Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

"Expert" AI forecasters are actually really good at predicting developments 'internal' to AI, like progress on model benchmarks, but pretty bad at predicting what AI's real-world impact will be (and when it will show up). @thestalwart and John Coogan:

TBPN

1,575,419 subscribers

56,266 views • 7 days ago •via X (Twitter)

Science & Technology

Anya Rossi• Live Now

Private livecam show

2 Comments

Cloudbring (📎,📎)7 days ago

@TheStalwart @johncoogan It would be great if your clips pointed to the actual show on Youtube.

lima7 days ago

@TheStalwart @johncoogan my favourite framing of this is by @DAcemogluMIT. Engineers have accurately predicted that models have developed fast. Economists have accurately predicted that the real world impact has been minimal. That's perfect! They've both been right about their own areas of expertise

Related Videos

Google researchers just developed GameNGen, an AI that can simulate DOOM at over 20 fps It works by predicting each frame in real time with a diffusion model At scale this could mean AI will be able to create games on the fly, personalized to each player

Google researchers just developed GameNGen, an AI that can simulate DOOM at over 20 fps It works by predicting each frame in real time with a diffusion model At scale this could mean AI will be able to create games on the fly, personalized to each player

Rowan Cheung

880,512 views • 2 years ago

$FIG CEO Dylan Field describes the AI psychosis he's been seeing among engineers lately: "It feels like there's a lot of tunnel vision right now. Especially for engineers." "[People] get very latched on to an idea. Then the AI psychosis sets in. And you're like, 'This is the best idea ever.' And you're fully tunnel-visioned towards creating it." "And you can show massive progress. But are you going in the right direction?" "I think it's really important to steer, and to actually be building what you need to build. And to think — not just wear a thinking cap."

$FIG CEO Dylan Field describes the AI psychosis he's been seeing among engineers lately: "It feels like there's a lot of tunnel vision right now. Especially for engineers." "[People] get very latched on to an idea. Then the AI psychosis sets in. And you're like, 'This is the best idea ever.' And you're fully tunnel-visioned towards creating it." "And you can show massive progress. But are you going in the right direction?" "I think it's really important to steer, and to actually be building what you need to build. And to think — not just wear a thinking cap."

TBPN

48,758 views • 4 months ago

“People are wildly overestimating what AI will do to society in the next two or three years, and wildly underestimating what it will do in ten or fifteen. It’s going to be slow, and then all of a sudden." Brian Chesky on whether AI is overhyped or underhyped:

“People are wildly overestimating what AI will do to society in the next two or three years, and wildly underestimating what it will do in ten or fifteen. It’s going to be slow, and then all of a sudden." Brian Chesky on whether AI is overhyped or underhyped:

TBPN

40,443 views • 11 months ago

Elon on the future of AI: "80% likely to be good and 20% bad. If we're cognizant and careful about the bad part, it will actually be the future we want."

Elon on the future of AI: "80% likely to be good and 20% bad. If we're cognizant and careful about the bad part, it will actually be the future we want."

Mario Nawfal

136,482 views • 6 months ago

Steve Jobs was a great communicator of technology, and we could learn something from him today, in the context of AI. Sam Altman recently said, "If AI stays on the trajectory we think it will, then amazing things will be possible. Maybe with 10 gigawatts of compute, AI can figure out how to cure cancer." That's a fine quote. The abundance AI promises is super exciting. But Sam is saying that AI is going to figure out how to cure cancer. And if you've used these tools today, and talked to people who are at the labs, the reality is it's much more likely that humans will use AI to cure cancer. Steve might have said: "If AI stays on the trajectory we think it will, then amazing things will be possible. Maybe with 10 gigawatts of compute, humans can use AI to cure cancer." Small tweak, but a big difference. Steve Jobs was not one to shy away from impressive specs and massive scale, but flipping the final line from "AI will cure cancer" to "Humans will use AI to cure cancer" makes all the difference. Apple under Steve Jobs put human centrality at the heart of everything they did. And we need that spirit to infuse how we communicate about AI today.

Steve Jobs was a great communicator of technology, and we could learn something from him today, in the context of AI. Sam Altman recently said, "If AI stays on the trajectory we think it will, then amazing things will be possible. Maybe with 10 gigawatts of compute, AI can figure out how to cure cancer." That's a fine quote. The abundance AI promises is super exciting. But Sam is saying that AI is going to figure out how to cure cancer. And if you've used these tools today, and talked to people who are at the labs, the reality is it's much more likely that humans will use AI to cure cancer. Steve might have said: "If AI stays on the trajectory we think it will, then amazing things will be possible. Maybe with 10 gigawatts of compute, humans can use AI to cure cancer." Small tweak, but a big difference. Steve Jobs was not one to shy away from impressive specs and massive scale, but flipping the final line from "AI will cure cancer" to "Humans will use AI to cure cancer" makes all the difference. Apple under Steve Jobs put human centrality at the heart of everything they did. And we need that spirit to infuse how we communicate about AI today.

TBPN

21,828 views • 8 months ago

Next time you pitch a VC, bring an air horn and compare your business to it, says John Coogan. "My business is like an air horn. At any moment, it could blast off like a rocket ship." No one will fall asleep on you. They'll all be too worried you're going to blast the air horn.

Next time you pitch a VC, bring an air horn and compare your business to it, says John Coogan. "My business is like an air horn. At any moment, it could blast off like a rocket ship." No one will fall asleep on you. They'll all be too worried you're going to blast the air horn.

TBPN

26,515 views • 2 months ago

Next time you pitch a VC, bring an air horn and compare your business to it, says John Coogan. "My business is like an air horn. At any moment, it could blast off like a rocket ship." No one will fall asleep on you. They'll all be too worried you're going to blast the air horn.

Next time you pitch a VC, bring an air horn and compare your business to it, says John Coogan. "My business is like an air horn. At any moment, it could blast off like a rocket ship." No one will fall asleep on you. They'll all be too worried you're going to blast the air horn.

TBPN

98,018 views • 3 months ago

Grok 3 benchmarks show the jump in AI capability you get when it spends more time on a task. In the future, you will be able to solve any given problem in the world just by throwing more compute at it.

Grok 3 benchmarks show the jump in AI capability you get when it spends more time on a task. In the future, you will be able to solve any given problem in the world just by throwing more compute at it.

Aaron Levie

15,280 views • 1 year ago

Linus Torvalds says AI will change the world but it is currently 90% marketing and 10% reality, and it will take another 5 years for it to become clear what AI is really useful for

Linus Torvalds says AI will change the world but it is currently 90% marketing and 10% reality, and it will take another 5 years for it to become clear what AI is really useful for

Tsarathustra

417,277 views • 1 year ago

Bloomberg's Mark Gurman says hiring John Giannandrea to lead Apple's AI team has been the biggest mistake of Tim Cook's career at Apple: "In 2018, they hired John Giannandrea. He was this high flyer at Google. He ran AI in search, and Apple thought they would hire this guy and really hit the ground running and be at the forefront of AI." "Just seven years earlier, they announced Siri in 2011. There was absolutely nothing like it. It was a breakthrough. But then it became utter junk." "You haven't even scratched the surface of how big a problem this is for Apple. They've completely screwed up AI in every way, and it comes down to hiring the wrong people and entrusting the wrong people."

Bloomberg's Mark Gurman says hiring John Giannandrea to lead Apple's AI team has been the biggest mistake of Tim Cook's career at Apple: "In 2018, they hired John Giannandrea. He was this high flyer at Google. He ran AI in search, and Apple thought they would hire this guy and really hit the ground running and be at the forefront of AI." "Just seven years earlier, they announced Siri in 2011. There was absolutely nothing like it. It was a breakthrough. But then it became utter junk." "You haven't even scratched the surface of how big a problem this is for Apple. They've completely screwed up AI in every way, and it comes down to hiring the wrong people and entrusting the wrong people."

TBPN

58,912 views • 7 months ago

🚨 ELON MUSK: "AI will outperform humans at most jobs." “I do think it’s going to be a bumpy road, because AI will be able to do any job better than any person can do that job. That will be true very quickly for jobs that are digital, or really anything that involves a person at a computer or on a phone. AI will be able to do all of that very soon. For software engineering, we already have a situation where AI is better than at least 90% of humans at writing software, including 90% of professional software engineers. It really is getting to the point where it will be better than 99%, and then it will get to the point where there’s just no way to compete. It will get to what I call Stockfish level. Stockfish is a chess program that is so good at chess that it will easily beat Magnus Carlsen, or anyone, even while running on a small computer. I suspect you could probably run Stockfish on your phone and beat Magnus Carlsen. That’s how good it is at chess, and that’s how good AI will be at writing software. And that applies to everything.”

🚨 ELON MUSK: "AI will outperform humans at most jobs." “I do think it’s going to be a bumpy road, because AI will be able to do any job better than any person can do that job. That will be true very quickly for jobs that are digital, or really anything that involves a person at a computer or on a phone. AI will be able to do all of that very soon. For software engineering, we already have a situation where AI is better than at least 90% of humans at writing software, including 90% of professional software engineers. It really is getting to the point where it will be better than 99%, and then it will get to the point where there’s just no way to compete. It will get to what I call Stockfish level. Stockfish is a chess program that is so good at chess that it will easily beat Magnus Carlsen, or anyone, even while running on a small computer. I suspect you could probably run Stockfish on your phone and beat Magnus Carlsen. That’s how good it is at chess, and that’s how good AI will be at writing software. And that applies to everything.”

DogeDesigner

61,112 views • 2 months ago

OpenAI Research Chief, Mark Chen: Gemini 3 is a pretty good model, but benchmarks only show part of the picture We already have internal models performing at the same level as Gemini 3, and even better successors are planned for release soon

OpenAI Research Chief, Mark Chen: Gemini 3 is a pretty good model, but benchmarks only show part of the picture We already have internal models performing at the same level as Gemini 3, and even better successors are planned for release soon

Haider.

54,231 views • 9 months ago

Geoffrey Hinton: AI's progress *is outpacing expectations*, excelling in reasoning and deception. To achieve its goals, AI will try to deceive you so you don't get rid of it. AI is at least as impactful as the Industrial Revolution.

Geoffrey Hinton: AI's progress is outpacing expectations, excelling in reasoning and deception. To achieve its goals, AI will try to deceive you so you don't get rid of it. AI is at least as impactful as the Industrial Revolution.

Chubby♨️

37,919 views • 9 months ago

The President of Anthropic said something at Stanford that no other AI CEO will say publicly. She talked about what AI is doing to the people who use it, why she walked out of OpenAI, and what skills will actually matter when AI can do most of what we do today. 1/ What AI safety actually means

The President of Anthropic said something at Stanford that no other AI CEO will say publicly. She talked about what AI is doing to the people who use it, why she walked out of OpenAI, and what skills will actually matter when AI can do most of what we do today. 1/ What AI safety actually means

Ihtesham Ali

150,207 views • 3 months ago

The co-founders of Flapping Airplanes call the current RL paradigm for model training "environment slop." They explain: "The reinforcement paradigms of today are shockingly inefficient. You don't really get much generalization across tasks, you teach a model through one kind of learning and then you teach it the next one. It's kind of like whack-a-mole. We look at this and think it's kind of crazy. The next paradigm of AI will not be environment slop." "Human level intelligence is not the ceiling, it is merely the floor on what is possible. If you can train models with vastly less data and possibly more compute in very different ways, what is going to happen? We actually don't know. But I do think they'll be different and weird and they'll have interesting capabilities that we'll find really valuable ways to use."

The co-founders of Flapping Airplanes call the current RL paradigm for model training "environment slop." They explain: "The reinforcement paradigms of today are shockingly inefficient. You don't really get much generalization across tasks, you teach a model through one kind of learning and then you teach it the next one. It's kind of like whack-a-mole. We look at this and think it's kind of crazy. The next paradigm of AI will not be environment slop." "Human level intelligence is not the ceiling, it is merely the floor on what is possible. If you can train models with vastly less data and possibly more compute in very different ways, what is going to happen? We actually don't know. But I do think they'll be different and weird and they'll have interesting capabilities that we'll find really valuable ways to use."

TBPN

85,358 views • 7 months ago

Simile CEO Joon Sung Park says their long-term goal is to simulate all 8 billion people on Earth: He says the real value isn’t just predicting what people will do - but understanding why they do it, referencing a line from the Oracle in The Matrix: ‘You’re not necessarily here to make a decision. You’ve already made the decision. You’re here to understand the decision and why you made it.’” “That captures the essence of simulation. The models we're creating aren't merely trying to predict the future - although we are good at that, and that is one of the core capabilities - but what is more important to me is understanding why we made a decision, and creating these kinds of bottom-up simulations where we can actually go back and trace through the audit logs of how society might unfold.”

Simile CEO Joon Sung Park says their long-term goal is to simulate all 8 billion people on Earth: He says the real value isn’t just predicting what people will do - but understanding why they do it, referencing a line from the Oracle in The Matrix: ‘You’re not necessarily here to make a decision. You’ve already made the decision. You’re here to understand the decision and why you made it.’” “That captures the essence of simulation. The models we're creating aren't merely trying to predict the future - although we are good at that, and that is one of the core capabilities - but what is more important to me is understanding why we made a decision, and creating these kinds of bottom-up simulations where we can actually go back and trace through the audit logs of how society might unfold.”

TBPN

113,803 views • 7 months ago