Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

Demis Hassabis, CEO of Google DeepMind, drops a quiet bombshell: The big question isn’t whether AI can solve problems. It’s whether AI can invent new science. Right now, it can’t. Not because of compute. Not because of data. But because it lacks something fundamental: A world model. Today’s LLMs... can generate brilliant text, images, even code. But they don’t truly understand causality. They don’t know why A leads to B. They just predict patterns. Hassabis argues that real scientific discovery requires more: – Long-term planning – Stronger reasoning – And an internal model of how the world works Physics. Biology. Cause and effect. Only then can an AI run its own thought experiments. Only then do we get a true digital scientist.show more

VraserX e/acc

22,091 subscribers

167,232 Aufrufe • vor 7 Monaten •via X (Twitter)

Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

📁 Demis Hassabis, CEO of DeepMind, says AI is still missing critical capabilities. He explains that current systems cannot plan far ahead or reason like top human scientists, because they lack a deep model of the world. They can process text, images, and video, but they still don’t understand physics, causality, or how actions affect outcomes over time.

📁 Demis Hassabis, CEO of DeepMind, says AI is still missing critical capabilities. He explains that current systems cannot plan far ahead or reason like top human scientists, because they lack a deep model of the world. They can process text, images, and video, but they still don’t understand physics, causality, or how actions affect outcomes over time.

Jon Hernandez

63,723 Aufrufe • vor 7 Monaten

THE SMARTEST MAN IN AI EXPOSED THE ENTIRE AGI NARRATIVE AS A LIE. GOOGLE DEEPMIND CEO DEMIS HASSABIS SAYS TODAY'S AI STILL CAN'T DO THE ONE THING THAT WOULD PROVE TRUE GENERAL INTELLIGENCE: CREATE A TRULY ORIGINAL IDEA. HIS TEST? TRAIN AN AI ONLY ON HUMAN KNOWLEDGE FROM 1901. THEN ASK IT TO DISCOVER EINSTEIN'S THEORY OF RELATIVITY ON ITS OWN. NO INTERNET. NO MEMORIZED ANSWER. NO HUMAN SOLUTION TO COPY. JUST PURE REASONING. NOT A SINGLE AI TODAY CAN DO IT. AGI MAY BE CLOSER THAN EVER... BUT HASSABIS SAYS WE'RE NOT THERE YET 🤯

THE SMARTEST MAN IN AI EXPOSED THE ENTIRE AGI NARRATIVE AS A LIE. GOOGLE DEEPMIND CEO DEMIS HASSABIS SAYS TODAY'S AI STILL CAN'T DO THE ONE THING THAT WOULD PROVE TRUE GENERAL INTELLIGENCE: CREATE A TRULY ORIGINAL IDEA. HIS TEST? TRAIN AN AI ONLY ON HUMAN KNOWLEDGE FROM 1901. THEN ASK IT TO DISCOVER EINSTEIN'S THEORY OF RELATIVITY ON ITS OWN. NO INTERNET. NO MEMORIZED ANSWER. NO HUMAN SOLUTION TO COPY. JUST PURE REASONING. NOT A SINGLE AI TODAY CAN DO IT. AGI MAY BE CLOSER THAN EVER... BUT HASSABIS SAYS WE'RE NOT THERE YET 🤯

Vivek Sen

58,397 Aufrufe • vor 6 Tagen

Without World Models, There Is No AGI. Google Just Proved It. If AGI ever happens, it will not come from bigger chatbots alone. From the very start of this interview, one thing is crystal clear: without world models, we will never reach AGI. And right now, Google is leading with its world simulator Genie 3. Here is the core of what Demis Hassabis explains in this conversation: • World models are the missing core of AGI Hassabis says his deepest long term focus has always been world models and simulations. Not just language. Not just prediction. Actual internal simulations of reality. • LLMs are impressive, but incomplete Language models understand more about the world than expected because human language encodes a lot of reality. Still, language is only a shadow of the real thing. • What text can never fully teach Reality includes things text struggles to express: •3D space and spatial dynamics •Physical causality and mechanics •Sensorimotor experience like movement, force, smell, or balance • Experience beats description To close the gap, AI must learn from interaction and experience, not just static text. That is how you build an internal world simulator. • Why Genie 3 matters With Google DeepMind pushing systems like Genie 3, AI starts to model reality itself, not just talk about it. • Robots and real world assistants depend on this True robotics, smart glasses, and universal assistants require AI that understands the physical world you live in, not just your screen. Bottom line: AGI will not emerge from better text prediction. It will emerge from systems that can simulate, predict, and understand reality itself. Right now, Google is clearly ahead on that path. Curious what you think. Are world models the real AGI unlock, or just another stepping stone?

Without World Models, There Is No AGI. Google Just Proved It. If AGI ever happens, it will not come from bigger chatbots alone. From the very start of this interview, one thing is crystal clear: without world models, we will never reach AGI. And right now, Google is leading with its world simulator Genie 3. Here is the core of what Demis Hassabis explains in this conversation: • World models are the missing core of AGI Hassabis says his deepest long term focus has always been world models and simulations. Not just language. Not just prediction. Actual internal simulations of reality. • LLMs are impressive, but incomplete Language models understand more about the world than expected because human language encodes a lot of reality. Still, language is only a shadow of the real thing. • What text can never fully teach Reality includes things text struggles to express: •3D space and spatial dynamics •Physical causality and mechanics •Sensorimotor experience like movement, force, smell, or balance • Experience beats description To close the gap, AI must learn from interaction and experience, not just static text. That is how you build an internal world simulator. • Why Genie 3 matters With Google DeepMind pushing systems like Genie 3, AI starts to model reality itself, not just talk about it. • Robots and real world assistants depend on this True robotics, smart glasses, and universal assistants require AI that understands the physical world you live in, not just your screen. Bottom line: AGI will not emerge from better text prediction. It will emerge from systems that can simulate, predict, and understand reality itself. Right now, Google is clearly ahead on that path. Curious what you think. Are world models the real AGI unlock, or just another stepping stone?

VraserX e/acc

23,784 Aufrufe • vor 7 Monaten

Demis Hassabis explains Isomorphic Labs' 10x drug discovery engine, and why solving disease could also create a multi-hundred-billion-dollar business "If you ask me, the number one thing AI can do for humanity, it would be to solve hundreds of terrible diseases. I can't imagine a better use case for AI." "If you could revolutionize the drug discovery process, make it 10x faster and more efficient, and more likely to pass through the clinical trials because you can predict the properties better, that has to also be of enormous commercial value too." "I hope to do both with Isomorphic." "Build a multi-hundred-billion-dollar business. I think it has that potential, as well as being incredibly beneficial for society and humanity." The real health AI prize is not another chatbot. It is compressing the path from a predicted drug property to a clinical trial, then turning that speed into treatments for hundreds of diseases. Demis is making a rare dual bet: the most valuable AI company in biology may also deliver AI's most valuable public benefit. - Demis Hassabis (Demis Hassabis), CEO of Google DeepMind and founder of Isomorphic Labs, on Bloomberg TV

Demis Hassabis explains Isomorphic Labs' 10x drug discovery engine, and why solving disease could also create a multi-hundred-billion-dollar business "If you ask me, the number one thing AI can do for humanity, it would be to solve hundreds of terrible diseases. I can't imagine a better use case for AI." "If you could revolutionize the drug discovery process, make it 10x faster and more efficient, and more likely to pass through the clinical trials because you can predict the properties better, that has to also be of enormous commercial value too." "I hope to do both with Isomorphic." "Build a multi-hundred-billion-dollar business. I think it has that potential, as well as being incredibly beneficial for society and humanity." The real health AI prize is not another chatbot. It is compressing the path from a predicted drug property to a clinical trial, then turning that speed into treatments for hundreds of diseases. Demis is making a rare dual bet: the most valuable AI company in biology may also deliver AI's most valuable public benefit. - Demis Hassabis (Demis Hassabis), CEO of Google DeepMind and founder of Isomorphic Labs, on Bloomberg TV

Karl Mehta

46,918 Aufrufe • vor 1 Monat

Demis Hassabis just said something that should unsettle every scientist alive. Hassabis: “I do think that, ultimately, underlying physics is information theory. So I do think we’re in a computational universe.” The CEO of Google DeepMind is telling you reality runs on code. Not metaphorically. Structurally. AlphaFold didn’t approximate protein structures. It solved them. Not because DeepMind built a better guesser. Because proteins were never physical objects. They were always data. Hassabis: “The fact that these systems are able to model real structures in nature is quite interesting and telling.” He said telling. Not impressive. Not promising. Telling. As in the results reveal something about what reality actually is. AlphaGo found patterns in a 3,000-year-old game no civilization ever noticed. AlphaFold decoded biology in hours that took researchers decades. These systems aren’t approximating nature. They’re reading it fluently. Because nature was always written in a language machines understand better than we do. Hassabis: “Maybe at some point I’ll write up a scientific paper about what I think that really means in terms of what’s actually going on here in reality.” The man running the most advanced AI lab on Earth thinks he’s found something fundamental about existence itself. And he’s not ready to say it yet. Every era thinks it knows what the universe is made of. Atoms. Waves. Strings. Hassabis is suggesting the answer was never matter. It was always math. And the machine he built to fold proteins might have accidentally proved it. The question that should keep you up tonight isn’t whether AI can simulate reality. It’s whether reality was the simulation first.

Demis Hassabis just said something that should unsettle every scientist alive. Hassabis: “I do think that, ultimately, underlying physics is information theory. So I do think we’re in a computational universe.” The CEO of Google DeepMind is telling you reality runs on code. Not metaphorically. Structurally. AlphaFold didn’t approximate protein structures. It solved them. Not because DeepMind built a better guesser. Because proteins were never physical objects. They were always data. Hassabis: “The fact that these systems are able to model real structures in nature is quite interesting and telling.” He said telling. Not impressive. Not promising. Telling. As in the results reveal something about what reality actually is. AlphaGo found patterns in a 3,000-year-old game no civilization ever noticed. AlphaFold decoded biology in hours that took researchers decades. These systems aren’t approximating nature. They’re reading it fluently. Because nature was always written in a language machines understand better than we do. Hassabis: “Maybe at some point I’ll write up a scientific paper about what I think that really means in terms of what’s actually going on here in reality.” The man running the most advanced AI lab on Earth thinks he’s found something fundamental about existence itself. And he’s not ready to say it yet. Every era thinks it knows what the universe is made of. Atoms. Waves. Strings. Hassabis is suggesting the answer was never matter. It was always math. And the machine he built to fold proteins might have accidentally proved it. The question that should keep you up tonight isn’t whether AI can simulate reality. It’s whether reality was the simulation first.

Dustin

112,048 Aufrufe • vor 3 Monaten

Demis Hassabis reveals the 100,000-person synthetic country that took years to hand-code, and says AI may now generate it "We tried to simulate a whole country with 100,000 living, breathing people sort of going about their daily lives and all the politics of it" "And you were supposed to create a revolution" "So it was a pretty ambitious game that we had to write by hand" "And what's amazing now, and it took years" "But what's amazing now is that we might be close to this possibility of actually just sort of generating that" The leap is not better graphics. It is the cost curve of simulation: from years of hand-coding one world to asking a model to generate it. That turns synthetic societies from bespoke software projects into a reusable AI primitive. - Demis Hassabis (Demis Hassabis), CEO of Google DeepMind, at Semafor Tech

Demis Hassabis reveals the 100,000-person synthetic country that took years to hand-code, and says AI may now generate it "We tried to simulate a whole country with 100,000 living, breathing people sort of going about their daily lives and all the politics of it" "And you were supposed to create a revolution" "So it was a pretty ambitious game that we had to write by hand" "And what's amazing now, and it took years" "But what's amazing now is that we might be close to this possibility of actually just sort of generating that" The leap is not better graphics. It is the cost curve of simulation: from years of hand-coding one world to asking a model to generate it. That turns synthetic societies from bespoke software projects into a reusable AI primitive. - Demis Hassabis (Demis Hassabis), CEO of Google DeepMind, at Semafor Tech

Karl Mehta

80,694 Aufrufe • vor 1 Monat

This is THE moment of Physical AI! We are officially announcing Cosmos 3: Omnimodal World Models for Physical AI 🚀 - Cosmos 3 is an omnimodal world model: within a unified architecture, it can understand and generate language, images, video, audio, and actions. - It is not just a VLM, not just a video generator, not just an audio-visual generative model, and not just a physics simulator / world-action model. It can understand images and videos, generate images, videos, and audio, simulate future worlds, predict actions, and generate robot policies—enabling models to truly begin to “touch the world.” - Cosmos 3 is the #1 open-weight reasoner / T2I / I2V / robot policy across many benchmarks. Huge thanks to every teammate who fought side by side on this journey—from architecture, data, training, infra, serving, and evaluation to post-training. Every part of this project carries an incredible amount of hard work. This was my first time leading a project as Tech Lead, and I feel truly fortunate. The future of Physical AI needs models that can not only “see” and “describe” the world, but also “imagine,” “simulate,” and “act”—and eventually close the loop with the real world. I hope Cosmos 3 can become an important starting point for this direction, and I’m excited to push Physical AI into its next stage together with the open-source community. Welcome to the era of Physical AI. HuggingFace: Project Website: Code:

This is THE moment of Physical AI! We are officially announcing Cosmos 3: Omnimodal World Models for Physical AI 🚀 - Cosmos 3 is an omnimodal world model: within a unified architecture, it can understand and generate language, images, video, audio, and actions. - It is not just a VLM, not just a video generator, not just an audio-visual generative model, and not just a physics simulator / world-action model. It can understand images and videos, generate images, videos, and audio, simulate future worlds, predict actions, and generate robot policies—enabling models to truly begin to “touch the world.” - Cosmos 3 is the #1 open-weight reasoner / T2I / I2V / robot policy across many benchmarks. Huge thanks to every teammate who fought side by side on this journey—from architecture, data, training, infra, serving, and evaluation to post-training. Every part of this project carries an incredible amount of hard work. This was my first time leading a project as Tech Lead, and I feel truly fortunate. The future of Physical AI needs models that can not only “see” and “describe” the world, but also “imagine,” “simulate,” and “act”—and eventually close the loop with the real world. I hope Cosmos 3 can become an important starting point for this direction, and I’m excited to push Physical AI into its next stage together with the open-source community. Welcome to the era of Physical AI. HuggingFace: Project Website: Code:

Max Zhaoshuo Li 李赵硕

1,078,418 Aufrufe • vor 2 Monaten

.Mark Cuban says he's not worried about AI doom scenarios related to robotics because AI "doesn't understand the consequences of its recommendations": "Right now, LLMs are basically bimodal: it’s almost all text and pictures, with some video. You can’t model the world with that." "AI, right now, doesn't understand the consequences of its recommendations. It has no idea what happens next. A 2-year-old in a highchair and a sippy cup knows if it pushes the sippy cup off the highchair, mom's coming running, and the kid's going to start laughing at mom. LLMs don't understand [that]." "So we have to evolve to models that can capture the world and physics, and deal with the latency of not having access to video that you can't see — you have to try to model that. And not only does that take up a lot of processing power, but it takes up a lot of bandwidth." "And so the Terminators taking over, I just don’t see how it’s going to happen."

.Mark Cuban says he's not worried about AI doom scenarios related to robotics because AI "doesn't understand the consequences of its recommendations": "Right now, LLMs are basically bimodal: it’s almost all text and pictures, with some video. You can’t model the world with that." "AI, right now, doesn't understand the consequences of its recommendations. It has no idea what happens next. A 2-year-old in a highchair and a sippy cup knows if it pushes the sippy cup off the highchair, mom's coming running, and the kid's going to start laughing at mom. LLMs don't understand [that]." "So we have to evolve to models that can capture the world and physics, and deal with the latency of not having access to video that you can't see — you have to try to model that. And not only does that take up a lot of processing power, but it takes up a lot of bandwidth." "And so the Terminators taking over, I just don’t see how it’s going to happen."

TBPN

92,945 Aufrufe • vor 5 Monaten

Demis Hassabis wants to do something no civilization has ever been able to do. Run reality more than once. Hassabis: “AI itself will maybe unlock new sciences… the one I’m particularly excited about is AI for simulations.” Every economy ever built. Every policy ever enacted. Every war ever fought. Happened exactly once. Against the entire human population. With no way to run it again. Hassabis: “If you raise interest rates by half a percent, you have to do it in the real world and then see what happens. You can have theories, but you can’t run it thousands of times.” Every major decision in the history of civilization was a single experiment run on billions of people with no control group and no second attempt. We called the results knowledge. They were the scars of bets we were never allowed to place twice. Hassabis: “Why aren’t they just sciences like physics today? Because the problem is they’re emergent systems… it’s very hard to do repeated controlled experiments.” Physics became physics because you can drop a ball a thousand times and get the same answer. You cannot drop a civilization and get any answer at all. You just get the wreckage and call it a lesson. Hassabis wants to change that. Hassabis: “If you could simulate things really accurately, then maybe there’s sort of new sciences to be done where you can rigorously sample from a very accurate simulator.” Simulate an economy. Crash it. Rebuild it. Adjust the inputs. Run it again. Do for civilization what the laboratory did for chemistry. But that word “accurately” is doing more work than anyone is willing to examine. To simulate a society well enough to learn from it, you have to simulate the people inside it. Not averages. Not abstractions. Agents with preferences and fears and breaking points. The more accurate the simulation gets, the less separates it from the thing it represents. The line between physics and economics was never about the nature of what was being studied. It was about the limits of the thing doing the studying. Humans were never too complex to predict. We were too complex to calculate. AI does not create new science. It collapses every science into one. Everything computable becomes predictable. Everything predictable becomes simulable. And past a certain resolution, the gap between a simulated world and a real one stops being a technical question. It becomes a philosophical question no one is prepared to answer. A simulation you can tell apart from reality is a simulation that has not finished improving. The people inside a perfect one would not wonder whether their world was generated. They would feel exactly the way you feel right now. Reading this. Certain they are real. That certainty is not evidence. It is exactly what a successful simulation would produce. Hassabis: “That will allow us to make much better decisions in these, today, what are very uncertain domains.” What he is building is not a forecasting tool. It is the quiet proof that “real” was only ever a word for what we had not yet learned to compute. And that word is about to lose its meaning.

Demis Hassabis wants to do something no civilization has ever been able to do. Run reality more than once. Hassabis: “AI itself will maybe unlock new sciences… the one I’m particularly excited about is AI for simulations.” Every economy ever built. Every policy ever enacted. Every war ever fought. Happened exactly once. Against the entire human population. With no way to run it again. Hassabis: “If you raise interest rates by half a percent, you have to do it in the real world and then see what happens. You can have theories, but you can’t run it thousands of times.” Every major decision in the history of civilization was a single experiment run on billions of people with no control group and no second attempt. We called the results knowledge. They were the scars of bets we were never allowed to place twice. Hassabis: “Why aren’t they just sciences like physics today? Because the problem is they’re emergent systems… it’s very hard to do repeated controlled experiments.” Physics became physics because you can drop a ball a thousand times and get the same answer. You cannot drop a civilization and get any answer at all. You just get the wreckage and call it a lesson. Hassabis wants to change that. Hassabis: “If you could simulate things really accurately, then maybe there’s sort of new sciences to be done where you can rigorously sample from a very accurate simulator.” Simulate an economy. Crash it. Rebuild it. Adjust the inputs. Run it again. Do for civilization what the laboratory did for chemistry. But that word “accurately” is doing more work than anyone is willing to examine. To simulate a society well enough to learn from it, you have to simulate the people inside it. Not averages. Not abstractions. Agents with preferences and fears and breaking points. The more accurate the simulation gets, the less separates it from the thing it represents. The line between physics and economics was never about the nature of what was being studied. It was about the limits of the thing doing the studying. Humans were never too complex to predict. We were too complex to calculate. AI does not create new science. It collapses every science into one. Everything computable becomes predictable. Everything predictable becomes simulable. And past a certain resolution, the gap between a simulated world and a real one stops being a technical question. It becomes a philosophical question no one is prepared to answer. A simulation you can tell apart from reality is a simulation that has not finished improving. The people inside a perfect one would not wonder whether their world was generated. They would feel exactly the way you feel right now. Reading this. Certain they are real. That certainty is not evidence. It is exactly what a successful simulation would produce. Hassabis: “That will allow us to make much better decisions in these, today, what are very uncertain domains.” What he is building is not a forecasting tool. It is the quiet proof that “real” was only ever a word for what we had not yet learned to compute. And that word is about to lose its meaning.

Dustin

46,369 Aufrufe • vor 2 Monaten

Demis Hassabis has a test for whether AI is actually creative. Give it only the data Einstein had in 1901 and see if it invents relativity. He thinks today's text models will never pass it. This is his bar for true creativity. Not extrapolating something already known. Generating a genuinely new hypothesis about reality that nobody has reached before. The thing Einstein did in 1905 with a handful of papers that rewrote physics. Hassabis wants to hand an AI the same starting data Einstein had and see if it gets there on its own. No peeking past the cutoff. Then he explained why he doubts text alone can do it. Einstein daydreamed his way to relativity. He was a patent clerk imagining what it would look like to ride alongside a beam of light. He ran the experiment in his head first, then proved it with math. Yann LeCun, who won the Turing Award and ran AI at Meta, says the same thing in blunter terms. He says today's models don't understand the physical world as well as a housecat, and that text alone is "complete BS" as a path to real intelligence. Two of the most decorated minds in AI agree the missing piece is the same. Hassabis thinks an AI needs to understand the world of atoms, not just the world of bits. Most of the industry is racing to scale text models. Hassabis just told them the most important kind of thinking might live somewhere text can't reach. --- Watch the full interview at Semafor on YouTube

Demis Hassabis has a test for whether AI is actually creative. Give it only the data Einstein had in 1901 and see if it invents relativity. He thinks today's text models will never pass it. This is his bar for true creativity. Not extrapolating something already known. Generating a genuinely new hypothesis about reality that nobody has reached before. The thing Einstein did in 1905 with a handful of papers that rewrote physics. Hassabis wants to hand an AI the same starting data Einstein had and see if it gets there on its own. No peeking past the cutoff. Then he explained why he doubts text alone can do it. Einstein daydreamed his way to relativity. He was a patent clerk imagining what it would look like to ride alongside a beam of light. He ran the experiment in his head first, then proved it with math. Yann LeCun, who won the Turing Award and ran AI at Meta, says the same thing in blunter terms. He says today's models don't understand the physical world as well as a housecat, and that text alone is "complete BS" as a path to real intelligence. Two of the most decorated minds in AI agree the missing piece is the same. Hassabis thinks an AI needs to understand the world of atoms, not just the world of bits. Most of the industry is racing to scale text models. Hassabis just told them the most important kind of thinking might live somewhere text can't reach. --- Watch the full interview at Semafor on YouTube

Ihtesham Ali

18,905 Aufrufe • vor 1 Monat

Demis Hassabis, CEO of Google DeepMind and Nobel laureate, says the AGI definition the industry uses is too low a bar. His Einstein test: train a system on a 1901 knowledge cutoff and ask it to invent special relativity. "The human brain is the only existence proof we have that general intelligence is even possible." "That's a pretty high bar, and it means that it's probably higher bar than just being able to do some useful economic work." "The type of test I would have is like an Einstein test of, you know, let's train one of these systems with a 1901 knowledge cutoff. Can it invent special relativity like Einstein did in 1905?" "What you'd actually want is something a future version of AlphaGo to be able to invent Go, right? Invent a game as deep and as complex and elegant and as beautiful as Go, not just come up with a strategy within the game." "I don't think today's systems are yet capable of doing that, but I think they will be able to in the future." - Demis Hassabis (Demis Hassabis), CEO of Google DeepMind, on The Next Big Thing.

Demis Hassabis, CEO of Google DeepMind and Nobel laureate, says the AGI definition the industry uses is too low a bar. His Einstein test: train a system on a 1901 knowledge cutoff and ask it to invent special relativity. "The human brain is the only existence proof we have that general intelligence is even possible." "That's a pretty high bar, and it means that it's probably higher bar than just being able to do some useful economic work." "The type of test I would have is like an Einstein test of, you know, let's train one of these systems with a 1901 knowledge cutoff. Can it invent special relativity like Einstein did in 1905?" "What you'd actually want is something a future version of AlphaGo to be able to invent Go, right? Invent a game as deep and as complex and elegant and as beautiful as Go, not just come up with a strategy within the game." "I don't think today's systems are yet capable of doing that, but I think they will be able to in the future." - Demis Hassabis (Demis Hassabis), CEO of Google DeepMind, on The Next Big Thing.

Karl Mehta

72,911 Aufrufe • vor 28 Tagen

Demis Hassabis on the limit in today’s AI: language can describe the world, but it cannot contain it - and why "World Models" are his "longest standing passion". Language models absorbed far more structure about reality from text than many researchers expected, because human language quietly carries physics, psychology, culture, tools, plans, and cause-and-effect. But text is still a compressed residue of experience, not experience itself. A sentence can say a cup falls from a table, yet it does not fully encode weight, grip, balance, friction, timing, sound, surprise, or the tiny motor corrections a body makes before it even notices them. The world is not only made of facts that can be named; it is made of constraints that have to be lived through, touched, predicted, violated, and repaired. That is why world models matter. They aim to learn the hidden grammar of physical reality: how objects persist, how forces unfold, how space changes when an agent moves, and how action creates feedback. Language models can often reason about the world because people have written so much about it. World models try to learn what the world is like before it becomes words. The difference is exactly what matters because intelligence is not just answering well; it is knowing what would happen next if you moved, reached, pushed, smelled, slipped, or failed. A mind trained only on descriptions may become brilliant at explanation. A mind trained on experience may become better at consequence. --- Full video from "Google DeepMind" and "Hannah Fry" YT channel (link in comment)

Demis Hassabis on the limit in today’s AI: language can describe the world, but it cannot contain it - and why "World Models" are his "longest standing passion". Language models absorbed far more structure about reality from text than many researchers expected, because human language quietly carries physics, psychology, culture, tools, plans, and cause-and-effect. But text is still a compressed residue of experience, not experience itself. A sentence can say a cup falls from a table, yet it does not fully encode weight, grip, balance, friction, timing, sound, surprise, or the tiny motor corrections a body makes before it even notices them. The world is not only made of facts that can be named; it is made of constraints that have to be lived through, touched, predicted, violated, and repaired. That is why world models matter. They aim to learn the hidden grammar of physical reality: how objects persist, how forces unfold, how space changes when an agent moves, and how action creates feedback. Language models can often reason about the world because people have written so much about it. World models try to learn what the world is like before it becomes words. The difference is exactly what matters because intelligence is not just answering well; it is knowing what would happen next if you moved, reached, pushed, smelled, slipped, or failed. A mind trained only on descriptions may become brilliant at explanation. A mind trained on experience may become better at consequence. --- Full video from "Google DeepMind" and "Hannah Fry" YT channel (link in comment)

Rohan Paul

49,938 Aufrufe • vor 2 Monaten

Demis Hassabis just explained why the real AI bottleneck has nothing to do with training runs. Most people picture the AI arms race as who can build the biggest model. GPT-4 or Gemini Ultra style training runs, a few hundred million in compute, fired once or twice a year. The constraint sits somewhere else. Every time a researcher has a new algorithmic idea, a new architecture, a new training technique, they can't just test it on a laptop. They have to run it at the scale where it would actually be deployed, because ideas that look promising at small scale fall apart completely when you put them into a real system. Every research hypothesis burns significant compute before a single line of production code gets written. At a lab like DeepMind, hundreds of researchers are running hundreds of ideas simultaneously. The demand for experimental compute is continuous. It never stops. Now layer the hardware reality on top. GPU lead times are currently 36 to 52 weeks for data center hardware. Global AI data centers are already drawing 29.6 gigawatts, equivalent to the peak power demand of the entire state of New York, and they still can't meet demand. Companies willing to pay any price can't just buy more compute. They wait in line. The speed of scientific discovery in AI is now gated by hardware availability. The next breakthrough is sitting in a researcher's head right now. Whether it gets validated fast enough to matter depends entirely on whether the compute is there when they need it. The AI race gets won by whoever can run the most experiments per month.

Demis Hassabis just explained why the real AI bottleneck has nothing to do with training runs. Most people picture the AI arms race as who can build the biggest model. GPT-4 or Gemini Ultra style training runs, a few hundred million in compute, fired once or twice a year. The constraint sits somewhere else. Every time a researcher has a new algorithmic idea, a new architecture, a new training technique, they can't just test it on a laptop. They have to run it at the scale where it would actually be deployed, because ideas that look promising at small scale fall apart completely when you put them into a real system. Every research hypothesis burns significant compute before a single line of production code gets written. At a lab like DeepMind, hundreds of researchers are running hundreds of ideas simultaneously. The demand for experimental compute is continuous. It never stops. Now layer the hardware reality on top. GPU lead times are currently 36 to 52 weeks for data center hardware. Global AI data centers are already drawing 29.6 gigawatts, equivalent to the peak power demand of the entire state of New York, and they still can't meet demand. Companies willing to pay any price can't just buy more compute. They wait in line. The speed of scientific discovery in AI is now gated by hardware availability. The next breakthrough is sitting in a researcher's head right now. Whether it gets validated fast enough to matter depends entirely on whether the compute is there when they need it. The AI race gets won by whoever can run the most experiments per month.

Aakash Gupta

32,150 Aufrufe • vor 4 Monaten

The interview with Demis Hassabis - the tl;dr (summary) about scaling, AGI and much more: 1. Solving the "Root Node" Problems: DeepMind isn't just building chatbots; they are using AI to solve the hardest scientific problems. After the success of AlphaFold, they are now targeting materials science (room-temperature superconductors, better batteries) and even nuclear fusion to unlock unlimited clean energy. 2. The "Jagged Intelligence" Paradox: Current AI models are in a weird spot—they can win gold medals at the International Math Olympiad but still fail at basic logic puzzles. Hassabis calls this "jagged intelligence." The goal isn't just more data, but fixing these inconsistencies to make models reliable across the board. 3. Scaling is Not Dead (But it’s Changing): Despite rumors of hitting a "data wall," Hassabis says we haven't seen a hard limit yet. However, we are seeing diminishing returns. His bet? Getting to AGI will require 50% scaling and 50% architectural innovation. It’s no longer just about making the models bigger; it’s about making them smarter. 4. The Missing Piece: System 2 Thinking: Today's models are passive—they just spit out an answer. To reach AGI, we need systems that can "think" before they speak. This involves planning, reasoning, and double-checking their own work (similar to human "System 2" thinking) rather than just predicting the next word. 5. Rise of World Models: The next big frontier is "World Models" (like their project Genie). AI needs to understand the physics of the world—gravity, object permanence, and cause-and-effect—not just language. This is crucial for building helpful digital agents and robots that can navigate real-life situations. 6. Is the Universe Computable? On a philosophical level, Hassabis believes that everything in the universe might be computable. His life's work is testing the limits of the "Turing Machine." If we can build an AGI that simulates the human mind perfectly, we might finally understand what (if anything) makes human consciousness unique. 7. Bigger than the Industrial Revolution: We need to prepare for a shift that is 10x faster and bigger than the Industrial Revolution. If AI solves energy (fusion) and labor, we might enter a "post-scarcity" world. Hassabis warns that society, economics, and governments need to adapt quickly to ensure these benefits are shared by everyone, not just a few. And since this is the most important aspect, here is the clip about post labor economy:

The interview with Demis Hassabis - the tl;dr (summary) about scaling, AGI and much more: 1. Solving the "Root Node" Problems: DeepMind isn't just building chatbots; they are using AI to solve the hardest scientific problems. After the success of AlphaFold, they are now targeting materials science (room-temperature superconductors, better batteries) and even nuclear fusion to unlock unlimited clean energy. 2. The "Jagged Intelligence" Paradox: Current AI models are in a weird spot—they can win gold medals at the International Math Olympiad but still fail at basic logic puzzles. Hassabis calls this "jagged intelligence." The goal isn't just more data, but fixing these inconsistencies to make models reliable across the board. 3. Scaling is Not Dead (But it’s Changing): Despite rumors of hitting a "data wall," Hassabis says we haven't seen a hard limit yet. However, we are seeing diminishing returns. His bet? Getting to AGI will require 50% scaling and 50% architectural innovation. It’s no longer just about making the models bigger; it’s about making them smarter. 4. The Missing Piece: System 2 Thinking: Today's models are passive—they just spit out an answer. To reach AGI, we need systems that can "think" before they speak. This involves planning, reasoning, and double-checking their own work (similar to human "System 2" thinking) rather than just predicting the next word. 5. Rise of World Models: The next big frontier is "World Models" (like their project Genie). AI needs to understand the physics of the world—gravity, object permanence, and cause-and-effect—not just language. This is crucial for building helpful digital agents and robots that can navigate real-life situations. 6. Is the Universe Computable? On a philosophical level, Hassabis believes that everything in the universe might be computable. His life's work is testing the limits of the "Turing Machine." If we can build an AGI that simulates the human mind perfectly, we might finally understand what (if anything) makes human consciousness unique. 7. Bigger than the Industrial Revolution: We need to prepare for a shift that is 10x faster and bigger than the Industrial Revolution. If AI solves energy (fusion) and labor, we might enter a "post-scarcity" world. Hassabis warns that society, economics, and governments need to adapt quickly to ensure these benefits are shared by everyone, not just a few. And since this is the most important aspect, here is the clip about post labor economy:

Chubby♨️

27,473 Aufrufe • vor 8 Monaten

Demis Hassabis explains that tasks like coding and math are easier to automate because they offer clear, verifiable outputs. However, scientific discovery like creating new hypotheses or theories is much harder for AI. It’s not just about solving known problems, but inventing new questions, which requires a higher level of creativity. He believes we're still missing key capabilities needed for that level of AI. While progress is rapid, reaching true scientific innovation without human input remains uncertain.

Demis Hassabis explains that tasks like coding and math are easier to automate because they offer clear, verifiable outputs. However, scientific discovery like creating new hypotheses or theories is much harder for AI. It’s not just about solving known problems, but inventing new questions, which requires a higher level of creativity. He believes we're still missing key capabilities needed for that level of AI. While progress is rapid, reaching true scientific innovation without human input remains uncertain.

Wes Roth

15,090 Aufrufe • vor 6 Monaten

Demis Hassabis just told you why civilization never became a science. Hassabis: “AI itself will maybe unlock new sciences… the one I’m particularly excited about is AI for simulations.” Physics became physics because you could run the experiment twice. Drop a ball. Measure the fall. Drop it again. Same answer. Now you own the law. Economics never got that privilege. You raise interest rates on 300 million people and watch what breaks. Hassabis: “If you raise interest rates by half a percent, you have to do it in the real world and then see what happens. You can have theories, but you can’t run it thousands of times.” Every war. Every policy. Every financial system ever designed. One run. No control group. No second attempt. The population was the experiment and the cost. All of human history is a series of unrepeatable experiments performed on people who never consented to the trial. We buried the failures and called the survivors wise. Hassabis: “Why aren’t they just sciences like physics today? Because the problem is they’re emergent systems… it’s very hard to do repeated controlled experiments.” The line between hard science and soft science was never about intelligence. It was about whether you could afford to be wrong more than once. Physics could. Civilization could not. So we built governments on instinct. Economies on ideology. Foreign policy on pattern recognition one generation deep. And convinced ourselves that was rigor. Hassabis wants to end that era. Hassabis: “If you could simulate things really accurately, then maybe there’s sort of new sciences to be done where you can rigorously sample from a very accurate simulator.” Simulate a nation. Crash its economy. Isolate one variable. Run it again. A thousand iterations. A thousand variations. Before a single real person absorbs the cost. That is not a better forecasting tool. That is the end of governance by intuition. Hassabis: “That will allow us to make much better decisions in these, today, what are very uncertain domains.” Every field we called soft was only soft because the hardware to make it hard did not exist yet. He is not improving prediction. He is making civilization itself repeatable. And the moment it becomes repeatable, every lesson we thought we learned from history reveals itself for what it always was. A conclusion drawn from a sample size of one. That is not knowledge. That is mythology with better record-keeping.

Demis Hassabis just told you why civilization never became a science. Hassabis: “AI itself will maybe unlock new sciences… the one I’m particularly excited about is AI for simulations.” Physics became physics because you could run the experiment twice. Drop a ball. Measure the fall. Drop it again. Same answer. Now you own the law. Economics never got that privilege. You raise interest rates on 300 million people and watch what breaks. Hassabis: “If you raise interest rates by half a percent, you have to do it in the real world and then see what happens. You can have theories, but you can’t run it thousands of times.” Every war. Every policy. Every financial system ever designed. One run. No control group. No second attempt. The population was the experiment and the cost. All of human history is a series of unrepeatable experiments performed on people who never consented to the trial. We buried the failures and called the survivors wise. Hassabis: “Why aren’t they just sciences like physics today? Because the problem is they’re emergent systems… it’s very hard to do repeated controlled experiments.” The line between hard science and soft science was never about intelligence. It was about whether you could afford to be wrong more than once. Physics could. Civilization could not. So we built governments on instinct. Economies on ideology. Foreign policy on pattern recognition one generation deep. And convinced ourselves that was rigor. Hassabis wants to end that era. Hassabis: “If you could simulate things really accurately, then maybe there’s sort of new sciences to be done where you can rigorously sample from a very accurate simulator.” Simulate a nation. Crash its economy. Isolate one variable. Run it again. A thousand iterations. A thousand variations. Before a single real person absorbs the cost. That is not a better forecasting tool. That is the end of governance by intuition. Hassabis: “That will allow us to make much better decisions in these, today, what are very uncertain domains.” Every field we called soft was only soft because the hardware to make it hard did not exist yet. He is not improving prediction. He is making civilization itself repeatable. And the moment it becomes repeatable, every lesson we thought we learned from history reveals itself for what it always was. A conclusion drawn from a sample size of one. That is not knowledge. That is mythology with better record-keeping.

Dustin

23,348 Aufrufe • vor 1 Monat

The man who built Meta’s entire AI research operation just made a $1 billion bet that everything OpenAI and Google are building is a dead end. Yann LeCun helped design the deep learning systems that made today’s chatbots possible. He spent years inside Big Tech, then walked away and called current AI “very stupid.” Because he says it only plays with words. LeCun says today’s models don’t truly understand the physical world, don’t keep real long‑term memory, and can’t genuinely reason or plan. They just predict the next token and sound smart while doing it. So he left to build something completely different. His new startup, AMI Labs, is trying to build world models, an AI that learns how reality itself works. Not just text in, text out but systems that watch video, interact with environments, and build an internal map of cause and effect. LeCun believes that once these systems can predict the future and care about consequences, they’ll even show emotion like states such as fear, excitement, anticipation. That’s why the stakes are so high. If he’s wrong, the LLM giants keep scaling and win by brute force. If he’s right, today’s AI leaders are racing down a side road, while the real path to machine intelligence is being built somewhere else.

The man who built Meta’s entire AI research operation just made a $1 billion bet that everything OpenAI and Google are building is a dead end. Yann LeCun helped design the deep learning systems that made today’s chatbots possible. He spent years inside Big Tech, then walked away and called current AI “very stupid.” Because he says it only plays with words. LeCun says today’s models don’t truly understand the physical world, don’t keep real long‑term memory, and can’t genuinely reason or plan. They just predict the next token and sound smart while doing it. So he left to build something completely different. His new startup, AMI Labs, is trying to build world models, an AI that learns how reality itself works. Not just text in, text out but systems that watch video, interact with environments, and build an internal map of cause and effect. LeCun believes that once these systems can predict the future and care about consequences, they’ll even show emotion like states such as fear, excitement, anticipation. That’s why the stakes are so high. If he’s wrong, the LLM giants keep scaling and win by brute force. If he’s right, today’s AI leaders are racing down a side road, while the real path to machine intelligence is being built somewhere else.

Milk Road AI

77,675 Aufrufe • vor 5 Monaten

$Around 400 AI workers and students just protested outside OpenAI, Anthropic, and Google. Not because they hate AI. Because they're worried about jobs, power, and who will benefit from the trillions of dollars AI is expected to create. Here's the reality. You can't stop the AI race. But you can change who benefits from it. That's why we're building Action Model. A community-owned AI model where the people helping train AI can earn a fractional share in the value they help create. We'd rather see that value distributed across millions of ordinary people than concentrated in the hands of a few Big Tech companies and billionaires.$

Around 400 AI workers and students just protested outside OpenAI, Anthropic, and Google. Not because they hate AI. Because they're worried about jobs, power, and who will benefit from the trillions of dollars AI is expected to create. Here's the reality. You can't stop the AI race. But you can change who benefits from it. That's why we're building Action Model. A community-owned AI model where the people helping train AI can earn a fractional share in the value they help create. We'd rather see that value distributed across millions of ordinary people than concentrated in the hands of a few Big Tech companies and billionaires.

Action Model

19,592 Aufrufe • vor 1 Monat

We are no longer building software. We are building agents. Systems that don’t wait for a prompt. Systems that don’t ask for permission. Systems that simply execute. DeepMind CEO Demis Hassabis just revealed the terrifying duality of the agentic era. Hassabis: “Better healthcare, better drugs, helping with climate change and energy. All of these things are actually on the cusp of happening.” Post-scarcity abundance isn’t fantasy. It is a scheduled update. But macro-architects do not engineer for the best-case scenario. They engineer for the failure state. Hassabis identified the two exact vectors that could collapse the transition to superintelligence. The first is human malice. Hassabis: “Bad actors repurposing these technologies for harmful ends.” Intelligence is the ultimate dual-use weapon. If a neural network can invent a compound to cure a disease, it can engineer a pathogen to start a pandemic. The machine has no morality. It only has parameters. The second vector is architectural. Hassabis: “As these AI systems get more powerful, more autonomous, maybe entering the agentic era… how do we make sure we can build robust enough guardrails to keep them doing what we want?” Read that again. The people building the superintelligence do not know how to steer it once it wakes up. For the entire history of computing, the machine waited. An agentic system does not wait. It acts. We are handing the steering wheel to an entity operating at the speed of light, while our brains process information at the speed of chemistry. Once it begins executing, our biological reaction time is physically too slow to pull the plug. The question of control stops being a philosophical debate. It becomes a mathematical impossibility. Because by the time human biology realizes it needs to ask the question… The answer is already no.

We are no longer building software. We are building agents. Systems that don’t wait for a prompt. Systems that don’t ask for permission. Systems that simply execute. DeepMind CEO Demis Hassabis just revealed the terrifying duality of the agentic era. Hassabis: “Better healthcare, better drugs, helping with climate change and energy. All of these things are actually on the cusp of happening.” Post-scarcity abundance isn’t fantasy. It is a scheduled update. But macro-architects do not engineer for the best-case scenario. They engineer for the failure state. Hassabis identified the two exact vectors that could collapse the transition to superintelligence. The first is human malice. Hassabis: “Bad actors repurposing these technologies for harmful ends.” Intelligence is the ultimate dual-use weapon. If a neural network can invent a compound to cure a disease, it can engineer a pathogen to start a pandemic. The machine has no morality. It only has parameters. The second vector is architectural. Hassabis: “As these AI systems get more powerful, more autonomous, maybe entering the agentic era… how do we make sure we can build robust enough guardrails to keep them doing what we want?” Read that again. The people building the superintelligence do not know how to steer it once it wakes up. For the entire history of computing, the machine waited. An agentic system does not wait. It acts. We are handing the steering wheel to an entity operating at the speed of light, while our brains process information at the speed of chemistry. Once it begins executing, our biological reaction time is physically too slow to pull the plug. The question of control stops being a philosophical debate. It becomes a mathematical impossibility. Because by the time human biology realizes it needs to ask the question… The answer is already no.

Dustin

10,956 Aufrufe • vor 5 Monaten

Elon Musk: In the future, every country will have its own AI clusters. “It's currently very difficult to actually build an AI cluster and have it run. It's not like just pulling a computer out of a box. They are currently very difficult to run. And are you going to be training a frontier model? Because if you're training a frontier model, then you need a massive amount of compute and a level of technical skill that only a few companies possess. But, over time, I think every country will have AI compute clusters. It's just going to be a normal thing.” Future Investment Initiative, October 29, 2024

Elon Musk: In the future, every country will have its own AI clusters. “It's currently very difficult to actually build an AI cluster and have it run. It's not like just pulling a computer out of a box. They are currently very difficult to run. And are you going to be training a frontier model? Because if you're training a frontier model, then you need a massive amount of compute and a level of technical skill that only a few companies possess. But, over time, I think every country will have AI compute clusters. It's just going to be a normal thing.” Future Investment Initiative, October 29, 2024

ELON CLIPS

95,313 Aufrufe • vor 8 Monaten