neural nets, and piece them apart and understand

138
00:52:51.721 --> 00:53:16.696
like how the information is flowing and how the decisions are being made, so to speak. The problem is just, it's a very inherently hard problem. If you have 10 trillion connections to look at, you can look at any particular group of them and be like, okay, so this is how like this particular connection works. But like, how do you get a sense of the whole, you know, how do you get a sense of like what's happening at a high level? And the answer is, well, it might be impossible, but people are working on it and they are making progress. And

139
00:53:16.696 --> 00:53:36.068
If they can make enough progress, then we're in a very different and much brighter world. I think that it would be much less likely for us to get into those loss of control scenarios if we could just actually see what our AIs were thinking and why and how at any given time. Right? Yeah. So we would still have the other problems to worry about, but at least we could mostly solve that one.

140
00:53:36.136 --> 00:53:59.997
It is pretty crazy to think that we're building a technology, a brain, that we don't understand. Yeah, it's pretty crazy. I mean, it's one of those things where like... In a movie, like a sci-fi movie, a bunch of scientists stood around this big brain and they're all just like, they're making it more, they're feeding it. Yeah. But they don't really know what the fuck. Yeah, I mean, it's kind of just like obviously a dangerous thing to be doing. Yeah. But we're doing it anyway because of this history of how the field has developed in the last 10 years, where...

141
00:53:59.997 --> 00:54:20.973
You know, people were like, oh, wow, yeah, that's obviously dangerous. Oh, no, what if someone else did it and did a bad job of it? Therefore, we should do it and do a good job of it. And now they're in this race where they're racing each other. And they're also under all sorts of political pressure to pretend that it's not as bad as it seems because they don't want to anger their investors. They don't want to anger the White House.

142
00:54:21.428 --> 00:54:44.480
One of the key questions we had from our audience was which, and I kind of asked you this in part, but which jobs are genuinely likely to survive AI and what skills should people slash students focus on over the next 10 years? That's kind of like, like imagine if you were someone living in Mexico in like 1500 and then you hear that like the conquistadors are coming.

143
00:54:45.762 --> 00:54:58.739
You could be asking yourself, like, okay, well, what sort of job should I be switching to, to, like, survive this transition? But, like, you have a lot more to worry about besides that. But yes, I think I would say that, like, if we manage to avoid the loss of control problem,

144
00:54:59.650 --> 00:55:22.482
and we end up with humans still in charge of the AIs and humans can like say what the AI's goals and values are supposed to be, even as they become much smarter than humans and even as they run the whole economy, then probably there will be regulation that protects some areas. And you can try to guess at what those areas might be. Maybe stuff that's more like judges, potentially. What about podcasters?

145
00:55:24.271 --> 00:55:50.478
I'll be honest. Probably not podcasters, I think. Stuff like, you know, being a nanny, maybe, right? Like, I think that even if there's a robot nanny that's like really, really good, I think a bunch of people might prefer to have an actual human because they might be creeped out by the idea of a really good robot nanny. So you can sort of reason like that. There's also like stuff that might be legally protected. Like maybe judges, for example, like are going to be legally required to be humans and not robots.

146
00:55:50.478 --> 00:56:00.113
Some people say, though, there's going to be so many jobs created that we can't foresee right now, like there was in the Industrial Revolution or the Internet boom or whatever. The problem with that is that...

147
00:56:00.181 --> 00:56:27.198
past technological advancements have been more narrow. They've like automated some things, but not everything. But we are talking about a hypothetical future situation in which everything gets automated. So there isn't any new job that you could do that the AI couldn't also do, except if it's like protected by regulation or something. That's also a thing. But so like, for example, right now there's this sort of like cycle where...

148
00:56:27.198 --> 00:56:42.807
You know, the AIs learned to do a certain thing, like write copy or like draft code or like debug something. And then humans who used to do that thing switch to managing AIs or switch to doing the other stuff that the AIs can't do.

149
00:56:42.807 --> 00:57:00.104
And that's why there's been this dynamic historically of, you know, new jobs opening up and people flooding to them. But if it gets to the point where the AIs can do everything that humans can do and better and faster and cheaper, then whatever that new job is that you might have switched to, the AIs can switch to that too. And they'll already be better at it than you.

150
00:57:00.931 --> 00:57:24.438
Because we haven't seen widespread unemployment yet in the economy, do you think people are getting a little bit complacent? Because what I'm seeing on my timeline is a lot of people saying, I told you so. I told you everything would be fine. And when you look at the US unemployment rate, currently it's flat to slightly down. If you look at the UK, it is up. The trend is up compared to last year. We're at about 5% unemployment. The US is at 4.2% unemployment.

151
00:57:24.438 --> 00:57:50.392
Yeah. Basically, nobody has said that there would be mass unemployment by now, or at least we didn't say that. And we were historically one of the more bullish people on AI progress. In AI-2027, because of the dynamics that we just described, the mass unemployment doesn't happen until 2028 or 2029 after they already have superintelligence. Because, again, the companies aren't trying to cause mass unemployment as step one. That's like step three after, you know, it's like step one, automate themselves.

152
00:57:50.392 --> 00:58:04.128
Step two, have this recursive self-improvement to get to superintelligence. Step three, expand out into the economy and out of everything. And so this is really unfortunate from humanity's perspective because one might have hoped that

153
00:58:04.668 --> 00:58:30.082
If there was this broad wave of automation going through the economy, people would sit up and pay attention and think about where all this is headed and demand good regulations from the government. But that's not actually what the strategy the companies are taking. You know, they're going to be getting the superintelligence first and then doing the broad wave of automation, which means that by the time they're actually doing all of that, well, it's already going to be moving very fast and the AI is already going to be very powerful.

154
00:58:30.082 --> 00:58:42.569
In your 2027 report, so you wrote that in 2025, but it is called AI 2027. You said that in mid 2025, we'd have the autonomous employee, which is sort of like AI agents taking instructions over Slack or Teams.

155
00:58:44.138 --> 00:59:01.368
That happened. I've actually got an AI agent in my WhatsApp, which I talked to, of course. Yeah, I've got Clawbot exploded, obviously around the world. And now, you know, Clawd have talked about their new Slack integration, but lots of people are using agents now. And that happened, I'd say for us at the time, we really sort of caught onto it at the start of 2026.

156
00:59:01.368 --> 00:59:18.614
You also said by 2026, companies begin replacing entire corporate departments with AI agent subscriptions. 2027, the final job. AI automates the job of the human AI researchers themselves and begins the machine learning research to upgrade and build the next generation of AIs. Yeah, yeah. So again, timelines.

157
00:59:18.817 --> 00:59:46.643
We are uncertain about how long it will take to achieve these milestones. In this scenario, they happen at those times. But by the time we had actually published the scenario, our timelines had shifted back a little bit, specifically mine had. So, like, my 50% mark was 2028 for the full automation of AI research milestone, not 2027. And then other people on my team had more like 2030, 2031, things like that. So I kind of want to, like...

158
00:59:47.453 --> 01:00:08.682
maybe try to illustrate this with, you know, we have like this probability distribution. It's like a smeared out probability mass. And like the 50% mark is this particular year, but there's like a lot of possibility that it happens years earlier or years later, right? What is this AI 2040? So AI 2047 was a best guess prediction as to how things would actually go. Yeah.

159
01:00:08.682 --> 01:00:29.472
AI 2040 Plan A is our recommendation for how things should go. So we called it AI 2040 because in this scenario, they build superintelligence in 2040 instead of much sooner because they delay things. Why do they delay things? To manage the risks and make sure that power is distributed equitably. They basically like...

160
01:00:30.637 --> 01:00:45.757
regulate AI development so that it still continues, but at a slower and more reasonable pace in a more transparent and safe way and spread out over more countries and companies. And as a result, they get to superintelligence in 2040 instead of in, say, 2030.

161
01:00:46.297 --> 01:01:06.850
And then we call it plan A because, well, it's our recommendation. Like we've come up with a plan for what government should do. And the scenario is an illustration of what it might look like to implement that plan. In a similar way to how AI-2027 is kind of an illustration of what it might look like to do what the companies are currently planning to do. If that makes sense.

162
01:01:07.711 --> 01:01:28.248
And is this wishful thinking, or is this what you think is going to happen? No, it's definitely not what we think is going to happen. It's not what you think is going to happen? No, no. What we think is going to happen is still something more like this, right? We don't expect the world to listen to us, right? This is our recommendation. But we hope that people do something like this, and we think it's possible. But it's not our, like...

163
01:01:28.248 --> 01:01:44.195
prediction for what's going to happen by default, you know? So I do want to run through the plans, the potential plans, and also plan A. But just to close off on how things might look after the year, because I think I wanted to touch on robotics too, and I've got this graph here, which talks about share of labor output.

164
01:01:44.195 --> 01:02:11.482
Yeah. Which I found to be quite striking. I've been sat here wondering, as an employer who employs hundreds and hundreds of people, when all this stuff is going to happen. And, you know, we're still hiring more people as things stand. There are some roles where our consideration is changing, shifting considerably. And I'd have to say that, you know, we're probably in the phase where our teams are AI powered and they're using agents to do some of their work now. But I'm wondering as an employer, like, when is it, when does this happen? Yeah.

165
01:02:11.701 --> 01:02:26.720
Yeah, great question. So if we could maybe zoom in on this a little bit. Yeah, we'll put it on the screen. So this is in the AI 2040 Plan A scenario. And notably, in that scenario, there's significant regulation introduced in 2029 that slows down the pace of AI development.

166
01:02:26.720 --> 01:02:48.235
In the scenario, they do that sort of at the last moment. So in the scenario, if they hadn't done that, then it was about to take off similar to how it does in AI 20.07. But as you can see, like in the scenario, there's still a bunch of jobs at the point that they implement it. And this gets back to what I was saying earlier, is that if you wait until most people have lost their jobs...

167
01:02:48.235 --> 01:03:05.954
to regulate the AI companies, that's already too late because they will probably already have super intelligent AI by then because their strategy is to first get super intelligent AI and do all that stuff. And I think you say that it would collapse the economy and cause even more harm to suddenly regulate something that all of us and all of our lives were then at that point relying on.

168
01:03:06.207 --> 01:03:35.519
Oh, but it's a risk well worth taking. I mean, it's true that right now a lot of people use AI for a lot of things, but if we could somehow slow or halt AI development now to set up a better way to do it, that would be well worth it, even though there would be significant costs. But you can't over here, right, can you? At this point where AI and robotics are doing most of the labor output. That's right. But in this scenario, in the AI 2040 Plan A scenario, they put in the regulations in 2029, and then they slowly and carefully develop AI

169
01:03:35.519 --> 01:04:02.907
in a way that avoids all the problems, which we can get into in a little bit. And so eventually, yes, eventually the AIs take the jobs. Eventually basically the whole economy is run by AIs and robots, but it happens gradually over the course of the 2030s instead of happening in this sort of crazy shock, you know, a year later. Right. Because in this scenario, they don't let the companies recursively self-improve and get to super intelligence as fast as possible.

170
01:04:02.907 --> 01:04:30.464
Instead, they regulate AI development so that the core capabilities of the AIs are improving at a more reasonable pace and also in a more transparent way so that the scientific community can see what's going on and help make it safe. But it's, I guess I noticed here that in both your scenarios, eventually AI and robotics do pretty much all the jobs. Yes. So you kind of side there with Elon when Elon says that working will be a choice. Yeah.

171
01:04:32.911 --> 01:05:00.198
Yes. I mean, by definition, if it can do all the things, then it can do all the things. I think that there's a question of like, should we allow there to be AIs that can do all the things, right? Some people think that the answer is no, and we should just shut it all down and prevent these types of AIs from being created in the first place. And we're actually kind of sympathetic to that. We have our, should we bring up the plans diagram? Yeah. Thanks. Yeah. So,

172
01:05:01.615 --> 01:05:21.275
Our scenario is called AI2040 Plan A. It's a scenario in which they slow down AI development to make a superintelligence happen in 2040 instead of earlier. And Plan A is our recommendation, so this is sort of illustrating our recommendation. But for comparison, we made like mini scenarios illustrating different alternative plans, which we call Plan S, Plan B, Plan C, and Plan D. Okay.

173
01:05:21.275 --> 01:05:44.714
Plan D is basically the same thing that happens in AI-2027. The race continues. There's very little regulation. You can read about that in AI-2027. Plan C, also very similar to what happens in the slowdown ending of AI-2027, where they solve the alignment problems. So in that ending, they slow down a little bit, pivot more resources to AI alignment and AI safety research,

174
01:05:45.136 --> 01:05:59.598
get lucky and succeed. And now they have aligned AIs. And then they speed up again and take all the jobs and beat China and all those things. Plan B is... It's kind of like plan C in that...

175
01:05:59.986 --> 01:06:27.863
Well, basically in plan B, you're being more aggressive towards China and you're taking actions to sabotage or cyber attack them to keep them behind so that you have more breathing room to solve the alignment problems yourself. Plan A is our recommendation. It's domestic regulation and then an international deal to continue building AI, but in a much better way. Plan S is shut it all down. If you want to have a future where...

176
01:06:28.741 --> 01:06:57.800
There aren't AIs running around that can do everything better and faster than humans. You kind of want something like Plan S. What do you want? Plan A is our recommendation. I think that I'm sympathetic to Plan S, but for reasons we explain, we recommend Plan A instead. And what do you think is most probable, if you're being honest? Plan D. Which is that they just carry out. Just a 20.07 type of thing where they keep racing. They don't really slow down significantly. And things happen extremely fast.

177
01:06:58.272 --> 01:07:28.124
The diagram sort of explains like roughly the reasoning behind this too. So like there's this high level thing of like, do you want to keep racing as fast as possible to make the AI smarter and smarter to put them in charge of more things so that we can be China? You know, if you're happy with that, then you can get down and says, this variation of Op-Ins here. If you are worried about that, well, you get to something like this. There's more different options besides these, but this is kind of like the ones that we could compress onto a screen.

178
01:07:29.980 --> 01:07:59.478
Do you have children? Yeah, I have two children. It's kind of sad. Like, I think that one way or another this will probably all be over by the time they're old enough to join the workforce. So I don't think they'll ever join the workforce. When you say this will be all over by the time they join the workforce, what do you mean by this will be all over? So these milestones that I described, like AI is automating the AI research, AI is getting super intelligent, um,

179
01:08:00.608 --> 01:08:21.837
AI is then exploding out into the economy, taking the jobs, building robot factories to build more robots, to build more factories, etc. GDP starting to go vertical. That sort of thing is what I mean. Like all of those events transpiring. Maybe there's like 10-20% chance or something that hits a wall and none of this comes to pass, even if you don't do anything.

180
01:08:23.457 --> 01:08:47.791
How old's your oldest? Six. Six. Boy, girl? Girl. Girl. So your daughter comes to you and says, Dad, what should I study in school? I mean, again, if these radical transformations happen, then the world would just look completely different. And what sort of jobs you set yourself up for basically won't matter that much, probably. I would say that the thing to do is...

181
01:08:47.791 --> 01:09:13.424
will A, try to make it actually go well. Like if you can exert any influence at all on history and how this all develops, you should be trying very hard to steer the future in better directions. And then separately from that, on a personal level, you should focus on, well, being a good person and doing things that are sort of good for their own sake rather than good because they'll set you up for later employment because that later employment is going to be very uncertain, basically.

182
01:09:13.913 --> 01:09:37.876
Elon talks about this age of abundance we're heading towards. Age of abundance. There'll definitely be abundance. The question is who controls the abundance and what do they do with it? Right? Are the AIs controlled by anyone or are they doing their own thing? And then if they are controlled by people, who controls them and what do they do? And what's the sort of like political structure governing how they make those decisions?

183
01:09:38.062 --> 01:10:03.712
I think it was Geoffrey Hinton that said to me, he said, there's no example in nature where a more intelligent species has less control than a less intelligent species. Thus saying that we're quite arrogant to think that in a world where there's this artificial brain that's a gazillion times the size of mine, that I'm going to give it orders. Yeah. I mean, that's the thing is I think it's like,

184
01:10:03.914 --> 01:10:29.632
That should be our default assumption is that like, well, there's these brains. We can't see exactly what they're thinking. We're going to make them smarter than us and put them in charge of everything. And then we're going to give them bodies. Yeah. And then they're going to be autonomously building new factors and so forth. And like, how is this supposed to end well again? Like, isn't this just exactly like us picking a new species that's then going to outcompete us when it doesn't need us anymore? Like, I think that is just the default trajectory. Yeah.

185
01:10:29.632 --> 01:10:56.632
Now, there's a whole argument we can get into about ways that we could get off of that default trajectory. So, for example, there's research into interpretability that I described previously. And if that research bears fruit, then you will be able to actually see what they're thinking. And then that would be an excellent tool for shaping them and controlling them and making sure that they do what we want, right? There's other sorts of AI alignment research agendas that are making progress. And if enough of those agendas succeed sufficiently, we can avoid this problem.

186
01:10:56.632 --> 01:11:22.973
Of course, also there's the regulatory side too, where part of what makes this difficult is that we're building these AIs in race conditions. The companies are secretive about their recipes for making these AIs because it's secrets that they want to protect so that other people can't copy them. And so a lot of it is happening behind closed doors. Only a few people can really see the recipes that they're using to train these AIs and so forth. And then oftentimes when the AIs...

187
01:11:22.973 --> 01:11:52.943
behave in unexpected ways or even just like blatantly misaligned ways, sometimes that information doesn't really flow out to the public because the companies are not really incentivized to tell everyone about how they messed up and how their AI is evil. It's just not very conducive to scientific progress on these issues. If the regulatory system was different, then perhaps we could be in a better situation, make faster progress. Also, of course, we wouldn't be planning to put these AIs in charge of everything as fast as possible. And we wouldn't be planning to like let themselves improve.

188
01:11:52.943 --> 01:12:09.987
not make, you know. I don't speak Vietnamese, but this show can because of AI video technology from our sponsor, Hey Gen. I get messages every single week from those of you listening to The Diary of a CEO all around the world and you express how much impact it's had on you and your life.

189
01:12:09.987 --> 01:12:38.050
If that's true, then those conversations shouldn't only reach people in English. Hey Jen can take one recording of me and deliver it in any language while keeping my voice, timing and expressions intact. But you don't need a studio like this to make it work for you. Record 15 seconds of yourself and get an AI avatar that delivers studio quality video in over 175 languages. We're up to 20 languages now and we're not the only ones using it. Hey Jen is already used by 30

190
01:12:38.050 --> 01:12:57.693
million people, including 85% of the Fortune 100. Whether you're building an audience on social media, launching an online course, or rolling out training across your team, check out HeyGen now. Your first three videos are totally free at heygen.com slash doac. That's H-E-Y-G-E-N.com slash doac. See you there.

191
01:12:58.840 --> 01:13:26.060
Ilya was, as you said, he was one of the leaders at OpenAI and he left and he started his own company now, Safe Superintelligence. Very curious name of a company, Safe Superintelligence after leaving OpenAI. Did you ever get to work with him? I wasn't directly working with him. I had a couple of chats with him. Do you think he's genuinely concerned as well? I think he is, but I think he's similar to these other CEOs where, I mean, just think about the sort of incentives that they're under, right? Like,

192
01:13:27.578 --> 01:13:46.782
They can sort of see the problem and then they can be like, okay, but like if I don't, if I stop, if I quit my job and or do something else, that's not going to solve the problem because the other CEOs are going to keep going. And even if all of us didn't go, then maybe China would keep going. So like, man, it seems like this is just going to happen one way or another, whether I do anything about it or not.

193
01:13:48.014 --> 01:14:09.226
I guess I should be involved, you know, and like maybe I can make it go well. And at any rate, like I don't want to be out in the cold while these other people I don't trust are in charge of everything. So they all sort of like reason through all of this and then convince themselves that like the thing to do is for them to build it and to do it better. And I think Ilya is just the latest example of this. Elon's another example. Dario is another example.

194
01:14:09.226 --> 01:14:25.882
You know, arguably OpenAI at the beginning, Sam, was an example, although like Elon and Dario were at OpenAI early on. So what do you think they should all do then? So I think what should happen is some sort of international regulation or at least domestic regulation similar to what we described in Plan A. Okay, so walk me through Plan A. Yeah.

195
01:14:25.882 --> 01:14:50.283
So in this scenario, AI takes longer to get to recursive self-improvement and full automation of AI research than it does in AI 20.07. We figured that we should try to illustrate a range of different possibilities because we do have those sort of uncertainty intervals. So we chose 20.30 as the moment when full automation would finally be achieved and things would really kick off. And then working backwards from that...

196
01:14:51.008 --> 01:15:18.785
When's the last moment you could really have good regulation? 2029. So in this scenario, AI progress slows down a little bit naturally, and the AI companies keep racing, but they don't quite succeed in automating themselves in 2027 or in 2028 or in 2029, but they're getting really close and they're going to do it in 2030. And then in 2029, the government steps in and regulates them. What regulations do they do? Well, they basically just shut it down temporarily. Can I ask...

197
01:15:19.527 --> 01:15:46.595
How does the elections overlay with your timeframes here? Because there's going to be a big election, isn't there, in 2028? And it seems now that sentiment has really, really turned against AI in the general public, and that it will be one of the big ticket items on the ballot. We think that it'll be maybe the most important issue in the presidential election in 2028. I think a lot of people, most people, will be quite concerned about where things are headed. And that's part of why we chose...

198
01:15:46.595 --> 01:16:14.489
to depict things the way they were doing in this scenario, because that helps explain why they might do this sort of regulation in 2029, is that the voters have been demanding it and the presidential candidates have been promising it. And in this scenario, and in 2027, would the general public have felt the consequences of AI much more severely than they have now by then? Yes. Although still, even in 2029 in this scenario, they still mostly have the jobs as depicted here, right? So in 2029 in this scenario, lots of jobs now involve managing AI agents. Right.

199
01:16:14.489 --> 01:16:41.388
You mentioned you have an AI agent, right? Well, in 2029, in this scenario, the AI agents will be much better. Still, though, not enough to just completely do everything. That was the sort of thing that would come in 2030 in this timeline. Again, we're uncertain about timelines. Things could go faster than depicted in this scenario. And in fact, I think things probably will go a bit faster than depicted in this scenario. But we're uncertain. We already did the very fast timeline scenario. So now we're doing a slower timeline scenario. But maybe we should talk about the high-level goals. So...

200
01:16:42.839 --> 01:16:59.883
They want to have AI continue, but in a slower pace. Who's that? So they can make it safe. The politicians, you know, the president and the people who voted for the president and, you know, the heads of other governments and so forth. So goal one, slow things down. Goal two, make it more transparent.

201
01:16:59.883 --> 01:17:18.648
So that the scientific community can catch up to this stuff and make more progress. And also so that we don't have to take the company's word for it when they say that their systems are safe. And when they say that they haven't, you know, put in any biases into their systems, for example. That's a constitution of power issue. We also want to avoid a situation where there's an intense concentration of power.

202
01:17:18.648 --> 01:17:48.584
So in addition to the transparency and the slowdown, we actually think it's actively good for there to be multiple AI companies across multiple different countries that have similar levels of very advanced AI capability and for there to be like broad diffusion of AI into society rather than, you know, a single mega project that has all the best AIs, for example. And the nice thing about that is you kind of get that by default if you do the first two things. If you slow it down and if you make it more transparent, then that means there's breathing room

203
01:17:48.584 --> 01:18:18.571
for other projects to sort of catch up, right? And the transparency just like literally helps them catch up because then they can like copy some of the ideas. And then I think the fourth thing would be reversibility. So in what follows in this scenario, we are going to be building up a lot of data centers, a lot of robots. We're going to be transforming the world at a sort of like slower pace, though still a very fast pace, but slower. And if things go wrong and the deal breaks down and everyone starts racing each other again to get to superintelligence as fast as possible,

204
01:18:18.571 --> 01:18:37.893
that would be very scary. And so the fourth principle is basically build the new data centers in such a way that if everything breaks down and everyone starts racing again, the newly built data centers get destroyed so that we're sort of back to square one again instead of in an even worse race where there's even more AI's and robots and compute everywhere.

205
01:18:37.893 --> 01:18:54.110
Um, so I can sort of walk you through the timeline if you're interested, or the president talks to China, talks to the leaders of a bunch of other countries and says, we're going to basically halt AI development until we can figure out a plan for how to do it in ways that achieve these goals. Um,

206
01:18:54.110 --> 01:19:15.153
So they basically send inspectors to each other's data centers. Like Chinese inspectors come to U.S. data centers, U.S. inspectors go to Chinese data centers and verify that they are doing inference and not training. Developing new AIs, that involves training them. But just taking existing AIs and using them to serve customers, that's called inference. Yeah.

207
01:19:15.153 --> 01:19:35.082
And so the sort of like solution they come up with here in this scenario is we'll allow them to keep doing inference but not training for now until we can get the new training data center set up. So they retrofit the existing data centers to serve inference. People can still keep talking to their agents, but they're going to stop getting better and better.

208
01:19:35.082 --> 01:19:57.003
for like six months to a year while they build the new data centers that are going to be the transparent data centers. And that's where the training is going to happen. Once they get those new data centers set up in 2030, then AI research continues. This is a bit spicy. We advocate for total research transparency, which means that on the training data centers that are training the new models, they basically have to publish everything.

209
01:19:57.003 --> 01:20:24.931
which means you get to see all the details of the recipes for training these models. You get to see the architectures, et cetera. We think that sort of open science is really important for solving the alignment problem fast enough because you don't want to have these sort of biased companies making the decisions about whether the AIs are safe. And we also think it's important for just good regulations more generally because right now most of the expertise in the world on AI is sort of concentrated in Silicon Valley and the governments in particular kind of

210
01:20:24.931 --> 01:20:54.918
don't really understand AI that well. And imagine an alternative. Instead of total research transparency, you had like an auditor system where the government says, here are some rules for how to make the AI safe. And then we're going to have like an agency that like goes into the companies and ask them questions and tries to make sure that they're following the rules. That creates this sort of adversarial dynamic where the company is incentivized to like fool the regulator, you know? And also if they discover some new problem that's not even on the government's radar,

211
01:20:54.918 --> 01:21:15.303
there may be incentivized to like not tell the government about it. Yeah. Right. So if you have the total transparency, it helps the government make better decisions faster. But it kills that competitive advantage. Yes. Dropbox is not going to like this. You know, OpenAI is not going to like this. This would be probably bad for the valuations. I don't think it would kill them completely, but it means that it would commoditize more.

212
01:21:15.303 --> 01:21:38.759
Right. So it means that there'd be like a bunch of AI companies that would catch up to the frontier. They would train AIs that are like roughly similar, roughly equivalent. They could still make money by doing that and then selling their AIs, but they wouldn't have a monopoly. They wouldn't have anything close to a monopoly, which I think is good for humanity, although it's bad for the bottom line of those particular companies. Notably, it's good for the bottom line of lots of other companies. Like if you're a company that's behind.

213
01:21:38.759 --> 01:21:57.271
and you don't, you're not anthropic, you're not open AI, then you would love this because this helps you catch up, you know, or this helps you to, like, capture more of the value from the chips you're selling, for example, or from the, like, downstream product that you're making. And by 2031, then you have one-fifth of all cognitive labor done by AI that's done by AI.

214
01:21:57.271 --> 01:22:25.199
Yeah, so what's happening here is that we're imagining that the government of the United States and the government of these other countries that are involved in this agreement that are sort of implementing similar regulations, they don't have to be exactly the same. But that's another thing that's nice about the transparency is that if you have this sort of transparency, then if two governments are implementing different regulations, like if one of them is like telling their companies to go slower or like banning more stuff than the other one is, they can both see that.

215
01:22:25.199 --> 01:22:40.032
Like, oh, you're letting them do that sort of thing, and you're not? Like, maybe we should let them do this too, you know? So it helps to sort of naturally equalize the regulations to some extent without having there to be a central power that just gets to make regulations for everybody.

216
01:22:40.285 --> 01:23:06.644
So anyhow, we're imagining that when they get this transparency set up, they basically agree to ban the dangerous stuff, to allow the not-so-dangerous stuff. And there's a constant ongoing conversation about like, well, what's dangerous and what's not? What should we ban? What should we allow? What about this country? What about that country? That conversation evolves over time. But the gist of it is, at least if they do it the way that we recommend it, is that they don't do an intelligence explosion. They don't let the AI's, you know, autonomy to the self improve. Instead...

217
01:23:06.644 --> 01:23:23.266
They slowly and carefully scale up the AIs that they currently have and invest lots into finding ways to make them more interpretable, to make them more easy to control, to understand better how they work and so forth. The result is that AI progress continues, but it's not quite as fast.

218
01:23:23.266 --> 01:23:50.367
And it's much, much, much safer and more transparent. But still through these, we see job disruption. It is continuing because they are building more data centers, right? Like this whole time, they're building more and more data centers, more and more chips. And they're continuing to like make there be a larger and larger population of AIs, so to speak. And that causes this huge transformation over the course of the 2030s. Sort of a big thing that we sort of want people to take away is that even if you heavily restrict AI progress...

219
01:23:51.363 --> 01:24:06.753
you still get this sort of crazy transformation. In this scenario, they basically allow progress to continue, but at a slower, more safe pace here in 2030. And then as a result, it takes until 2035 to get to top expert level AI. So...

220
01:24:06.753 --> 01:24:24.252
Remember, they were on track to do that in 2030, but then sort of at the last moment they stopped. But because it was sort of so close to the last moment, that means that they can sort of get there pretty soon if they want to, and it's just a matter of how long they allow it to go, right? So they sort of slow it down, spread it out,

221
01:24:24.252 --> 01:24:44.654
leisurely arrive at this level after five years. By this point, they've built up massive amounts of data centers everywhere. So it's not just that the AIs are smarter and able to do all the things that humans can do, but also there's a lot more of them. And there's a lot of robots and so forth. So by this point, you kind of have the economy that a lot of people would have imagined with AGI.

222
01:24:44.654 --> 01:25:14.287
where there's AIs, there's lots of them, they're able to do all sorts of jobs. There's robots, there's lots of them, they're able to do all sorts of physical work. And basically, the economy is being run by these machines. So in 2031, you have the one-fifth of all cognitive labor done by AI. In 2023, you have 60 million AIs running at 100x speed. In 2033, there's cash dividends to all Americans. I've got to explain this to me.

223
01:25:15.232 --> 01:25:44.172
Yeah, so if the AI is going to be taking people's jobs, then it's very important that people not starve to death and still have money. And if companies are going to be using AI and robots to take all these jobs, then that means that there needs to be some sort of taxation scheme or something to make sure that people still have a slice of that pie. The pie is going to grow huge, but you still need to actually give people a slice of the pie. And our proposal for how to do that, we call it the citizen's dividend, basically...

224
01:25:44.172 --> 01:26:08.287
People have shares in an agency that sells permits to the robot companies and to the compute companies and makes profit from selling those permits. And then those are people have shares in that entity. It starts off small. It starts off something like $25,000 per person. And then by the end, it's something like $10 million per citizen. Per person? Per person per year. Yeah.

225
01:26:08.911 --> 01:26:35.962
Factoring in inflation? Yes, factoring in inflation. So we're all going to be multimillionaires? Yes, if this happens, which it probably won't. But if it happens, this is where we go. And again, this is the thing I want to emphasize is that if you get to the point where your AIs are close to being able to do all the research and then you sort of pause and slow down, that means that you still have a lot of transformation ahead of you. Because if you allow those AIs to still proceed slowly and start to automate various jobs and so forth,

226
01:26:35.962 --> 01:26:59.300
After some years, they will in fact have done that. And they will have built huge amounts of new data centers, huge amounts of new chip fabs, huge amounts of new robots, robot factories, etc. We're not sure obviously how fast this will go exactly, but we've thought about it a lot and we have our guesses. And this is sort of like our median guess. What does this mean, 2037, the apocalyptic arrival of truth on Earth? Yeah, so like...

227
01:26:59.300 --> 01:27:22.452
This is the point where we say they get to top expert level AI. So it's not super intelligence in the sense that it's not like vastly smarter than humans at things because they deliberately pause it at the level of top experts. So here they're going slow. Here they've just actually stopped. But they've stopped at a point where the AIs are just actually really good at everything. So kind of they've definitely got AGI. Maybe they got like weak super intelligence.

228
01:27:23.296 --> 01:27:53.030
Because they have so many of these AIs and because they think faster than humans, you know, they just run much faster, that's going to transform society dramatically. So we talk about some of the ways in which it transforms society. Like this is sort of life after work. We talk about what it would be like to be living on your citizen's dividend and not have a job anymore in this sort of world. Here we talk about all the scientific changes and all the social changes that would come from all of the intellectual progress and activity that would be generated by all of these AIs.

229
01:27:53.030 --> 01:28:09.162
So, for example, here is things like cancer cures and like, you know, people living in apartments that were built by robots two years ago. 2036. Providing again we stop in 2029. And providing, I mean, a conservative, this is a conservative time frame.

230
01:28:09.719 --> 01:28:32.247
Yeah, like, unfortunately, I actually think that things will happen faster than this by default. And that if we don't slow down, things will happen much faster than this. Once you get to the point where you've got, you know, a billion AIs running day and night, and they're each better than the best humans at everything. And so they're doing a lot of science. They're doing a lot of talking to each other. They're doing a lot of thinking. Everyone's constantly talking to their AI assistants and so forth.

231
01:28:32.602 --> 01:28:58.268
There's going to be a lot of scientific progress. There's going to be a lot of changes to politics, to ideologies. It's going to be very disruptive and crazy. And we get into some of the ways in which it is later, basically. I'm still not super clear on what this means, the apocalyptic arrival of truth on Earth. It's just because there's so many AIs that are so smart that they're uncovering, making new discoveries in sciences. Let me give you an example. Lie detectors. Yeah. So...

232
01:28:58.268 --> 01:29:23.108
That's an example of a technology that might be invented. Yeah. You know, right now we don't have good lie detectors. We have very bad lie detectors that like sort of work, but don't fully work. But once you've had these top expert level AIs thinking for many years at 100x human speed, and there's billions of them and they have access to robot factories to do research and stuff, they'll probably invent a ton of technologies. Maybe they'll invent lie detectors that actually work on real humans. Yeah.

233
01:29:23.108 --> 01:29:40.456
That'll have big social effects, right? Imagine a presidential candidate who's like, those allegations are false. And to prove them, I will go under a lie detector and say that they're false. I was just thinking about the whole justice system and how that would be overturned. In fact, you could, you know, theoretically walk down the street and be...

234
01:29:41.603 --> 01:30:10.493
It's both terrifying and exciting. One thing that we talk about in this section is, the invention of light detectors could be really bad. It could be that it enables a new form of totalitarianism, where the powerful people, the CEOs and politicians, force the people under them to go into light detectors and say, yes, I'm loyal to the dear leader. I would never do anything against the dear leader. And if you're lying, then you're in. And then if you're lying, you get fired. So there's a ton of very harmful uses of light detector technology.

235
01:30:10.493 --> 01:30:27.284
There's also the good uses. And broadly speaking, I would say the good uses are when lie detectors are used on the powerful instead of by the powerful. What's this? 2040 passing the torch to AIs. Yeah, great. So here they pause at the top expert AI level. And the reason why they pause is because

236
01:30:27.284 --> 01:30:56.613
their safety cases aren't good enough for going beyond that level. So in the sort of regulatory systems that they set up over the course of these years, roughly speaking, the way they would work is when you're making a new AI, and then when you're trying to deploy the AI into something, you have to have some sort of safety case explaining what your intentions are and why you think it's going to work the way that you want it to work. And in particular, why the AI is going to do as it's told, for example, and why nothing super terrible is going to happen, like AI takeover.

237
01:30:57.220 --> 01:31:03.160
It's relatively easy to make safety cases like this when your AIs are still not capable of automating everything.

238
01:31:04.409 --> 01:31:29.047
But the more powerful they get, the more difficult it is to actually argue that things are going to be fine. Because the AIs are just more capable and they can get up to more stuff. And if they're actually untrustworthy, the possible downsides are bigger. So that's why they stop at this level, is that they realize that if they keep going, then they might actually lose control of everything. But at the current level, they're convinced by safety cases that it's fine. But they don't want to go further. So they stop there.

239
01:31:29.401 --> 01:31:54.427
And then what happens in 2040 is they've made significant progress scientifically, including on alignment. And they figured out how to make AIs that are actually aligned in a robust way. With humans. With humans. So they can actually trust those AIs and they can allow them to become much smarter again. So that's why we call the whole thing AI 2040. Because in 2040, they sort of let off the brakes and allow the AIs to become significantly smarter than humans.

240
01:31:54.697 --> 01:32:17.360
I guess, you know, this is a plan and this is a hope. Yes. But in reality, this is not what you think probabilistically if you had to... That's right. It's important to distinguish, like, this is what we recommend, this is what we want to happen from, like, this is what we actually think will happen by default. Now, we do think it's possible for this to happen, but, you know, that will require a lot of people to sort of wake up...

241
01:32:17.360 --> 01:32:38.318
and pay more attention and advocate for something like this to happen. So our main scenario is mostly talking about the policy choices made and the broad-scale effects on society. We figured it would also be nice to accompany this with a little mini-scenario that describes what it would actually feel like to live through this from an ordinary person's perspective. Okay.

242
01:32:38.318 --> 01:33:00.931
So 2029, everyone's yelling at each other. The presidents are negotiating something. And they've paused AI, but you still have access to the existing AI. So it doesn't really feel that different, although it definitely is like something exciting happening. 2031, they've started progress again. The AIs are really smart. More people have lost their jobs. It's like really starting to actually affect things. But I think still most people have their jobs, but their jobs are sort of transformed. So like by 2031, it's like...

243
01:33:00.931 --> 01:33:17.317
Most white collar jobs involve working with AIs to a large extent, or managing teams of AIs or collaborating with them somehow. Also, there are some things like robo taxis that are basically just working. Citizens dividend, you know. Ideally, this would happen sooner. Like, in our scenario, they kind of do things at the last minute.

244
01:33:17.317 --> 01:33:38.393
You know, so like a lot of these policy things are like happening kind of like just in time. Obviously, we would recommend that you do them sooner and do a better job of them too. But so 2033, you start getting your checks from your dividend. So you're forecasting that there will be a citizen's check. Your model says it could be around 25,000 at the start per person. And then it would grow as the economy grows.

245
01:33:38.562 --> 01:34:07.317
But also as I guess job displacement takes hold, they're going to need to grow that check and make sure you can improve. And that's why it's kind of the last possible moment, because if you waited to implement this until like 2037, then like everyone would have already lost their jobs by the time that happens, right? People losing their jobs, especially if it happens quickly, like we see on this sort of graph here, is going to cause lots of problems in terms of civil unrest, social unrest, purpose, mental health, these kinds of things, theoretically. Yes. How do you think about that?

246
01:34:07.739 --> 01:34:30.723
it's going to be rough. And hopefully we can navigate that well. We think that at a high level, people need to have money and also people need to have power. And I think these are like somewhat different things. It's like, why are jobs important? Well, there's a lot of reasons why jobs are important, but I think the main ones are, well, it's how people get money so they can survive and get the things that they want by buying the things that they want. So if people are going to be losing their jobs, you need some other way of people getting money.

247
01:34:31.364 --> 01:34:44.442
And then there's also the power thing, which is that right now people have political power in part due to their economic power. People can threaten to go on strike, for example, or, you know, countries that are ruled by dictators can't.

248
01:34:45.843 --> 01:35:10.565
just completely genocide an entire subpopulation. Or they can, but it's costly for them to do so because then they'll have less money because that subpopulation is contributing to their economy and contributing tax revenue and so forth. But if you end up in a world where actually nobody's contributing tax revenue except for the AI companies and the robot companies, then you, the government, are less incentivized to care about what...

249
01:35:10.565 --> 01:35:38.830
you know, the common people think. So when people lose their jobs, they're not just threatened with loss of income, they're also threatened with loss of political power. And so we think that it's important to like do things to push against that. What does that look like? How do people have power in such a world? Well, in democracies, at least, they still have votes. Okay. So I think that it's very important for there to be regulations on the use of AI that help make

250
01:35:39.083 --> 01:36:07.889
the public discourse more sane and more actually giving the people what is in their interest and what they want and avoiding a sort of opposite outcome where, you know, the masses are easily manipulated by AI-powered media, for example, or where everyone's talking all day to their AI advisors and the AI advisors are like subtly steering them away from voting for the candidate that would

251
01:36:07.889 --> 01:36:37.876
not be what the AI companies want because the AI companies have this other candidate that they like better and they're like secretly biasing their AIs to like steer people towards voting for that candidate right so so we want to be in a situation where um people have AIs that are actually trustworthy and that are truth-seeking AIs honest AIs and that don't have any sort of like political agendas put into them by the AI companies or by the government you know you want to avoid a situation where the AI company where the government has issued some sort of secret order that like

252
01:36:37.876 --> 01:37:04.471
The AIs have to be such and such a way. You know, the Department of War dispute versus Anthropic is like an interesting sort of foreshadowing of this, right? Where Anthropic was giving their AIs to the Department of War. The Department of War wanted to use them for certain things and was upset that Anthropic's AIs were like not supposed to be used for those things. That things in particular were domestic surveillance and autonomous robots. Mm-hmm.

253
01:37:04.977 --> 01:37:32.939
There's going to be a lot more issues like that coming up, and you want it to be the case that people know what they're getting, and that if people are spending hours a day talking to their chatbot, that chatbot doesn't have political biases put into it or a secret agenda or things like that, and instead has been trained to give honest, true answers to things. And I think if you can do that, it can improve the discourse and help people to use their votes to put even better regulations and even better politicians in place and so forth. And you can sort of potentially bootstrap this to having something where people are

254
01:37:32.939 --> 01:38:01.306
people's power is even more secure than it is today. A lot of the stuff we've covered in part, so, you know, the wars and drones and missiles, we're already seeing this around the world at the moment, which is really, really interesting. And we've talked about robots outnumbering humans as well, which is part of this prediction. Some of the ones down here I found to be really curious, which is people will be protected by AIs wherever they go. Yeah. In this scenario, they delay the creation of superintelligence until 2040.

255
01:38:01.508 --> 01:38:22.518
And in fact, they paused from 2035. But then they let it go after that. And then they let the AIs become vastly super intelligent. And we think that once the AIs are vastly super intelligent, the world will transform even more radically than what happens in the 2030s in this scenario. So in the 2030s in this scenario, it's more like human level. You know, the AIs are not...

256
01:38:23.682 --> 01:38:39.275
They're doing the same sorts of things that human experts would have done. They're just doing it a bit better, a bit faster, and a lot cheaper. And there's a lot more of them. And the robots are still, you know, doing the same sorts of things that human workers would have done. There's just more of them, and they're cheaper. And because of exponential growth...

257
01:38:39.275 --> 01:39:05.549
you start with a world that looks not that different from today in 2029. And then by 2039, you end in a world that's radically transformed, where everyone's living in these fancy new apartments that were built by robots two years ago. There's giant special economic zones that are full of robots and solar panels and factories, producing more robots and solar panels and factories and so forth. Most of the economy is AIs and robots, and people don't have jobs anymore. That sort of transformation is what you get if you pause at human level.

258
01:39:06.342 --> 01:39:22.880
But if you go beyond the superintelligence, there's a whole nother transformation coming that's going to look more like magic. Think about how the technology of today would look like magic to someone from 500 years ago, you know? And that's without even like a qualitative improvement in intelligence, right? Like,

259
01:39:22.880 --> 01:39:50.943
The humans of today aren't like qualitatively smarter than the humans from 500 years ago. It's just that we've had more time to do research and we have more like money and resources to build, you know, prototypes and experiments and run experiments and so forth. But if you had a point where there were billions and billions of AIs that were not only faster than humans, but like qualitatively way, way, way better at everything, and in particular at doing scientific research, we should expect that some of the things that they develop will seem like magic to us.

260
01:39:50.943 --> 01:40:19.040
And we'll just completely, like, we did not think that was even possible, you know? People don't want to die. People don't want to be hit by cars. People don't want to be, like, attacked by a random mass murderer. Cancer's gone? I mean, not just cancer. Like, you know, a lot of the stuff that happens in science fiction will probably have happened by then. So things like people scanning their brains and uploading into computers, right? Or self-replicating robots in the asteroid belt, creating more and more satellites to...

261
01:40:20.052 --> 01:40:47.086
produce more and more power, to produce more and more self-replicating robots and so forth. Most people still live on Earth, but the trend is to move to space. That's right. Yeah. So like if you end up in a situation where the entire human economy is just like a tiny drop in the bucket that is the entire economy, and it's just like huge amounts of robots and AIs that are moving incredibly quickly, then what you want is Earth to be

262
01:40:47.086 --> 01:41:14.272
mostly left as something like a preserve. You know, I think a lot of people are worried about the environment being destroyed, which it totally would be if it wasn't protected. And, you know, there's a lot of people who sort of like their lives as it is and don't want to be uploaded or live in some crazy new future thing. And it seems to us like the reasonable solution to these issues is create new living spaces off the planet with some of that vast...

263
01:41:14.272 --> 01:41:35.551
economic wealth and activity that's happening for the people who want that sort of thing. And then that way the earth can be preserved. Data center, picture here of data centers in the ocean. I mean, there's three images there of different environments where humans might live. Again, like our proposal was you preserve like 99% of the earth and

264
01:41:35.551 --> 01:42:00.796
mostly as is, as historic or environmental, for historic or environmental reasons. But then like some parts of it you designate as special economic zones where the robots can go crazy and dig giant pit mines and produce factories and so forth. We were thinking it would be good to build the data centers on the ocean instead of on land for a variety of reasons. Although later space would be better and we could see that being reasonable as well.

265
01:42:00.796 --> 01:42:27.425
What about immortality in a world of AI? Well, 3045, you say you've lived a dozen lifetimes and are immortal, passing from life to life as if by reincarnation. I mean, there's a lot of billionaires at the moment that are focused on longevity. I mean, Brian Johnson said he's got this central rule, which is do not die right now because we're in the age of AI. And it's conceivable that with superintelligence, we'll be able to choose when we die.

266
01:42:27.425 --> 01:42:42.376
Yep, I think that's probably right. We don't depict that happening in this part because at this part they only have, you know, human level AIs. But that's one of those things that seems quite plausible that superintelligence could achieve through a variety of means.

267
01:42:45.599 --> 01:43:05.579
What is your hope with all of this stuff? And why did you do this? Why did you make this 2040 Plan A? In the first week after we published AI 2027, it blew up a lot bigger than we expected, by the way. After we published AI 2027, it blew up a lot bigger than we expected, by the way. We actually made forecasts beforehand of like,

268
01:43:05.917 --> 01:43:28.732
how many views it would get and stuff like that. And it was like 90th percentile outcome. So like, um, very much not what we expected. Um, but in like the Twitter storm that happened, various people were like, are, why are you giving us all this like doom and gloom, uh, predictions? Like, how about a more positive vision of like what you think we should do instead? And I think that that seed sort of like, uh,

269
01:43:28.732 --> 01:43:44.594
implanted in us and then we were like yeah that's reasonable like we've sort of depicted what we think the default path looks like and why we think it's pretty scary now maybe we should switch tax and come up with some actual recommendations and then depict that as well even though you don't believe they're probable

270
01:43:44.965 --> 01:44:14.783
Yeah, I mean, you can vote for a political candidate even if you aren't confident that they're going to win, you know? And you can say, like, here's what I think we should do, even if you think that people are probably not going to do it. You shouldn't say this if you think it's completely unlikely. Like, if you think there's no chance, then, like, maybe you shouldn't bother. But we think there's a chance. Like, in particular, for the reasons that we describe in this scenario, we think that people are going to wake up to the power of AI over the next few years. Because of something happens? The companies are saying that they're going to do this.

271
01:44:14.783 --> 01:44:34.780
And they are kind of on track. And it just sort of makes sense that, like, if they get anywhere close to this level of AI, then there's, like, big issues and big problems. And, like, we need to, like, do something about this. And so I think that even if there's not any, like, very dramatic warning shot or something...

272
01:44:34.780 --> 01:44:47.858
I think that just naturally people are going to start paying more attention to this and reasoning through the implications and trying to predict what's going to happen. And so naturally people are going to be more interested in regulation of AI, for example. And in fact...

273
01:44:49.411 --> 01:45:13.627
There's actually more of this happening than we predicted. More of what happening? Serious interest in AI regulation. So at the time that we published AI 2027, the sort of like mainstream position of the tech companies and in the government was kind of like AI regulation, bad idea. Free for all. Free for all. Yeah. In fact, there was even an attempt to preemptively ban states from regulating AI. Yeah. Yeah.

274
01:45:14.268 --> 01:45:26.722
You remember that? Now it seems like the conversation has changed a lot. Like now the U.S. government just told Anthropic they have to shut down their AI because they were worried that bad actors would use it for cyber attacks. You know, the government...

275
01:45:27.481 --> 01:45:56.995
is like waking up and doing more stuff than we expected already. And we're actually hopeful that that trend will just continue and that before it's actually too late, there will be very serious conversations happening inside the government and outside the government and in the broader society about all of these issues and trying to chart a course that avoids the loss of control and concentration of power risk that we mentioned. You've spent, what, must be almost coming up to 15 years thinking about this stuff.

276
01:45:58.025 --> 01:46:19.507
If this here was a button, and if you press that button, your plan S would occur, and it would shut down every data center that is currently training a frontier AI model for good. There would never be any other AI labs working on these problems. Would you press that button?

277
01:46:20.165 --> 01:46:46.169
I was about to slam it until you said for good. Okay. Like, I think if it was a sort of temporary shutdown, I would totally slam that button because we are not ready to do this, you know? Like, civilization is not ready to have these companies automate themselves and then get smarter and smarter and then have the superintelligence. Like, no, there's a bunch of reasons why that's really dangerous. But I would be at least hesitant to press this button

278
01:46:46.372 --> 01:47:05.525
if it permanently foreclosed the possibility of ever doing it again, for sure. But if you think that plan D is probable, which is this race we're on to superintelligence... If I had a choice between D and S, I think I would press it. Well, it comes down to what you think, right? Because if you think that is what's going to happen, plan D, and the only alternative...

279
01:47:05.879 --> 01:47:35.258
I didn't say this is what's going to happen. Probabilistically. Yeah, yeah. Like I'd be like, this is the most likely. Maybe this is the second most likely. Maybe this is the third most likely. They are all possible. So with your current perspective on whatever one you think is going to happen, would you press the button? I'm giving you an S, a definite S or whatever you think is going to happen. That's tough. What is the scope of the shutdown? So is it... It's no one can train an AI model again. Ever again. Ugh.

280
01:47:37.081 --> 01:48:02.545
That's real rough because, like I said, there's loads of benefits that we could get from AI if we do it right. I think I've almost put you in the position of Sam Altman to some degree. Yeah. Do you mind if I just take a moment to think about this? I prefer you to think. Yeah. I think I would not press...

281
01:48:03.389 --> 01:48:32.583
the button but I'm I feel very torn about it um the reason why I think I would not press the button is that I still have substantial hope that we can get something much better than this something more like this and I think that basically I think that if we don't build powerful AI systems eventually then we're probably going to die as a civilization eventually you know like 100 years from now 200 years from now something like that like nuclear war pandemic something you know

282
01:48:34.102 --> 01:49:02.941
I don't think human civilization right now is like super, super stable. And so I think that basically what I was about to say was the possible benefits for posterity and for all the billions and billions of people who could live in the future outweigh the like the current level of risk. But actually... I've heard that narrative before. Yeah, I don't know. Like, yeah, like maybe, maybe it's just like, nope.

283
01:49:03.076 --> 01:49:30.835
The people right now are the people we should prioritize. People right now are in grave danger. They're going to be fine for at least the next couple of decades. So never mind posterity. Prioritize the people right now. People right now definitely don't want to do this lottery, I would say. Yeah, you've really asked me a tough question. So would you press the button if that was the button?

284
01:49:32.675 --> 01:49:52.098
Probably not, but I would feel very torn. Okay. So what I always think about the personas of the audience that are watching, and these are, you know, they're very curious people, especially on the subject of AI, as we've seen, but they want to know what it means for them. I think a lot of them also want to know what they can do. Ah, yes. Yeah, what can people do? Well, yeah.

285
01:49:52.402 --> 01:50:20.161
I think that if you either have talent or passion, you can get directly involved. There's lots of organizations that are worried about these things and that are trying to do something about it, like political advocacy or technical research or building useful tools that will hopefully help people be better and stuff. But if you don't want to make any major career changes or things like that, then I would say just pay more attention to these issues and talk about it more.

286
01:50:20.161 --> 01:50:40.546
with people. Do stuff like, you know, emailing your congressman or whatever. It doesn't change things that much, but it does help. I think that, especially for this particular issue, the core problem is that people aren't taking it seriously yet. Like, if the sorts of things that I was just saying to you for the last hour or two were just, like, top of everybody's mind—

287
01:50:42.183 --> 01:51:07.006
we wouldn't even be here. Like, there would already be much more significant regulation in place, you know? And not only would there be more heavy regulation in place, but there would have been better regulation in place that's less, you know, less like a cudgel and more like a scalpel, and that's, like, more sensitive to what's actually bad and what's not so bad and so forth. And there'd be more expert people in the government and advising the

288
01:51:08.086 --> 01:51:28.100
The more people wake up to these concerns and to these projections, I think the more likely it is that we can do good stuff before it's too late. What about how they should vote at the polls? We've got an election coming up in the United States in a couple of years' time, but there's elections happening all over the world, all the time. You should ask your candidates what they think about all this AI stuff. You should...

289
01:51:28.100 --> 01:51:47.692
try to get them to like have opinions and then you should vote for the candidates whose opinions are better on this topic. This is the most important thing happening in our lifetimes, probably in all of history, in fact. And it's very important that it go well. And so it's what all the leaders of all the countries should be thinking about and making plans for. Isn't it such a weird thing to be alive at this moment in time?

290
01:51:48.097 --> 01:52:15.974
Like I was thinking about all the times that I could have been born. And I guess my ancestors probably thought the same. But I was thinking as you were speaking, I was like, I think it's when you referred to it as like the final show. Yeah. What was the phraseology you used? I said the run up to the climax or something. Yeah. I mean, what a crazy thing to be born in the run up to the climax where everything you're describing here is within my lifetime conceivably, hopefully. Yeah. Or maybe not, hopefully. What a crazy time to be alive. Certainly. Yeah.

291
01:52:15.974 --> 01:52:34.047
I noticed that when I asked you if you had kids, your demeanor changed quite considerably. Well, it's, yeah. It's like you dropped into a different state. Obviously, that's been central to the rumination that you've been experiencing. Well, it is a sad topic, right? Like, when I had kids...

292
01:52:34.351 --> 01:52:59.900
Like the reason to have kids is in large part about the future, you know? Like it's not just like a cuddly thing to have with you in the moment. It's because you have all these hopes and dreams about how they'll grow up and how they'll go do their own thing and be their own person and stuff. And because of what's happening with AI, I think a lot of those dreams are in jeopardy. Presumably you still would have had kids. I've actually flip-flopped on this occasionally. Yeah. Basically the top line answer is I'm not sure. The...

293
01:52:59.900 --> 01:53:23.947
My first child, we had her when she was born in 2019. Yeah. So this was before my timeline shortened a lot. So at this point, I was interested in AI. I was tracking the field. I was making forecasts. But I didn't actually expect it to happen soon. You know? And then this caused, like, when I did start thinking, like, oh, my gosh, it's going to be happening real soon. Like, by 2030, you know?

294
01:53:24.554 --> 01:53:54.068
Um, that caused some reconsidering. And so I basically told my wife, like, let's not have any more kids. It's too uncertain, you know, but that turned out to be really hard because, especially for my wife, like we already had one kid and like no siblings. Um, so eventually I sort of gave in and was like, okay, well, you know what? We already have one. It's going to be all right. Like maybe, maybe the future will be good. And even if it's not like, well, we're all on the same boat together.

295
01:53:55.216 --> 01:54:19.246
It's quite chilling what you're saying. It's chilling because you know more than me. And if you're at home saying to your wife, listen, maybe we should pause on having more children and building a family because of what's going on with AI. To be clear, yes, I mean, yes, it's very concerning. I am chilled. This is bad. This is what I've been saying. I hope things go well. I think things might go well. I think that there's a lot we can do to steer things in a better direction.

296
01:54:19.381 --> 01:54:45.757
I mean, one of those things as well, I have to say is just speaking about it. It's, I think a lot of the progress we've seen with governments waking up and, you know, we've seen certain things with people booing certain people at certain events, um, is, is a downstream from people like yourself actually coming on shows like this and all the other podcasts and telling us what's going on because elsewhere, to be fair, we're going to be gaslighted by the people that have the biggest PR machines. Yeah.

297
01:54:45.757 --> 01:55:07.762
So I often, I think it's probably worth me saying, I find myself kind of in two minds because I'm an entrepreneur and I'm an investor. I'm an investor in probably more than 100 companies now. And so many of those companies are using AI. I invested in Grok, the inference chip company. I've invested in SpaceX, which now own another Grok and they're doing AI. I use AI every day in my life. I've been using it through this conversation to understand different things that you've said.

298
01:55:07.762 --> 01:55:32.078
So that's one side of me, which is like business builder, entrepreneur, who has seen the benefits of AI in my own life. And then there's the other side of me. And it's funny, because I think sometimes people think you have to pick a camp. But through all of my life, even when I was a social media CEO, and I was saying, by the way, listen, I'm building a social media business, but I think there's some downsides to social media. I find myself at the same moment where I'm like, I build with AI, I have AI investments. And at the same time, as a civilian, I'm like...

299
01:55:32.534 --> 01:55:55.552
Yeah. I mean, I think that is a tension. I think that there's different ways you can draw the line. And I know lots of people who draw the line in lots of different ways. So there's some people who just like, I'm not going to use AI. I think this stuff is bad and on a bad trajectory. So I'm going to like boycott AI, right? I'm not one of those people. I use AI a lot. We all do at AI Futures Project. It's helpful for a lot of our work.

300
01:55:55.552 --> 01:56:23.159
The opposite end of the spectrum is people being like, well, it seems like it's on a trajectory to happen. So the thing to do to make it go well is to like get involved and accumulate power and try to like steer it from the inside. And so I'm going to go work at OpenAI or Anthropic and like try to like climb the ranks and then like, you know, be someone who matters when the important decisions are being made. And I know loads of people like that. That was like what I was doing when I was, that wasn't what I was doing exactly, but like.

301
01:56:23.311 --> 01:56:47.257
That was the path. That was like, that was, I mean, in some sense, this is what the whole narrative of the companies are, right? Like, this is why they tell themselves it's okay to do what they're doing, is that they're worried about the other guys, you know? And so, like, all these people are deciding, like, we're going to, like, lean really hard into it. We're going to, like, be there in the room when the important decisions are being made, you know? So there's a whole spectrum. And I'm sort of, like, somewhere in the middle. Like, I'm not at the AI companies. I'm not helping them go faster.

302
01:56:47.257 --> 01:57:13.717
instead i'm talking to the broad public and trying to advocate for what i think is the my current best guess as to the way out you know the way forward um but i'm not like boycotting all the aIs i'm not like you know uh trying to i'm not refusing to like engage with it in that way do you think it's too late no i don't think it's too late if i thought it was too late i wouldn't be here where would you be with my family

303
01:57:16.366 --> 01:57:27.875
What's your closing message to the general public, if you had to have a closing statement to them? Maybe I would say that, like, you're going to hear a lot of things, and you already have been hearing a lot of things about...

304
01:57:28.077 --> 01:57:52.765
AI, and it's going to sound like science fiction. But sometimes things which sound like science fiction happen in reality. And in fact, many times, historically, things which used to be science fiction have then become reality. And people need to stop thinking about what does or doesn't sound like science fiction and just start thinking about the trends and the actual trends that this technology is on. And

305
01:57:52.765 --> 01:58:16.357
reading and forecasting how it's going to go, and then taking seriously the possibility that it could go something like this, and then thinking about what should be done about that. And where would you direct them to get more information? You can go to ai2047.com to read our previous scenario. You can go to ai2040.com plan A to read our new proposal for what it should be done. These things are not just...

306
01:58:16.357 --> 01:58:37.012
a sci-fi story they also have lots of like explainers and links to other things and so they're kind of like a nice jumping off point to to learn about all of this stuff um if you want i could um after this is over like give a reading list of like other papers and articles and please do blogs to follow and so forth

307
01:58:37.012 --> 01:58:52.064
And I'll link them all below in the comment section. So if you're listening now, go ahead and take a look at the description of this episode and you'll see a bunch of links, which is Daniel's recommendations of what you should read. You know, I think it's just a really, really great moment in time to get educated on this stuff.

308
01:58:52.064 --> 01:59:20.060
um humans have an inclination because of cognitive dissonance where we feel uncomfortable about something to bury our heads in the sand and avoid it yeah but actually i think this is one such time to do the very opposite for many reasons to to inform yourself so you know what actions to take but also because ai you know unavoidably is going to be a huge part of all of our lives and careers yeah yeah thank you and that that's a good way to say it it's going to matter a lot it's going to it's going to be everywhere soon and um

309
01:59:20.060 --> 01:59:22.675
you need to do something about it before it's too late

310
01:59:22.777 --> 01:59:52.763
And what about AI Future Project? That's our organization. We spent a year writing AI 2047 after I left OpenAI, and then we spent another year writing AI 2040 Plan A. Daniel, thank you. Thank you. Thank you for all the work that you do. I can see how much you care about this stuff. And it's your care. It's funny, care itself makes others feel care. And seeing how personal this is for you and seeing how much you've dedicated your life to this, but also hearing that you basically walked away from $2 million to be able to speak to the public about this information is incredibly admirable.

311
01:59:52.763 --> 02:00:12.575
I think voices like yours are more important now than they've ever been on this subject. So please do keep fighting the fight that you're fighting. And that's one of information. It is of honesty. And it is of saying what is often the quiet part out loud. Thank you. Doing really, really smart research. I'll link everything we've discussed today below. And I hope we can chat again sometime soon. Thank you.