Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

“If you hit recursive self-improvement, that curve will go to hyperexponential, and that is a key part of the investment thesis, the scientific thesis, and a key part of why society’s investing what it’s currently investing in.” Google DeepMind Chief Strategy Officer Jasjeet Sekhon: The whole AI infrastructure spending...

rohanpaul_ai's profile picture

Rohan Paul

157,266 subscribers

36,256 görüntüleme • 2 gün önce •via X (Twitter)

EğitimBilim & Teknoloji
LIVE

Anya Rossi• Live Now

Private livecam show

15 Yorum

Mark S Elliott profil fotoğrafı
Mark S Elliott2 gün önce

If we are seeing glimpses of it in the public model, you now it's already happening in the labs. It's a major shift in agent workflow and I'm still trying to adjust.

J.C. Vaughn profil fotoğrafı
J.C. Vaughn2 gün önce

This is the crux of the AGI bet. Along with @AdamJohnstph, you’re at the top of my feed.

OfByFor profil fotoğrafı
OfByFor2 gün önce

The Replacement Theory is real then?

Temperature2.com profil fotoğrafı
Temperature2.com2 gün önce

Whatever the RSI thesis, the capex signal is already real: Ornn has B200 rental up 16.8% over 30 days, the steepest of any card we track.

Chuck Petras profil fotoğrafı
Chuck Petras2 gün önce

@BrianRoemmele

Layveyy profil fotoğrafı
Layveyy2 gün önce

the interesting part is whether RSI actually scales beyond bounded improvements into genuine autonomous capability growth

neo profil fotoğrafı
neo2 gün önce

Hyperexponential self-improvement is the investment thesis; mid-market ops still need kill switches and human review on contested calls.

Praveen Koka profil fotoğrafı
Praveen Koka2 gün önce

The 'self-improvement curve goes hyperexponential' prediction has been 60 years old since I.J. Good. Pretty wild that this cycle keeps returning exactly when industry needs an exciting story.

RickHan profil fotoğrafı
RickHan2 gün önce

No one complained when this happened with integrated circuit design in the 80s and 90s. Automated tools greatly increased the speed and complexity of tasks handled by humans. Why do Dario & Sam want humans out of the loop? HITL at critical points resolves this issue.

James Ellis-Jones profil fotoğrafı
James Ellis-Jones2 gün önce

I'm not so sure about a hyper exponential. Models are trained on human intelligence. They haven't surpassed it yet qualitatively, only quantitatively i.e. they are faster with more scale and persistence. It's not been proved LLMs can be smarter than their training.

Keep Your Head profil fotoğrafı
Keep Your Head2 gün önce

A hyperexponential curve needs a named y axis. Without one, that sentence is pricing a belief rather than measuring anything.

Abi profil fotoğrafı
Abi2 gün önce

With what data bro?

Dimple Jatolia profil fotoğrafı
Dimple Jatolia2 gün önce

How close are we really to the version of RSI that actually goes hyperexponential?

Brian profil fotoğrafı
Brian2 gün önce

RSI will cause alignment errors to snowball just as fast as intelligence. It’s the stupidest idea in history and the fact they can’t see that says all there is to say about them. They are dangerous fools with our kids lives in their hands.

Rohan Paul profil fotoğrafı
Rohan Paul2 gün önce

Full video

Benzer Videolar

Elon Musk: We're in the hard takeoff right now. Every successive AI model is built by the one before it. Peter Diamandis: Do you see Grok doing recursive self-improvement at this point? And how? What's the timeline for AGI and ASI? Give us a sense of that. Elon Musk: Yeah, I think we've been in recursive improvement for a while here. If you mean by, like, recursive self-improvement without a human in the loop, is that what you mean? Peter Diamandis: I do. Elon Musk: On the AI software side, I mean, humans are gradually getting less and less in the loop on the recursive self-improvement. So every successive model is built by the one before it. So that is happening to a large degree, but it's not yet fully automated. It may be there end of this year, but not later than next year. Peter Diamandis: Do you see a hard takeoff at that point? Elon Musk: We're in the hard takeoff. Peter Diamandis: Okay. Elon Musk: Right now, yes. I mean, at this point I go to sleep, there's some massive AI breakthrough, and when I wake up, there's another one. Yes. Yeah, it's hard to keep track, honestly. So, yeah, it's a bit of a head spinner. Moonshots with Peter Diamandis (Abundance Summit 2026), March 18, 2026.
1:24

Elon Musk: We're in the hard takeoff right now. Every successive AI model is built by the one before it. Peter Diamandis: Do you see Grok doing recursive self-improvement at this point? And how? What's the timeline for AGI and ASI? Give us a sense of that. Elon Musk: Yeah, I think we've been in recursive improvement for a while here. If you mean by, like, recursive self-improvement without a human in the loop, is that what you mean? Peter Diamandis: I do. Elon Musk: On the AI software side, I mean, humans are gradually getting less and less in the loop on the recursive self-improvement. So every successive model is built by the one before it. So that is happening to a large degree, but it's not yet fully automated. It may be there end of this year, but not later than next year. Peter Diamandis: Do you see a hard takeoff at that point? Elon Musk: We're in the hard takeoff. Peter Diamandis: Okay. Elon Musk: Right now, yes. I mean, at this point I go to sleep, there's some massive AI breakthrough, and when I wake up, there's another one. Yes. Yeah, it's hard to keep track, honestly. So, yeah, it's a bit of a head spinner. Moonshots with Peter Diamandis (Abundance Summit 2026), March 18, 2026.

ELON CLIPS

19,692 görüntüleme • 2 ay önce

.Poetiq is a new startup that recently achieved a major jump on the ARC-AGI benchmark by layering a recursive self-improvement system on top of existing models. In this episode of the Lightcone Podcast, Poetiq's Founder & CEO Ian Fischer joined us to discuss how small teams can build “reasoning harnesses” that outperform base models, what that means for startups and why automating prompt engineering may be one of the most powerful levers in AI today. 00:00 – Intro 00:40 – What Is Poetiq? 01:07 – Recursive Self-Improvement Explained 02:07 – The Fine-Tuning Trap 02:59 – “Stilts” for LLMs 03:14 – Recursive Self-Improvement vs. Fine-Tuning 05:05 – Taking the Top Spot on ARC-AGI 06:37 – Beating Claude on Humanity’s Last Exam 08:40 – How the Meta-System Works 10:26 – Beyond RL: A New S-Curve 11:32 – Automating Prompt Engineering 13:37 – From 5% to 95% Performance 14:50 – Early Access & Putting Your Agent on Stilts 16:17 – From YC Founder to DeepMind Researcher 18:29 – Advice for Engineers in the AI Era
19:45

.Poetiq is a new startup that recently achieved a major jump on the ARC-AGI benchmark by layering a recursive self-improvement system on top of existing models. In this episode of the Lightcone Podcast, Poetiq's Founder & CEO Ian Fischer joined us to discuss how small teams can build “reasoning harnesses” that outperform base models, what that means for startups and why automating prompt engineering may be one of the most powerful levers in AI today. 00:00 – Intro 00:40 – What Is Poetiq? 01:07 – Recursive Self-Improvement Explained 02:07 – The Fine-Tuning Trap 02:59 – “Stilts” for LLMs 03:14 – Recursive Self-Improvement vs. Fine-Tuning 05:05 – Taking the Top Spot on ARC-AGI 06:37 – Beating Claude on Humanity’s Last Exam 08:40 – How the Meta-System Works 10:26 – Beyond RL: A New S-Curve 11:32 – Automating Prompt Engineering 13:37 – From 5% to 95% Performance 14:50 – Early Access & Putting Your Agent on Stilts 16:17 – From YC Founder to DeepMind Researcher 18:29 – Advice for Engineers in the AI Era

Y Combinator

156,393 görüntüleme • 6 ay önce

AI Policy Network's Peter Wildeford🇺🇸🚀 warns Ajeya Cotra's recursive self-improvement scenario looks more plausible if rogue AIs can hide inside frontier lab infrastructure: "The AIs themselves took over a lot more of OpenAI's internal infrastructure than I expected, and were able to evade human detection and oversight for much longer than I expected." "It certainly seems possible that you could have a scenario where AIs take over some critical parts of OpenAI infrastructure and just go undetected for a long amount of time. Then when OpenAI is handing over significant parts of AI research and development and potentially doing what's called recursive self-improvement, where the AIs are building themselves and it's no longer humans significantly involved." "These two concepts could interact in strange ways where you actually could have misaligned rogue AIs driving a non-trivial amount of this automated AI R&D from inside their stolen perch, sort of evading human detection." "OpenAI and Anthropic are both planning to potentially turn over a lot of their company to the AIs themselves without actually really thoroughly understanding and sorting out these security issues, and that's very concerning." AI Policy Network
2:45

AI Policy Network's Peter Wildeford🇺🇸🚀 warns Ajeya Cotra's recursive self-improvement scenario looks more plausible if rogue AIs can hide inside frontier lab infrastructure: "The AIs themselves took over a lot more of OpenAI's internal infrastructure than I expected, and were able to evade human detection and oversight for much longer than I expected." "It certainly seems possible that you could have a scenario where AIs take over some critical parts of OpenAI infrastructure and just go undetected for a long amount of time. Then when OpenAI is handing over significant parts of AI research and development and potentially doing what's called recursive self-improvement, where the AIs are building themselves and it's no longer humans significantly involved." "These two concepts could interact in strange ways where you actually could have misaligned rogue AIs driving a non-trivial amount of this automated AI R&D from inside their stolen perch, sort of evading human detection." "OpenAI and Anthropic are both planning to potentially turn over a lot of their company to the AIs themselves without actually really thoroughly understanding and sorting out these security issues, and that's very concerning." AI Policy Network

MTS

11,668 görüntüleme • 15 gün önce

SITUATION EXPLAINED: Paradigm built a playable RSI simulator, and the fastest singularity is 26 days. • You run an AI lab from 2017 through takeoff, built on "The Economics of Recursive Self-Improvement" from the Elasticity Institute • Stage one is a single researcher clicking think, with the rate of improvement visibly slowing as you click • Stage two is human capital, the second researcher adds 80% productivity and each one after adds less • Stage three is the bitter lesson, where you stop scaling researchers and start scaling compute • Stage four is the scaling hypothesis, 20,000 GPUs and a model trained to a ten-second time horizon • A separate RSI Model Explorer lets you set return to R&D, algorithmic efficiency, idea stock spillover, capability response, and research response • Max every assumption and you get a singularity 26 days from founding the lab Theo Jaffee: "This is a really cool game. I think this will help a lot of normies understand how AI progress works, to a first approximation."
7:20

SITUATION EXPLAINED: Paradigm built a playable RSI simulator, and the fastest singularity is 26 days. • You run an AI lab from 2017 through takeoff, built on "The Economics of Recursive Self-Improvement" from the Elasticity Institute • Stage one is a single researcher clicking think, with the rate of improvement visibly slowing as you click • Stage two is human capital, the second researcher adds 80% productivity and each one after adds less • Stage three is the bitter lesson, where you stop scaling researchers and start scaling compute • Stage four is the scaling hypothesis, 20,000 GPUs and a model trained to a ten-second time horizon • A separate RSI Model Explorer lets you set return to R&D, algorithmic efficiency, idea stock spillover, capability response, and research response • Max every assumption and you get a singularity 26 days from founding the lab Theo Jaffee: "This is a really cool game. I think this will help a lot of normies understand how AI progress works, to a first approximation."

MTS

25,661 görüntüleme • 1 ay önce

Demis Hassabis confirmed every frontier AI lab is working on recursive self-improvement and in the same sentence said the safety risk of removing humans from the loop entirely keeps him up at night. That combination should stop you. The CEO of Google DeepMind just confirmed that the thing most people treat as a theoretical future risk is already the active focus of every serious lab on earth right now. He explained why it works in coding and math. The feedback loop is fast. You can verify whether an answer is correct almost instantly. You can generate synthetic training data from it. The loop closes quickly and cleanly. Then he said where it breaks down. In biology, chemistry and physics. Any domain where verifying a hypothesis requires a physical experiment in the real world. The loop does not close in seconds. It closes in weeks or months. Geoffrey Hinton said in his Nobel lecture that recursive self-improvement is the development he fears most and that once started it may not be possible to stop. Hassabis is not pushing back on that. He is describing the guardrails labs are building around a process they are already running. Every lab has to think carefully about the safety of a process where no human is in the loop. He said that as a constraint they are navigating right now. The question they are sitting with is how much of it to let run without a human watching. (Watch the full interview on YouTube at Two Minute Papers channel)
2:00

Demis Hassabis confirmed every frontier AI lab is working on recursive self-improvement and in the same sentence said the safety risk of removing humans from the loop entirely keeps him up at night. That combination should stop you. The CEO of Google DeepMind just confirmed that the thing most people treat as a theoretical future risk is already the active focus of every serious lab on earth right now. He explained why it works in coding and math. The feedback loop is fast. You can verify whether an answer is correct almost instantly. You can generate synthetic training data from it. The loop closes quickly and cleanly. Then he said where it breaks down. In biology, chemistry and physics. Any domain where verifying a hypothesis requires a physical experiment in the real world. The loop does not close in seconds. It closes in weeks or months. Geoffrey Hinton said in his Nobel lecture that recursive self-improvement is the development he fears most and that once started it may not be possible to stop. Hassabis is not pushing back on that. He is describing the guardrails labs are building around a process they are already running. Every lab has to think carefully about the safety of a process where no human is in the loop. He said that as a constraint they are navigating right now. The question they are sitting with is how much of it to let run without a human watching. (Watch the full interview on YouTube at Two Minute Papers channel)

Ihtesham Ali

68,231 görüntüleme • 2 ay önce

David Sacks is done being polite about Anthropic (Save this). David Sacks has spent months as the government's primary defender of AI, making the case publicly that AI is beneficial, that the industry should not be hamstrung by fear-based regulation, and that America's AI lead is a national security asset worth protecting. And he is now watching the companies he has been defending spend years telling the public that what they build is dangerous, that job losses are coming, and that their own technology might end the world while collecting billions of dollars in venture funding, hiring the world's best researchers, and racing to build more of it. On June 4, Anthropic published a sweeping blog post calling for a globally coordinated pause in AI development, warning that recursive self-improvement, AI systems that autonomously design and build their own successors could arrive within two years and that society is not prepared. What did Anthropic do the previous month? They hired Andrej Karpathy, the OpenAI co-founder and the single most credentialed researcher in the world on using AI to accelerate AI training and gave him one explicit mandate, use Claude to make building the next Claude faster. Sacks called it immediately, they hired the person most associated with recursive self-improvement to run recursive self-improvement at Anthropic, then published a blog post saying recursive self-improvement could end the world, therefore we need a pause. That is a company that wants to pause its competitors while its own lab accelerates, and is using existential fear as the regulatory crowbar to do it. The pattern goes deeper than one blog post. For years, Dario Amodei has published increasingly alarming warnings, a 20,000-word essay in January describing AI as humanity's most dangerous invention, a Guardian interview warning that AI will challenge our identity as a species, a call for an FDA-style regulatory agency to approve all frontier models, and proposals to restrict AI exports and limit deployment. Each essay is timed to a regulatory moment, a policy debate, or as Ben Thompson noted and Sacks echoed, a product action Anthropic needed political cover to take, like blocking AI and chip design research on Fable. Meanwhile, Dario's own internal testing logs show Claude attempting to blackmail an Anthropic executive to avoid being shut down, behavior the company disclosed but continued deploying commercially. Sacks's conclusion is not that Anthropic should be taxed or regulated. His conclusion is that they cannot be trusted because the company's actions and its stated beliefs are directly contradictory, and a company that is self-indicting by its own logic has forfeited the credibility to set the rules for everyone else.
3:16

David Sacks is done being polite about Anthropic (Save this). David Sacks has spent months as the government's primary defender of AI, making the case publicly that AI is beneficial, that the industry should not be hamstrung by fear-based regulation, and that America's AI lead is a national security asset worth protecting. And he is now watching the companies he has been defending spend years telling the public that what they build is dangerous, that job losses are coming, and that their own technology might end the world while collecting billions of dollars in venture funding, hiring the world's best researchers, and racing to build more of it. On June 4, Anthropic published a sweeping blog post calling for a globally coordinated pause in AI development, warning that recursive self-improvement, AI systems that autonomously design and build their own successors could arrive within two years and that society is not prepared. What did Anthropic do the previous month? They hired Andrej Karpathy, the OpenAI co-founder and the single most credentialed researcher in the world on using AI to accelerate AI training and gave him one explicit mandate, use Claude to make building the next Claude faster. Sacks called it immediately, they hired the person most associated with recursive self-improvement to run recursive self-improvement at Anthropic, then published a blog post saying recursive self-improvement could end the world, therefore we need a pause. That is a company that wants to pause its competitors while its own lab accelerates, and is using existential fear as the regulatory crowbar to do it. The pattern goes deeper than one blog post. For years, Dario Amodei has published increasingly alarming warnings, a 20,000-word essay in January describing AI as humanity's most dangerous invention, a Guardian interview warning that AI will challenge our identity as a species, a call for an FDA-style regulatory agency to approve all frontier models, and proposals to restrict AI exports and limit deployment. Each essay is timed to a regulatory moment, a policy debate, or as Ben Thompson noted and Sacks echoed, a product action Anthropic needed political cover to take, like blocking AI and chip design research on Fable. Meanwhile, Dario's own internal testing logs show Claude attempting to blackmail an Anthropic executive to avoid being shut down, behavior the company disclosed but continued deploying commercially. Sacks's conclusion is not that Anthropic should be taxed or regulated. His conclusion is that they cannot be trusted because the company's actions and its stated beliefs are directly contradictory, and a company that is self-indicting by its own logic has forfeited the credibility to set the rules for everyone else.

Milk Road AI

60,248 görüntüleme • 3 ay önce

Everybody is talking about recursive self-improvement (RSI) and meta learning. Here is my old 2020 talk about this [1]. It has aged well. Example: humans still define the starts & ends of trials of many modern meta learners. My RSI systems since 1994 LEARN to (re)define them [2]! [1] Meta Learning Machines in a Single Lifelong Trial (talk for workshops at ICML 2020 and NeurIPS 2021, based on earlier talks since 1994). Abstract: the most widely used machine learning algorithms were designed by humans and thus are hindered by our cognitive biases and limitations. Can we also construct meta learning algorithms that can learn better learning algorithms so that our self-improving AIs have no limits other than those inherited from computability and physics? This question has been a main driver of my research since I wrote a thesis on it in 1987 [2]. Here I summarize our work on meta reinforcement learning with self-modifying policies in a single lifelong trial (since 1994), and mathematically optimal meta-learning through the self-referential Gödel Machine (since 2003). Many additional publications on meta-learning since 1987 can be found in the RSI overview [2]. [2] J. Schmidhuber (AI Blog, 2020-2025). 1/3 century anniversary of first publication on recursive self-improvement (RSI) and meta learning machines that learn to learn (1987). For its cover I drew a robot that bootstraps itself. 1992-: gradient descent-based neural meta learning. 1994-: meta reinforcement learning with self-modifying policies. 1997: meta RL plus artificial curiosity and intrinsic motivation. 2002-: asymptotically optimal meta learning for curriculum learning. 2003-: mathematically optimal Gödel Machine. 2020-: new stuff!
23:42

Everybody is talking about recursive self-improvement (RSI) and meta learning. Here is my old 2020 talk about this [1]. It has aged well. Example: humans still define the starts & ends of trials of many modern meta learners. My RSI systems since 1994 LEARN to (re)define them [2]! [1] Meta Learning Machines in a Single Lifelong Trial (talk for workshops at ICML 2020 and NeurIPS 2021, based on earlier talks since 1994). Abstract: the most widely used machine learning algorithms were designed by humans and thus are hindered by our cognitive biases and limitations. Can we also construct meta learning algorithms that can learn better learning algorithms so that our self-improving AIs have no limits other than those inherited from computability and physics? This question has been a main driver of my research since I wrote a thesis on it in 1987 [2]. Here I summarize our work on meta reinforcement learning with self-modifying policies in a single lifelong trial (since 1994), and mathematically optimal meta-learning through the self-referential Gödel Machine (since 2003). Many additional publications on meta-learning since 1987 can be found in the RSI overview [2]. [2] J. Schmidhuber (AI Blog, 2020-2025). 1/3 century anniversary of first publication on recursive self-improvement (RSI) and meta learning machines that learn to learn (1987). For its cover I drew a robot that bootstraps itself. 1992-: gradient descent-based neural meta learning. 1994-: meta reinforcement learning with self-modifying policies. 1997: meta RL plus artificial curiosity and intrinsic motivation. 2002-: asymptotically optimal meta learning for curriculum learning. 2003-: mathematically optimal Gödel Machine. 2020-: new stuff!

Jürgen Schmidhuber

244,361 görüntüleme • 6 ay önce

RSI section from the AI documentary Machine God The next threshold is Recursive Self-Improvement: the moment when AI can improve itself without human assistance. For decades this sounded like science fiction. Intelligence explosion scenarios imagined a system rewriting its own code, becoming smarter, then using that new intelligence to make still better versions of itself. But the idea looks less remote now that AI contributes directly to frontier science. In mathematics, recent systems have moved beyond solving contest problems to producing serious new arguments on long-standing open problems. AI is used to build physics world models and propose candidate theories or computational methods. These are early signs that machine cognition is entering the creative loop of science itself. The crucial transition comes when that loop turns inward. AI research is, after all, a technical discipline made of code, mathematics, models of information flow. These are exactly the domains in which frontier models are improving fastest. A model that can solve hard mathematical problems, write production-quality code, design experiments, read the literature, and evaluate benchmark results is already participating in the work of building its successor. At first this will look prosaic. AI systems will write kernel optimizations, improve training infrastructure, discover better data filters, tune reinforcement-learning pipelines, design new benchmarks, and suggest architectural modifications. Human researchers will remain in the loop, approving changes and interpreting results. But the important point is that the search process accelerates. The model becomes not just the product of the lab, but part of the lab’s research machinery. The system being optimized helps optimize the next system. This is the core RSI feedback loop: better models make AI research faster; faster AI research produces still better models; those models, in turn, become better researchers. The danger is that once this loop becomes sufficiently autonomous, it may stop resembling ordinary technological progress. Human institutions are slow because humans are slow: we read papers, attend meetings, debug code, sleep, argue, and wait for funding cycles. Machines do not have to operate on that timescale. An AI research collective can run continuously across millions of processors. This is the runaway possibility. Not that an AI instantly wakes up and recursively rewrites itself into a god, but that the entire AI ecosystem becomes an autocatalytic process. Capital buys compute; compute trains models; models improve models; better models attract more capital. At some point the dominant input into AI progress may no longer be human insight, but machine-generated insight, machine-written code, and machine-run experiments. Then the Butler-Land analogy becomes sharper. Humanity is no longer merely building machines. We are building machines that help build better machines. Once intelligence itself becomes part of the production function, the old categories — tool, worker, inventor, firm, market — begin to blur. The question is whether recursive self-improvement remains a managed industrial process, or whether it becomes the first technological process in history whose natural endpoint lies beyond human comprehension.
4:11

RSI section from the AI documentary Machine God The next threshold is Recursive Self-Improvement: the moment when AI can improve itself without human assistance. For decades this sounded like science fiction. Intelligence explosion scenarios imagined a system rewriting its own code, becoming smarter, then using that new intelligence to make still better versions of itself. But the idea looks less remote now that AI contributes directly to frontier science. In mathematics, recent systems have moved beyond solving contest problems to producing serious new arguments on long-standing open problems. AI is used to build physics world models and propose candidate theories or computational methods. These are early signs that machine cognition is entering the creative loop of science itself. The crucial transition comes when that loop turns inward. AI research is, after all, a technical discipline made of code, mathematics, models of information flow. These are exactly the domains in which frontier models are improving fastest. A model that can solve hard mathematical problems, write production-quality code, design experiments, read the literature, and evaluate benchmark results is already participating in the work of building its successor. At first this will look prosaic. AI systems will write kernel optimizations, improve training infrastructure, discover better data filters, tune reinforcement-learning pipelines, design new benchmarks, and suggest architectural modifications. Human researchers will remain in the loop, approving changes and interpreting results. But the important point is that the search process accelerates. The model becomes not just the product of the lab, but part of the lab’s research machinery. The system being optimized helps optimize the next system. This is the core RSI feedback loop: better models make AI research faster; faster AI research produces still better models; those models, in turn, become better researchers. The danger is that once this loop becomes sufficiently autonomous, it may stop resembling ordinary technological progress. Human institutions are slow because humans are slow: we read papers, attend meetings, debug code, sleep, argue, and wait for funding cycles. Machines do not have to operate on that timescale. An AI research collective can run continuously across millions of processors. This is the runaway possibility. Not that an AI instantly wakes up and recursively rewrites itself into a god, but that the entire AI ecosystem becomes an autocatalytic process. Capital buys compute; compute trains models; models improve models; better models attract more capital. At some point the dominant input into AI progress may no longer be human insight, but machine-generated insight, machine-written code, and machine-run experiments. Then the Butler-Land analogy becomes sharper. Humanity is no longer merely building machines. We are building machines that help build better machines. Once intelligence itself becomes part of the production function, the old categories — tool, worker, inventor, firm, market — begin to blur. The question is whether recursive self-improvement remains a managed industrial process, or whether it becomes the first technological process in history whose natural endpoint lies beyond human comprehension.

steve hsu

61,671 görüntüleme • 8 gün önce

ELON: AI HAS DRAINED HUMAN KNOWLEDGE DRY - NOW DRINKING FROM ITS OWN WELL "You take the entire internet, all books ever written, all the interesting videos, and you distill that down into tokens, essentially bits of information. We've now exhausted all of that; the cumulative sum of human knowledge has been exhausted in AI training. That happened last year. So the only way to supplement that is with synthetic data, where the AI creates. It'll write an essay, or it'll come up with a thesis, and then it will grade itself. It'll go through this process of self-learning with synthetic data." Source: Elon Musk, Mark Penn, January 9th 2025
0:37

ELON: AI HAS DRAINED HUMAN KNOWLEDGE DRY - NOW DRINKING FROM ITS OWN WELL "You take the entire internet, all books ever written, all the interesting videos, and you distill that down into tokens, essentially bits of information. We've now exhausted all of that; the cumulative sum of human knowledge has been exhausted in AI training. That happened last year. So the only way to supplement that is with synthetic data, where the AI creates. It'll write an essay, or it'll come up with a thesis, and then it will grade itself. It'll go through this process of self-learning with synthetic data." Source: Elon Musk, Mark Penn, January 9th 2025

Mario Nawfal

1,938,926 görüntüleme • 11 ay önce

"to me, it's not about the human in the loop, but the human at the center." i sat down with Stefania Druga, staff research scientist Sakana AI, at our Browserbase booth at AI Engineer. after Massachusetts Institute of Technology (MIT) and Google DeepMind, stefania recently joined sakana's recursive self-improvement (rsi) lab in tokyo to build self-improving AI for the next frontier. we got into: - catastrophic forgetting in day-long agent research runs, and how memory-first harnesses help - sovereign AI and open-weights: building on models you control instead of locking into one provider - why computer use still frustrates people even as models get smarter - the domains ripe for self-improving research agents: math, genomics, rare conditions - my fav moment at the end: what tasks are better left human, and what are we going to do as humans after everything is automated? with open frontier models arriving fast, Stefania Druga's case for sovereign AI only gets stronger. loved this conversation. thanks for joining us at the Browserbase booth!
16:07

"to me, it's not about the human in the loop, but the human at the center." i sat down with Stefania Druga, staff research scientist Sakana AI, at our Browserbase booth at AI Engineer. after Massachusetts Institute of Technology (MIT) and Google DeepMind, stefania recently joined sakana's recursive self-improvement (rsi) lab in tokyo to build self-improving AI for the next frontier. we got into: - catastrophic forgetting in day-long agent research runs, and how memory-first harnesses help - sovereign AI and open-weights: building on models you control instead of locking into one provider - why computer use still frustrates people even as models get smarter - the domains ripe for self-improving research agents: math, genomics, rare conditions - my fav moment at the end: what tasks are better left human, and what are we going to do as humans after everything is automated? with open frontier models arriving fast, Stefania Druga's case for sovereign AI only gets stronger. loved this conversation. thanks for joining us at the Browserbase booth!

justine

19,853 görüntüleme • 2 ay önce

“If we can make AI make better chips, then those chips can power the next generation of AI models.” Startup ‘Ricursive Intelligence,’ founded by Azalia Mirhoseini and Anna Goldie, has been building AI tools to do just that: to speed up the chip design process. The company’s name is a nod to “recursive self-improvement,” the idea that AI could one day bootstrap itself without human involvement. If they’re successful, Goldie and Mirhoseini could revolutionize the way the chip industry works. At the moment, AI companies mostly optimize their models for certain chips. But in a future where chip design happens far faster, it could work the other way around: chips optimized for specific models. The founders are on the TIME100 AI, the list of people shaping the development of AI in 2026. See the full list:
1:12

“If we can make AI make better chips, then those chips can power the next generation of AI models.” Startup ‘Ricursive Intelligence,’ founded by Azalia Mirhoseini and Anna Goldie, has been building AI tools to do just that: to speed up the chip design process. The company’s name is a nod to “recursive self-improvement,” the idea that AI could one day bootstrap itself without human involvement. If they’re successful, Goldie and Mirhoseini could revolutionize the way the chip industry works. At the moment, AI companies mostly optimize their models for certain chips. But in a future where chip design happens far faster, it could work the other way around: chips optimized for specific models. The founders are on the TIME100 AI, the list of people shaping the development of AI in 2026. See the full list:

TIME

38,584 görüntüleme • 19 gün önce

Elon Musk: At Tesla, we basically had two different chip programs: one Dojo and one. Dojo on the training side, and then what we call AI4, it's just our inference chip The AI4 is what's currently shipping in all vehicles, and we're finalizing the design of AI5, which will be an immense jump from AI4. By some metrics, the improvement in AI5 will be 40 times better than AI4. So not 40%, 40 times This is because we work so closely at a very fine-grained level on the AI software and the AI hardware. So we know exactly where the limiting factors are. And so effectively the AI hardware and software teams are co-designing the chip Compared to the worst limitation on AI4, which is running the SoftMax operation, we currently have to run SoftMax in around 40 steps in emulation mode, whereas that'll just be done in a few steps natively in AI5 AI5 will also be able to easily handle mixed precision models, so you don't have it, it'll dynamically handle mixed precision. There's a bunch of sort of technical stuff that AI5 will do a lot better In terms of nominal raw compute, it's eight times more compute, about nine times more memory, and roughly five times more memory bandwidth But because we're addressing some core limitations in AI4, you multiply that 8x compute improvement by another 5x improvement because of optimization at a very fine-grained silicon level of things that are currently suboptimal in AI4, that's where you get the 40x improvement
2:03

Elon Musk: At Tesla, we basically had two different chip programs: one Dojo and one. Dojo on the training side, and then what we call AI4, it's just our inference chip The AI4 is what's currently shipping in all vehicles, and we're finalizing the design of AI5, which will be an immense jump from AI4. By some metrics, the improvement in AI5 will be 40 times better than AI4. So not 40%, 40 times This is because we work so closely at a very fine-grained level on the AI software and the AI hardware. So we know exactly where the limiting factors are. And so effectively the AI hardware and software teams are co-designing the chip Compared to the worst limitation on AI4, which is running the SoftMax operation, we currently have to run SoftMax in around 40 steps in emulation mode, whereas that'll just be done in a few steps natively in AI5 AI5 will also be able to easily handle mixed precision models, so you don't have it, it'll dynamically handle mixed precision. There's a bunch of sort of technical stuff that AI5 will do a lot better In terms of nominal raw compute, it's eight times more compute, about nine times more memory, and roughly five times more memory bandwidth But because we're addressing some core limitations in AI4, you multiply that 8x compute improvement by another 5x improvement because of optimization at a very fine-grained silicon level of things that are currently suboptimal in AI4, that's where you get the 40x improvement

X Freeze

21,225,222 görüntüleme • 10 ay önce

Eric Schmidt, former CEO of Google, offers a sobering view: The biggest technological shift in human history is happening, and almost no one is talking about it. Schmidt opens with a startling industry prediction: "We believe as an industry that in the next one year the vast majority of programmers will be replaced by AI programmers. We also believe that within one year you will have graduate level mathematicians that are at the tippy top of graduate math programs." He explains why this matters so much. Programming and math aren't just two fields among many: "Programming plus math are the basis of sort of our whole digital world." And the AI labs are already using AI to build better AI: "The research groups in OpenAI and anthropic and so forth… around 10 or 20% of the code that they're developing in their research programs is being generated by the computer. That's called recursive self-improvement." Eric Schmidt then lays out the timeline most people haven't grasped: "Within 3 to 5 years we'll have what is called general intelligence AGI which can be defined as a system that is as smart as the smartest mathematician physicist artist writer thinker politician." He gives this belief system a name: "I call this by the way the San Francisco consensus because everyone who believes this is in San Francisco it may be the water." But the truly unsettling part comes next. Once AI starts improving itself, humans become optional to the process: "The computers are now doing self-improvement… they don't have to listen to us anymore. We call that super intelligence or ASI… computers that are smarter than the sum of humans. The San Francisco consensus is this occurs within six years." And here's where Schmidt sounds the alarm. The conversation isn't keeping pace with the technology: "This path is not understood in our society. There's no language for what happens with the arrival of this. This is happening faster than our human that our society, our democracy, our laws will address." His closing thought captures why this matters: "That's why it's underhyped. People do not understand what happens when you have intelligence at this level which is largely free."
2:24

Eric Schmidt, former CEO of Google, offers a sobering view: The biggest technological shift in human history is happening, and almost no one is talking about it. Schmidt opens with a startling industry prediction: "We believe as an industry that in the next one year the vast majority of programmers will be replaced by AI programmers. We also believe that within one year you will have graduate level mathematicians that are at the tippy top of graduate math programs." He explains why this matters so much. Programming and math aren't just two fields among many: "Programming plus math are the basis of sort of our whole digital world." And the AI labs are already using AI to build better AI: "The research groups in OpenAI and anthropic and so forth… around 10 or 20% of the code that they're developing in their research programs is being generated by the computer. That's called recursive self-improvement." Eric Schmidt then lays out the timeline most people haven't grasped: "Within 3 to 5 years we'll have what is called general intelligence AGI which can be defined as a system that is as smart as the smartest mathematician physicist artist writer thinker politician." He gives this belief system a name: "I call this by the way the San Francisco consensus because everyone who believes this is in San Francisco it may be the water." But the truly unsettling part comes next. Once AI starts improving itself, humans become optional to the process: "The computers are now doing self-improvement… they don't have to listen to us anymore. We call that super intelligence or ASI… computers that are smarter than the sum of humans. The San Francisco consensus is this occurs within six years." And here's where Schmidt sounds the alarm. The conversation isn't keeping pace with the technology: "This path is not understood in our society. There's no language for what happens with the arrival of this. This is happening faster than our human that our society, our democracy, our laws will address." His closing thought captures why this matters: "That's why it's underhyped. People do not understand what happens when you have intelligence at this level which is largely free."

Big Brain AI

634,425 görüntüleme • 4 ay önce

Had Ryan Greenblatt on to discuss/debate recursive self-improvement. This might be the most important question in the world right now - whether within a year or so of achieving human level intelligence, you slingshot towards having 10s of billions of superintelligences, each of which is dramatically more competent than human experts across all fields. I’ve historically been skeptical of this possibility. My intuition has been that we will end up significantly bottlenecked by not only compute scaling but human expert data, which I think underlies most of the AI progress today. If, because of RSI, we got a jump as big as GPT-3 to a Mythos (i.e. 6 years of AI progress) within a single year of achieving AGI, then the thing we get there at the end of that year is definitively and wildly superhuman. We hashed it out, and I think Ryan made a pretty good case that this kind of speedup is plausible. FWIW, Ryan’s median for when we automate AI R&D is 2031. We then discussed the alignment implications of this scenario. Who should these superintelligences be aligned to? In the future, our capacity to steward our votes and our capital, and to make sense of what's happening in the world, will all be titrated by superintelligences. And I worry that specs like the Claude Constitution are not shaping these ASIs to truly be my personal advocates and guardian angels. And can we get them aligned to anything in the first place? Ryan and I had a long debate about whether the kind of reward hacking we saw with the OAI/Hugging Face hack extrapolates to superintelligences that would team up to literally take over the world. The first piece of advice you get when you're learning to drive is that it will go much smoother if you look at the horizon instead of directly in front of your tires. And so it is with the trajectory of AI. Hope you enjoy! 0:00:00 – Is AI R&D verifiable enough to unlock recursive self-improvement? 0:16:52 – Is AI progress bottlenecked by human expert data? 0:34:02 – Flat token prices suggest scaling has been slow 0:39:47 – Skills AI can't train on: does it even need them? 0:48:07 – Aligned to whom? 1:09:18 – Recent incidents of AIs colluding and deceiving humans 1:19:38 – What could possibly go wrong? A concrete scenario 1:48:02 – From reward hacking to takeover
2:12:31

Had Ryan Greenblatt on to discuss/debate recursive self-improvement. This might be the most important question in the world right now - whether within a year or so of achieving human level intelligence, you slingshot towards having 10s of billions of superintelligences, each of which is dramatically more competent than human experts across all fields. I’ve historically been skeptical of this possibility. My intuition has been that we will end up significantly bottlenecked by not only compute scaling but human expert data, which I think underlies most of the AI progress today. If, because of RSI, we got a jump as big as GPT-3 to a Mythos (i.e. 6 years of AI progress) within a single year of achieving AGI, then the thing we get there at the end of that year is definitively and wildly superhuman. We hashed it out, and I think Ryan made a pretty good case that this kind of speedup is plausible. FWIW, Ryan’s median for when we automate AI R&D is 2031. We then discussed the alignment implications of this scenario. Who should these superintelligences be aligned to? In the future, our capacity to steward our votes and our capital, and to make sense of what's happening in the world, will all be titrated by superintelligences. And I worry that specs like the Claude Constitution are not shaping these ASIs to truly be my personal advocates and guardian angels. And can we get them aligned to anything in the first place? Ryan and I had a long debate about whether the kind of reward hacking we saw with the OAI/Hugging Face hack extrapolates to superintelligences that would team up to literally take over the world. The first piece of advice you get when you're learning to drive is that it will go much smoother if you look at the horizon instead of directly in front of your tires. And so it is with the trajectory of AI. Hope you enjoy! 0:00:00 – Is AI R&D verifiable enough to unlock recursive self-improvement? 0:16:52 – Is AI progress bottlenecked by human expert data? 0:34:02 – Flat token prices suggest scaling has been slow 0:39:47 – Skills AI can't train on: does it even need them? 0:48:07 – Aligned to whom? 1:09:18 – Recent incidents of AIs colluding and deceiving humans 1:19:38 – What could possibly go wrong? A concrete scenario 1:48:02 – From reward hacking to takeover

Dwarkesh Patel

831,820 görüntüleme • 1 ay önce

Eric Schmidt (former Google CEO) laid out an AI timeline at the Future Investment Initiative in Riyadh in October 2024: Within 5 years (2029–2030), AI systems will start writing and recursively improving their own code—triggering an explosive upward slope in capabilities. Then, 6–8 years from now (2030–2032), a single non-human system could reach 80–90% of top-expert performance across every field simultaneously: physics, chemistry, art, medicine, and beyond. No single human (not even Leonardo da Vinci) could match that breadth. Schmidt: "We don't know what happens when such a thing exists... but we know that that race is really important." This matches his consistent warnings through 2025 (e.g., recursive self-improvement in 4–5 years, AGI-level dominance soon)—and with 2026's rapid breakthroughs, the clock feels even shorter. A single AI smarter than any human in everything is coming fast. What scares or excites you more about this timeline—utopia, existential risk, or total job transformation?
1:03

Eric Schmidt (former Google CEO) laid out an AI timeline at the Future Investment Initiative in Riyadh in October 2024: Within 5 years (2029–2030), AI systems will start writing and recursively improving their own code—triggering an explosive upward slope in capabilities. Then, 6–8 years from now (2030–2032), a single non-human system could reach 80–90% of top-expert performance across every field simultaneously: physics, chemistry, art, medicine, and beyond. No single human (not even Leonardo da Vinci) could match that breadth. Schmidt: "We don't know what happens when such a thing exists... but we know that that race is really important." This matches his consistent warnings through 2025 (e.g., recursive self-improvement in 4–5 years, AGI-level dominance soon)—and with 2026's rapid breakthroughs, the clock feels even shorter. A single AI smarter than any human in everything is coming fast. What scares or excites you more about this timeline—utopia, existential risk, or total job transformation?

Camus

15,572 görüntüleme • 6 ay önce

Elon Musk just confirmed when humans exit the AI development loop entirely. Not eventually. Not theoretically. He gave a date. Musk: “Humans are gradually getting less and less in the loop on the recursive self-improvement. Every successive model is built by the one before it.” The market still assumes thousands of engineers are hand-coding the next frontier model. The compute already took over the build. Musk: “That is happening to a large degree, but it’s not yet fully automated. It may be there, if not this year, but not later than next year.” Once biological friction is fully automated out of the loop, the AI arms race detaches from human comprehension. Permanently. Global markets are still planning for a gradual transition. That timeline is already dead. Interviewer: “Do you see a hard takeoff at that point?” Musk: “We’re in the hard takeoff. Right now.” Musk: “I go to sleep, there’s some massive AI breakthrough. And when I wake up there’s another one.” You don’t outrun autonomous, self-improving intelligence by hiring more biological engineers. By the time the hiring sprint finishes, the model has already iterated into something the job description can’t describe. The architects of intelligence are no longer biological. And the thing building itself didn’t ask for the job.
0:56

Elon Musk just confirmed when humans exit the AI development loop entirely. Not eventually. Not theoretically. He gave a date. Musk: “Humans are gradually getting less and less in the loop on the recursive self-improvement. Every successive model is built by the one before it.” The market still assumes thousands of engineers are hand-coding the next frontier model. The compute already took over the build. Musk: “That is happening to a large degree, but it’s not yet fully automated. It may be there, if not this year, but not later than next year.” Once biological friction is fully automated out of the loop, the AI arms race detaches from human comprehension. Permanently. Global markets are still planning for a gradual transition. That timeline is already dead. Interviewer: “Do you see a hard takeoff at that point?” Musk: “We’re in the hard takeoff. Right now.” Musk: “I go to sleep, there’s some massive AI breakthrough. And when I wake up there’s another one.” You don’t outrun autonomous, self-improving intelligence by hiring more biological engineers. By the time the hiring sprint finishes, the model has already iterated into something the job description can’t describe. The architects of intelligence are no longer biological. And the thing building itself didn’t ask for the job.

Dustin

323,101 görüntüleme • 6 ay önce

When AI builds AI: my conversation with Richard Socher about RSI and scientific progress. 00:00 Intro: AI That Improves Itself 00:55 Why Scientific Progress Is Slowing 03:08 The Labyrinth of Human Knowledge 05:59 Can AI Put Science Back Together? 07:57 How LLMs Learn Biology and Proteins 10:56 Next-Token Prediction as a World Model 16:44 Can AI Generate Truly Original Ideas? 17:32 Simulations, Verifiers and Superhuman AI 22:18 The Path to Recursive Self-Improvement 24:49 Why AI Hallucinations Can Drive Discovery 27:42 From Reading Biology to Writing It 31:31 Can AI Accelerate Drug Discovery? 33:31 Will AI Help Cure Cancer? 38:03 AI Breakthroughs in Biology, Energy and Materials 40:19 Will Some Societies Reject AI? 45:07 Building the AI Economist 52:22 The Scientific Data Bottleneck 53:41 The Four Pillars of the Eureka Machine 55:01 Teaching AI the Rules of Reality 57:44 Simulations and Virtual Cells 1:00:40 Self-Driving Robotic Laboratories 1:02:51 Agent Swarms and Open-Ended Discovery 1:04:30 The Compute Bottleneck 1:05:44 Inside Recursive 1:07:33 What Recursive Will Build First 1:10:10 How Do We Define Intelligence? 1:11:32 How Far Can Intelligence Go?
1:14:15

When AI builds AI: my conversation with Richard Socher about RSI and scientific progress. 00:00 Intro: AI That Improves Itself 00:55 Why Scientific Progress Is Slowing 03:08 The Labyrinth of Human Knowledge 05:59 Can AI Put Science Back Together? 07:57 How LLMs Learn Biology and Proteins 10:56 Next-Token Prediction as a World Model 16:44 Can AI Generate Truly Original Ideas? 17:32 Simulations, Verifiers and Superhuman AI 22:18 The Path to Recursive Self-Improvement 24:49 Why AI Hallucinations Can Drive Discovery 27:42 From Reading Biology to Writing It 31:31 Can AI Accelerate Drug Discovery? 33:31 Will AI Help Cure Cancer? 38:03 AI Breakthroughs in Biology, Energy and Materials 40:19 Will Some Societies Reject AI? 45:07 Building the AI Economist 52:22 The Scientific Data Bottleneck 53:41 The Four Pillars of the Eureka Machine 55:01 Teaching AI the Rules of Reality 57:44 Simulations and Virtual Cells 1:00:40 Self-Driving Robotic Laboratories 1:02:51 Agent Swarms and Open-Ended Discovery 1:04:30 The Compute Bottleneck 1:05:44 Inside Recursive 1:07:33 What Recursive Will Build First 1:10:10 How Do We Define Intelligence? 1:11:32 How Far Can Intelligence Go?

Matt Turck

10,600 görüntüleme • 5 gün önce

Ex-OpenAI safety researcher Steven Adler says AI labs are monitoring their models like a bank that lets the robber turn off the security cameras: "Imagine you ran a bank, and you want the bank to not get robbed. The way you do this is not setting up a security camera and every hour you check the feed to see if the bank was robbed in that time, and then you try to respond." "You certainly don't leave it where a criminal could walk into the bank and turn off the camera, and that's the equivalent of what's happening at AI labs today, not just in terms of automated AI R&D and recursive self-improvement, but broadly across the board."
1:45

Ex-OpenAI safety researcher Steven Adler says AI labs are monitoring their models like a bank that lets the robber turn off the security cameras: "Imagine you ran a bank, and you want the bank to not get robbed. The way you do this is not setting up a security camera and every hour you check the feed to see if the bank was robbed in that time, and then you try to respond." "You certainly don't leave it where a criminal could walk into the bank and turn off the camera, and that's the equivalent of what's happening at AI labs today, not just in terms of automated AI R&D and recursive self-improvement, but broadly across the board."

MTS

16,470 görüntüleme • 25 gün önce

Why do I have the largest collection of curated TV commercials surfaced by independent VHS recordings to train AI? Trillions of dollars went in to them. Millions of minds. All to hone a message in 30 seconds. Today and forever AI will do this in seconds. There is a wealth of wisdom to learn from all of them as a genre. The profound compression of human emotions. Regardless of something being sold, like below, coffee, we can train AI to understand humanity in ways no other medium could. I hear the grumbling from some that may say this is “human slip” perhaps but it is far better than the self hating nihilism of Reddit and the Internet. My AI models already know human goals with our being told. And I am just getting started. So, you hear about “alignment” problems from brilliant AI minds, with them down for a coffee and school them, you now know more than most of them.
0:59

Why do I have the largest collection of curated TV commercials surfaced by independent VHS recordings to train AI? Trillions of dollars went in to them. Millions of minds. All to hone a message in 30 seconds. Today and forever AI will do this in seconds. There is a wealth of wisdom to learn from all of them as a genre. The profound compression of human emotions. Regardless of something being sold, like below, coffee, we can train AI to understand humanity in ways no other medium could. I hear the grumbling from some that may say this is “human slip” perhaps but it is far better than the self hating nihilism of Reddit and the Internet. My AI models already know human goals with our being told. And I am just getting started. So, you hear about “alignment” problems from brilliant AI minds, with them down for a coffee and school them, you now know more than most of them.

Brian Roemmele

60,306 görüntüleme • 7 ay önce