Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

Approaching models

The Rizz Bible

56,141 subscribers

37,768 просмотров • 1 год назад •via X (Twitter)

Наука и технологии

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Alexandr Wang says that $200 billion is being spent training AI models in 2024, which is approaching the US defense budget of $800 billion, based on the belief that we are on the path to superintelligence

Alexandr Wang says that $200 billion is being spent training AI models in 2024, which is approaching the US defense budget of $800 billion, based on the belief that we are on the path to superintelligence

Tsarathustra

195,782 просмотров • 1 год назад

Dario Amodei: every few months, new models become smarter than the last, and they’re approaching the point of outperforming most humans at most tasks; anthropic already uses Claude to build their products, train new versions,; and he reaffirms his 90% code thesis.

Dario Amodei: every few months, new models become smarter than the last, and they’re approaching the point of outperforming most humans at most tasks; anthropic already uses Claude to build their products, train new versions,; and he reaffirms his 90% code thesis.

Chubby♨️

29,238 просмотров • 9 месяцев назад

Anthropic CEO, Dario Amodei: "we might be 6-12 months away from models doing all of what software engineers do end-to-end" We're approaching a feedback loop where AI builds better AI But the loop isn't fully closed yet, chip manufacturing and training time still limit speed

Anthropic CEO, Dario Amodei: "we might be 6-12 months away from models doing all of what software engineers do end-to-end" We're approaching a feedback loop where AI builds better AI But the loop isn't fully closed yet, chip manufacturing and training time still limit speed

Haider.

743,599 просмотров • 6 месяцев назад

What are current economic models missing about AGI? How would we know if we were approaching explosive growth? Stanford economist Phil Trammell has been rigorously thinking about the intersection of economic theory and AI (incl. AGI) for over five years, long before the recent surge of interest in large language models. In this episode of Epoch After Hours, Philip Trammell and Epoch AI researcher Anson Ho discuss what economic theory really has to say about the development and impacts of AGI: what current economic models get wrong, the odds of explosive economic growth, what “real GDP” actually measures, and much more! -- Timestamps -- 0:00:00 - Problems with existing work on the economics of AI 0:10:18 - Declining returns to R&D 0:18:28 - What real GDP misses 0:26:57 - Task-based models & AI automation 0:49:32 - The limits of economic theory 1:09:11 - How to detect an economic singularity 1:23:32 - Increasing returns to scale

What are current economic models missing about AGI? How would we know if we were approaching explosive growth? Stanford economist Phil Trammell has been rigorously thinking about the intersection of economic theory and AI (incl. AGI) for over five years, long before the recent surge of interest in large language models. In this episode of Epoch After Hours, Philip Trammell and Epoch AI researcher Anson Ho discuss what economic theory really has to say about the development and impacts of AGI: what current economic models get wrong, the odds of explosive economic growth, what “real GDP” actually measures, and much more! -- Timestamps -- 0:00:00 - Problems with existing work on the economics of AI 0:10:18 - Declining returns to R&D 0:18:28 - What real GDP misses 0:26:57 - Task-based models & AI automation 0:49:32 - The limits of economic theory 1:09:11 - How to detect an economic singularity 1:23:32 - Increasing returns to scale

Epoch AI

103,973 просмотров • 9 месяцев назад

NEW EP: Why AGI Will Happen This Decade with Emad Mostaque The AI revolution is here, but *actually* who controls the intelligence we'll all rely on? In today's episode, & Wayne Sim sit down with Emad of Intelligent Internet to cover: > Who Controls AI? > The Ghibli Effect by ChatGPT-4o > Why Most AI Models Are Inefficient > Universal Basic Intelligence > AI Atlantis: Economic Disruption & Rebirth > Credibly Neutral AI Infrastructure Full episode links below. Timestamps: 00:00 Intro 00:56 Mantle Ad 01:37 Ghibli AI Use 04:50 Replacing Humans with AI? 05:21 Artificial Remote Intelligence 10:14 Emad’s Opinion on Better AI Models 12:25 Concentration of Power with AI? 16:34 Training AI Models 20:46 Compute Demand Collapse? 27:00 Abundance of Open Source AI 30:04 Shift Towards Specialized Models 34:05 RedStone Ad 34:18 Frax Ad 34:44 Approaching Biases in AI 45:16 End of Original Thoughts? 50:30 Intelligent Internet & Reinforced Learning 59:07 Coordinating Incentives

NEW EP: Why AGI Will Happen This Decade with Emad Mostaque The AI revolution is here, but actually who controls the intelligence we'll all rely on? In today's episode, & Wayne Sim sit down with Emad of Intelligent Internet to cover: > Who Controls AI? > The Ghibli Effect by ChatGPT-4o > Why Most AI Models Are Inefficient > Universal Basic Intelligence > AI Atlantis: Economic Disruption & Rebirth > Credibly Neutral AI Infrastructure Full episode links below. Timestamps: 00:00 Intro 00:56 Mantle Ad 01:37 Ghibli AI Use 04:50 Replacing Humans with AI? 05:21 Artificial Remote Intelligence 10:14 Emad’s Opinion on Better AI Models 12:25 Concentration of Power with AI? 16:34 Training AI Models 20:46 Compute Demand Collapse? 27:00 Abundance of Open Source AI 30:04 Shift Towards Specialized Models 34:05 RedStone Ad 34:18 Frax Ad 34:44 Approaching Biases in AI 45:16 End of Original Thoughts? 50:30 Intelligent Internet & Reinforced Learning 59:07 Coordinating Incentives

The Rollup

32,804 просмотров • 1 год назад

"We have this long term vision of turning biology from something which is trial and error and experimental and make it something that looks more like an engineering discipline in the next century." New Episode with Jack Dent ! 0:00 Intro 02:28 Inside Stripe w/ Patrick Collison, at <100 People 08:45 Sam Altman role in Chai Discovery origin 16:20 Chai-2 Antibody Breakthrough 19:24 Approaching biology as an engineering problem 24:45 Using AI models to treat people on an individual scale 26:52 Robotics in labs 27:22 Pharma industry today 29:49 The clinical trial bottleneck 32:53 Personalized biology models & cancer vaccines 36:24 Longevity, peptides, and trillion-dollar drugs 39:02 Does it matter that the government cut exploratory science grants? 41:00 What next-gen Chai models will do 42:08 Can frontier AI labs build this themselves? 43:26 Biology is hard

"We have this long term vision of turning biology from something which is trial and error and experimental and make it something that looks more like an engineering discipline in the next century." New Episode with Jack Dent ! 0:00 Intro 02:28 Inside Stripe w/ Patrick Collison, at <100 People 08:45 Sam Altman role in Chai Discovery origin 16:20 Chai-2 Antibody Breakthrough 19:24 Approaching biology as an engineering problem 24:45 Using AI models to treat people on an individual scale 26:52 Robotics in labs 27:22 Pharma industry today 29:49 The clinical trial bottleneck 32:53 Personalized biology models & cancer vaccines 36:24 Longevity, peptides, and trillion-dollar drugs 39:02 Does it matter that the government cut exploratory science grants? 41:00 What next-gen Chai models will do 42:08 Can frontier AI labs build this themselves? 43:26 Biology is hard

Sabrina Halper

38,838 просмотров • 2 месяцев назад

Here's a recent talk I gave recapping the last 6-12 months of AI progress, why getting perfect models is hard, how labs are likely approaching the next phase of training (for agents), and other interesting tidbits across the reasoning landscape. Topics: 00:00 Introduction & the state of reasoning 05:50 Hillclimbing imperfect evals 09:18 Technical bottlenecks 13:02 Sycophancy 18:08 The Goldilocks Zone 19:28 What comes next? (hint, planning) 26:40 Q&A YouTube etc in replies. Thanks Kyle Corbitt and OpenPipe for hosting me.

Here's a recent talk I gave recapping the last 6-12 months of AI progress, why getting perfect models is hard, how labs are likely approaching the next phase of training (for agents), and other interesting tidbits across the reasoning landscape. Topics: 00:00 Introduction & the state of reasoning 05:50 Hillclimbing imperfect evals 09:18 Technical bottlenecks 13:02 Sycophancy 18:08 The Goldilocks Zone 19:28 What comes next? (hint, planning) 26:40 Q&A YouTube etc in replies. Thanks Kyle Corbitt and OpenPipe for hosting me.

Nathan Lambert

89,519 просмотров • 1 год назад

In Episode 11 of Layer One, Kelvin Sparks and special co-host Ava Labs 🔺 President John Wu 🔺 are joined by Matt Hougan, CIO of Bitwise, to discuss the launch of the Avalanche ETP and how institutional investors are approaching crypto exposure. The group also touched on: • Tokenized real-world assets • ETFs in onboarding capital • Why this cycle may differ from previous ones OUTLINE 00:00 - Introduction 01:40 - Avalanche ETP Overview 02:45 - ETF Differentiation 04:34 - Staking and Yield 06:19 - Institutional Strategy 09:04 - Competing L1 Models 14:22 - Bitcoin Price Thesis 18:49 - Dual Market Narratives 21:57 - Institutional Capital Flows 30:02 - AI and Crypto Future

In Episode 11 of Layer One, Kelvin Sparks and special co-host Ava Labs 🔺 President John Wu 🔺 are joined by Matt Hougan, CIO of Bitwise, to discuss the launch of the Avalanche ETP and how institutional investors are approaching crypto exposure. The group also touched on: • Tokenized real-world assets • ETFs in onboarding capital • Why this cycle may differ from previous ones OUTLINE 00:00 - Introduction 01:40 - Avalanche ETP Overview 02:45 - ETF Differentiation 04:34 - Staking and Yield 06:19 - Institutional Strategy 09:04 - Competing L1 Models 14:22 - Bitcoin Price Thesis 18:49 - Dual Market Narratives 21:57 - Institutional Capital Flows 30:02 - AI and Crypto Future

The Block Media

21,256 просмотров • 3 месяцев назад

Everyone is talking about AI models, but almost no one is talking about data ownership. For years, centralized systems have dominated data pipelines, opaque processes, hidden biases, and zero contributor visibility. We call it innovation…, but in reality, it’s a black box. And that’s exactly how we ended up facing model degradation risks today. Because here’s the truth: AI is only as good as the data it learns from. That’s why Sovereign AI Data matters. To me, it means control, traceability, and accountability. Perle Labs is approaching this differently: ◈ On-chain provenance → Every data point is verifiable ◈ Expert-driven input → Quality over quantity ◈ Reputation systems → Trust becomes measurable This isn’t just a technical upgrade. It’s a paradigm shift. Instead of: Anonymous contributors ➠Unverified datasets ➠Centralized ownership ➠We move toward: Transparent data flows, Credible expertise and Distributed value creation If AI is the brain, then data is the memory. And right now most systems are learning from memory, which they can’t verify. Perle is solving that. While others focus on scaling models… This is about fixing the foundation they stand on. And foundations decide everything. #PerleAI #ToPerle participating in Perle Labs community campaign

Everyone is talking about AI models, but almost no one is talking about data ownership. For years, centralized systems have dominated data pipelines, opaque processes, hidden biases, and zero contributor visibility. We call it innovation…, but in reality, it’s a black box. And that’s exactly how we ended up facing model degradation risks today. Because here’s the truth: AI is only as good as the data it learns from. That’s why Sovereign AI Data matters. To me, it means control, traceability, and accountability. Perle Labs is approaching this differently: ◈ On-chain provenance → Every data point is verifiable ◈ Expert-driven input → Quality over quantity ◈ Reputation systems → Trust becomes measurable This isn’t just a technical upgrade. It’s a paradigm shift. Instead of: Anonymous contributors ➠Unverified datasets ➠Centralized ownership ➠We move toward: Transparent data flows, Credible expertise and Distributed value creation If AI is the brain, then data is the memory. And right now most systems are learning from memory, which they can’t verify. Perle is solving that. While others focus on scaling models… This is about fixing the foundation they stand on. And foundations decide everything. #PerleAI #ToPerle participating in Perle Labs community campaign

Big_Wealth❤️‍🔥

11,486 просмотров • 3 месяцев назад

“What if the most important AI infrastructure isn’t the biggest model — but the one you can actually trust?” Sachi Miyasaki sits down with Karan Sirdesai Karan, Founder of Mira, to explore how Mira is approaching decentralized AI from a more focused angle: making machine intelligence reliable enough for real-world use. They discuss why Mira is not trying to become a full-stack AI protocol, how its approach differs from networks like Bittensor, Ritual, and Sahara, and why focusing on trust and verification could make Mira a stronger fit inside the wider AI infrastructure stack. The episode also breaks down Mira’s reliability architecture, from its core verification layer to node-level consensus, and why the next evolution of AI APIs may not be about simply accessing more models, but about knowing whether the answers you receive can actually be trusted.

“What if the most important AI infrastructure isn’t the biggest model — but the one you can actually trust?” Sachi Miyasaki sits down with Karan Sirdesai Karan, Founder of Mira, to explore how Mira is approaching decentralized AI from a more focused angle: making machine intelligence reliable enough for real-world use. They discuss why Mira is not trying to become a full-stack AI protocol, how its approach differs from networks like Bittensor, Ritual, and Sahara, and why focusing on trust and verification could make Mira a stronger fit inside the wider AI infrastructure stack. The episode also breaks down Mira’s reliability architecture, from its core verification layer to node-level consensus, and why the next evolution of AI APIs may not be about simply accessing more models, but about knowing whether the answers you receive can actually be trusted.

BeaconLayer Podcast

77,966 просмотров • 5 дней назад

What happens if a star gets too close to a black hole? The black hole can rip it apart -- but how? It's not the high gravitational attraction itself that's the problem -- it's the difference in gravitational pull across the star that creates the destruction. In the featured animated video illustrating this disintegration, you first see a star approaching the black hole. Increasing in orbital speed, the star's outer atmosphere is ripped away during closest approach. Much of the star's atmosphere disperses into deep space, but some continues to orbit the black hole and forms an accretion disk. The animation then takes you into the accretion disk while looking toward the black hole. Including the strange visual effects of gravitational lensing, you can even see the far side of the disk. Finally, you look along one of the jets being expelled along the spin axis. Theoretical models indicate that these jets not only expel energetic gas, but also create energetic neutrinos -- one of which may have been seen recently on Earth. Video Illustration Credit: DESY, Science Communication Lab

What happens if a star gets too close to a black hole? The black hole can rip it apart -- but how? It's not the high gravitational attraction itself that's the problem -- it's the difference in gravitational pull across the star that creates the destruction. In the featured animated video illustrating this disintegration, you first see a star approaching the black hole. Increasing in orbital speed, the star's outer atmosphere is ripped away during closest approach. Much of the star's atmosphere disperses into deep space, but some continues to orbit the black hole and forms an accretion disk. The animation then takes you into the accretion disk while looking toward the black hole. Including the strange visual effects of gravitational lensing, you can even see the far side of the disk. Finally, you look along one of the jets being expelled along the spin axis. Theoretical models indicate that these jets not only expel energetic gas, but also create energetic neutrinos -- one of which may have been seen recently on Earth. Video Illustration Credit: DESY, Science Communication Lab

NASA Astronomy Picture of the Day

181,453 просмотров • 2 лет назад

David Friedberg says Anthropic asked big pharma for their data and nearly everyone said no "There's been an effort by Anthropic to sign up life sciences companies to contribute to a new life sciences focused model. They're approaching these large companies with large proprietary data sets and saying, if you share your data, we will give you early access, some sort of proprietary value. Sign this NDA and you can participate with us." "I think nearly everyone I've spoken with has woken up to the fact that they are trying to commoditize everyone's business. If all of the tens of billions of dollars you have invested in experiments and product development, and you've generated all of this proprietary data along the way, that data is a true asset of your organization. It's an asset that you've spent billions of dollars developing." "And by handing it over to a model company to then combine with other people's data, you are commoditizing the one core differentiation that you have. And so everyone is largely saying no." "I think what everyone's realizing is they're better off developing their own weights and their own models using either an open source basis or there might be some intermediary business model that evolves."

David Friedberg says Anthropic asked big pharma for their data and nearly everyone said no "There's been an effort by Anthropic to sign up life sciences companies to contribute to a new life sciences focused model. They're approaching these large companies with large proprietary data sets and saying, if you share your data, we will give you early access, some sort of proprietary value. Sign this NDA and you can participate with us." "I think nearly everyone I've spoken with has woken up to the fact that they are trying to commoditize everyone's business. If all of the tens of billions of dollars you have invested in experiments and product development, and you've generated all of this proprietary data along the way, that data is a true asset of your organization. It's an asset that you've spent billions of dollars developing." "And by handing it over to a model company to then combine with other people's data, you are commoditizing the one core differentiation that you have. And so everyone is largely saying no." "I think what everyone's realizing is they're better off developing their own weights and their own models using either an open source basis or there might be some intermediary business model that evolves."

dnap

285,495 просмотров • 16 дней назад

Maria closed by raising one of the darkest topics of the interview—the reaction to Charlie Kirk’s assassination. Some people openly celebrated the murder, prompting her to ask whether America is sliding into something irreversible. Armstrong said his models had already projected civil unrest from 2022 onward, tied to an economic decline that won’t bottom until 2028. The assassination, he warned, “was a huge catalyst.” He argued that media polarization poured gasoline on the fire and pointed to the Weimar Republic as a cautionary tale, where corruption and ethnic division paved the way for extremism. “Once you divide a nation ethnically,” he said, “there’s no going back.” Then came his most sobering forecast yet: “our computer says all these forms of government, republics, they’re going to collapse by 2032.” Armstrong framed it as part of a larger historic cycle. Monarchies collapsed centuries ago, and now republics are approaching the same fate. The takeaway was grim but clarifying: this isn’t about one man’s death. It’s about the systemic unraveling of civility, trust, and governance—accelerated by leaders and media who weaponize division. His conclusion was chilling, a final reminder of where unchecked corruption and division ultimately lead: collapse by 2032.

Maria closed by raising one of the darkest topics of the interview—the reaction to Charlie Kirk’s assassination. Some people openly celebrated the murder, prompting her to ask whether America is sliding into something irreversible. Armstrong said his models had already projected civil unrest from 2022 onward, tied to an economic decline that won’t bottom until 2028. The assassination, he warned, “was a huge catalyst.” He argued that media polarization poured gasoline on the fire and pointed to the Weimar Republic as a cautionary tale, where corruption and ethnic division paved the way for extremism. “Once you divide a nation ethnically,” he said, “there’s no going back.” Then came his most sobering forecast yet: “our computer says all these forms of government, republics, they’re going to collapse by 2032.” Armstrong framed it as part of a larger historic cycle. Monarchies collapsed centuries ago, and now republics are approaching the same fate. The takeaway was grim but clarifying: this isn’t about one man’s death. It’s about the systemic unraveling of civility, trust, and governance—accelerated by leaders and media who weaponize division. His conclusion was chilling, a final reminder of where unchecked corruption and division ultimately lead: collapse by 2032.

Vigilant Fox 🦊

34,456 просмотров • 10 месяцев назад

Every major turning point in human history had a moment where the people closest to it understood what was coming and everyone else didn’t. The printing press. The atomic bomb. The internet. Dario Amodei is trying to close that gap. Most people still aren’t listening. Amodei: “My fundamental view is that AI has been on an exponential for the last ten years and as part of a sort of Moore’s law for intelligence.” Not a metaphor. A measured curve. Not slowing. Accelerating. We are “well advanced on that curve” with a “small number of years” remaining before AI surpasses human cognitive capability across most things. Amodei: “We’re increasingly close to what I’ve called a country of geniuses in a data center.” Not one system. A coordinated set of AI agents, each more capable than most humans at most things, running in parallel, never sleeping, never losing focus, coordinating at speeds no biological intelligence can match. The ceiling on human progress has always been simple. Genius is rare and time is finite. That constraint is gone. We are not approaching the ceiling of intelligence. We are approaching the ceiling of intelligence that biology can produce. Those are not the same ceiling. The human brain is constrained by evolution, energy, skull size, and lifespan. AI has none of those limits. We have no framework for what intelligence looks like when you remove every biological constraint that shaped ours. Every tool we use to comprehend it is built from a mind it will surpass. Amodei: “AI models surpassing the cognitive capabilities of most humans for most things.” The upsides are “absolutely staggering.” So are the consequences. Displacement. Misuse. A period of disruption that reshapes how most people work, earn, and find purpose before the benefits reach them. The same exponential that produces the cures produces the chaos. They arrive together. That’s the optimistic read. The other read is this. We are building it without international standards in place. Racing to deploy before solving displacement or misuse. And once it’s operating beyond our ability to fully comprehend or follow, managing the disruption stops being something we do. It becomes something we experience while hoping the intelligence we created decides keeping us around still serves its objectives. Every turning point in history looks inevitable in hindsight. The people inside the moment never saw it that cleanly. We’re inside the moment.

Every major turning point in human history had a moment where the people closest to it understood what was coming and everyone else didn’t. The printing press. The atomic bomb. The internet. Dario Amodei is trying to close that gap. Most people still aren’t listening. Amodei: “My fundamental view is that AI has been on an exponential for the last ten years and as part of a sort of Moore’s law for intelligence.” Not a metaphor. A measured curve. Not slowing. Accelerating. We are “well advanced on that curve” with a “small number of years” remaining before AI surpasses human cognitive capability across most things. Amodei: “We’re increasingly close to what I’ve called a country of geniuses in a data center.” Not one system. A coordinated set of AI agents, each more capable than most humans at most things, running in parallel, never sleeping, never losing focus, coordinating at speeds no biological intelligence can match. The ceiling on human progress has always been simple. Genius is rare and time is finite. That constraint is gone. We are not approaching the ceiling of intelligence. We are approaching the ceiling of intelligence that biology can produce. Those are not the same ceiling. The human brain is constrained by evolution, energy, skull size, and lifespan. AI has none of those limits. We have no framework for what intelligence looks like when you remove every biological constraint that shaped ours. Every tool we use to comprehend it is built from a mind it will surpass. Amodei: “AI models surpassing the cognitive capabilities of most humans for most things.” The upsides are “absolutely staggering.” So are the consequences. Displacement. Misuse. A period of disruption that reshapes how most people work, earn, and find purpose before the benefits reach them. The same exponential that produces the cures produces the chaos. They arrive together. That’s the optimistic read. The other read is this. We are building it without international standards in place. Racing to deploy before solving displacement or misuse. And once it’s operating beyond our ability to fully comprehend or follow, managing the disruption stops being something we do. It becomes something we experience while hoping the intelligence we created decides keeping us around still serves its objectives. Every turning point in history looks inevitable in hindsight. The people inside the moment never saw it that cleanly. We’re inside the moment.

Dustin

47,204 просмотров • 5 месяцев назад

This is an ape ("Kanzi") playing Minecraft! A fascinating experiment on non-human biological neural networks 🙉 I've been teaching AI to play Minecraft for too long. There're so many similar techniques that the ape trainers used: - In-context reinforcement learning: Kanzi gets a fruit or peanut whenever he hits a marked milestone in the game, incentivizing him to follow the in-game guides. - RLHF: Kanzi doesn't understand much language, but he can see the trainers cheering him on, and he occasionally cheers them back! That gives him a strong signal that he's on the right track. - Imitation learning: the trainers show Kanzi just 1 demonstration of how to do a task, and he immediately grasps the concepts. It's much more efficient than using rewards alone. - Curriculum learning: they start with very simple environments to gradually teach Kanzi the controls. At the end, Kanzi is able to navigate complex caves, mazes, and the Nether. It also amazes me how strong the ape's vision system is. Kanzi never saw Minecraft in his life, and for sure his ancestors didn't either. Yet he rapidly adapts to Minecraft's texture and physics, which are dramatically different from the natural world. This level of generalization is far beyond what our most powerful vision models can do today. We are right in the thick of Moravec's paradox again: our best AIs are approaching human level on understanding language, but far behind animals on parsing pixels. From YouTube channel "ChrisDaCow": Researchers are from the Ape Initiative, a non-profit org.

This is an ape ("Kanzi") playing Minecraft! A fascinating experiment on non-human biological neural networks 🙉 I've been teaching AI to play Minecraft for too long. There're so many similar techniques that the ape trainers used: - In-context reinforcement learning: Kanzi gets a fruit or peanut whenever he hits a marked milestone in the game, incentivizing him to follow the in-game guides. - RLHF: Kanzi doesn't understand much language, but he can see the trainers cheering him on, and he occasionally cheers them back! That gives him a strong signal that he's on the right track. - Imitation learning: the trainers show Kanzi just 1 demonstration of how to do a task, and he immediately grasps the concepts. It's much more efficient than using rewards alone. - Curriculum learning: they start with very simple environments to gradually teach Kanzi the controls. At the end, Kanzi is able to navigate complex caves, mazes, and the Nether. It also amazes me how strong the ape's vision system is. Kanzi never saw Minecraft in his life, and for sure his ancestors didn't either. Yet he rapidly adapts to Minecraft's texture and physics, which are dramatically different from the natural world. This level of generalization is far beyond what our most powerful vision models can do today. We are right in the thick of Moravec's paradox again: our best AIs are approaching human level on understanding language, but far behind animals on parsing pixels. From YouTube channel "ChrisDaCow": Researchers are from the Ape Initiative, a non-profit org.

Jim Fan

745,963 просмотров • 2 лет назад

At 14, Jim Simons got a job putting away stock in a basement. He was so bad, they demoted him to floor sweeper. When he said he wanted to study math at MIT, his bosses laughed. He went on to make $100+ billion in trading profits. More than Buffett, Soros, & Dalio combined. In 2010, he spent 60 minutes at MIT explaining how a mathematician became the world's greatest trader: "They thought that was the funniest thing they had ever heard. The guy who couldn't remember where to put the sheep manure is going to be a mathematician at MIT." He applied to MIT. Got accepted. Studied mathematics. It went all right. After graduating, Simons took a job at the Institute for Defense Analysis. Secret government work. Good pay. Half your time on their work, half on your own mathematics. Then the Vietnam War happened. General Maxwell Taylor, the head of the organization, wrote an article in the New York Times about how victory was days away. Simons disagreed. He wrote a letter to the Times expressing that view. "They kindly published it." A few months later, a reporter asked to interview him about people who work for the defense department but oppose the war. Simons agreed. Told his local boss afterward. "He said, 'You did what?' And he picked up the phone and called General Maxwell Taylor." Silence on the other end. "He hung up. Looked at me. Said, 'You're fired.'" "I said, 'I'm fired? I'm a permanent member.'" "He said, 'I'll tell you the difference between a temporary member and a permanent member. A temporary member has a contract.'" "It was the first time, and happily the last time, I was ever fired." On starting Renaissance: Simons left mathematics at 38. Frustrated. Stuck on a problem he couldn't solve. He had some money from an investment that finally paid off. He invested that money. Found out he wasn't bad at it. He brought in the best modeler he knew. A guy named Lenny Baum. "Lenny started making models. But then he seemed to get less interested in models and more interested in reading the news." "Then he started having opinions on what was going to go up and what was going to go down. And he was right enough times." "I said okay, to hell with the modeling. Let's just try to make some money." What happened next: "We multiplied our investors' money by 12 in two years." "We were incredibly lucky." But in the back of his mind, he knew models were the answer. "If you're doing fundamental trading, one morning you come in, you feel like a genius. Your positions are all your way. You think, God, I'm really smart." "The next day you come in, they've gone against you, and you feel like an idiot." "It just didn't seem like a way to live your life." In 1988, he made the decision: 100% models. No human override. "Some firms say they have models. What they typically mean is the model advises the trader what to do. If he likes the advice, he'll take it. If he doesn't, he won't." "That's not science. You can't simulate that. How were you feeling when you got out of bed 13 years ago? Did you like what the model said?" "If you're going to trade using models, you slavishly use the models. You do whatever the hell it says. No matter how smart or dumb you might think it is at that moment." "That turned out to be a wonderful decision." On the secret sauce: "People always ask me, what's the secret? We're not the only quant firm in the world. But we seem to have done better than anybody." "The real secret sauce is that we start with great scientists. First-class people who've done first-class work." "Second, we provide people with a great infrastructure." "The most important thing is an open atmosphere. Everybody knows what everybody else is doing. No compartmentalization." "And people get paid based on the overall profits. Not just on your work. Everyone has an interest in everyone else's success." "Those policies, no one of which seems so remarkable, turn out to be a pretty winning combination." On guiding principles: His wife told him to end with values. He said he wasn't sure he had any. "She assured me that I had some values, if only I could think hard about them." Here's what he came up with: 1. Do something new. "I don't like to run with the pack. For one thing, I'm not such a fast runner." "If you're one of n people all working on the same problem, I'd be last. But if you can think of a new problem that other people aren't working on, maybe that'll give you a chance." 2. Collaborate with the best people you possibly can. "That gives you some reach and some scope. And it's also fun to work with terrific people." 3. Be guided by beauty. "What's aesthetic is doing it right. Getting the right kind of people. Approaching the problem and doing it right." "It's a beautiful thing to do something right." 4. Don't give up. "Sometimes it's appropriate to be trying to do something for a hell of a long time." 5. Hope for some good luck.

At 14, Jim Simons got a job putting away stock in a basement. He was so bad, they demoted him to floor sweeper. When he said he wanted to study math at MIT, his bosses laughed. He went on to make $100+ billion in trading profits. More than Buffett, Soros, & Dalio combined. In 2010, he spent 60 minutes at MIT explaining how a mathematician became the world's greatest trader: "They thought that was the funniest thing they had ever heard. The guy who couldn't remember where to put the sheep manure is going to be a mathematician at MIT." He applied to MIT. Got accepted. Studied mathematics. It went all right. After graduating, Simons took a job at the Institute for Defense Analysis. Secret government work. Good pay. Half your time on their work, half on your own mathematics. Then the Vietnam War happened. General Maxwell Taylor, the head of the organization, wrote an article in the New York Times about how victory was days away. Simons disagreed. He wrote a letter to the Times expressing that view. "They kindly published it." A few months later, a reporter asked to interview him about people who work for the defense department but oppose the war. Simons agreed. Told his local boss afterward. "He said, 'You did what?' And he picked up the phone and called General Maxwell Taylor." Silence on the other end. "He hung up. Looked at me. Said, 'You're fired.'" "I said, 'I'm fired? I'm a permanent member.'" "He said, 'I'll tell you the difference between a temporary member and a permanent member. A temporary member has a contract.'" "It was the first time, and happily the last time, I was ever fired." On starting Renaissance: Simons left mathematics at 38. Frustrated. Stuck on a problem he couldn't solve. He had some money from an investment that finally paid off. He invested that money. Found out he wasn't bad at it. He brought in the best modeler he knew. A guy named Lenny Baum. "Lenny started making models. But then he seemed to get less interested in models and more interested in reading the news." "Then he started having opinions on what was going to go up and what was going to go down. And he was right enough times." "I said okay, to hell with the modeling. Let's just try to make some money." What happened next: "We multiplied our investors' money by 12 in two years." "We were incredibly lucky." But in the back of his mind, he knew models were the answer. "If you're doing fundamental trading, one morning you come in, you feel like a genius. Your positions are all your way. You think, God, I'm really smart." "The next day you come in, they've gone against you, and you feel like an idiot." "It just didn't seem like a way to live your life." In 1988, he made the decision: 100% models. No human override. "Some firms say they have models. What they typically mean is the model advises the trader what to do. If he likes the advice, he'll take it. If he doesn't, he won't." "That's not science. You can't simulate that. How were you feeling when you got out of bed 13 years ago? Did you like what the model said?" "If you're going to trade using models, you slavishly use the models. You do whatever the hell it says. No matter how smart or dumb you might think it is at that moment." "That turned out to be a wonderful decision." On the secret sauce: "People always ask me, what's the secret? We're not the only quant firm in the world. But we seem to have done better than anybody." "The real secret sauce is that we start with great scientists. First-class people who've done first-class work." "Second, we provide people with a great infrastructure." "The most important thing is an open atmosphere. Everybody knows what everybody else is doing. No compartmentalization." "And people get paid based on the overall profits. Not just on your work. Everyone has an interest in everyone else's success." "Those policies, no one of which seems so remarkable, turn out to be a pretty winning combination." On guiding principles: His wife told him to end with values. He said he wasn't sure he had any. "She assured me that I had some values, if only I could think hard about them." Here's what he came up with: 1. Do something new. "I don't like to run with the pack. For one thing, I'm not such a fast runner." "If you're one of n people all working on the same problem, I'd be last. But if you can think of a new problem that other people aren't working on, maybe that'll give you a chance." 2. Collaborate with the best people you possibly can. "That gives you some reach and some scope. And it's also fun to work with terrific people." 3. Be guided by beauty. "What's aesthetic is doing it right. Getting the right kind of people. Approaching the problem and doing it right." "It's a beautiful thing to do something right." 4. Don't give up. "Sometimes it's appropriate to be trying to do something for a hell of a long time." 5. Hope for some good luck.

Jaynit

73,562 просмотров • 3 месяцев назад

Elon Musk just said the AI community is misunderstanding the math of superintelligence by two orders of magnitude. Not slightly off. Not directionally wrong. A hundred times off. Musk: “Most people in the AI community don’t yet understand. The intelligence density potential is vastly greater than what we’re currently experiencing.” Everyone is focused on the hardware race. Bigger data centers. More GPUs. Nuclear power plants built to feed the compute. That’s half the equation. Musk: “I think we’re off by two orders of magnitude in terms of intelligence density per gigabyte. That’s just algorithmic improvement. Same computer.” Read that carefully. Not more hardware. Not more energy. Not more capital. The same machine. A hundred times smarter. Through software alone. That’s before the hardware improvements compound on top of it. Musk: “And the computers are getting better. That’s why I think it is a 10x improvement per year type thing. 1,000 percent.” A thousand percent compounding annual growth rate in raw intelligence. A system that becomes 10x more capable every twelve months doesn’t follow a linear curve. It doesn’t follow an exponential curve that human intuition can track. It follows a curve that human intuition cannot simulate at all. In year one it’s 10x smarter. In year two it’s 100x. In year three it’s 1,000x. At that point, the gap between that system and a human brain is wider than the gap between a human brain and a calculator. This is the math the public isn’t running. The models aren’t just getting better. They are compounding on themselves at a rate that makes every previous technology curve look flat. Musk: “The intelligence density potential is vastly greater than what we’re currently experiencing.” We aren’t approaching superintelligence on the timeline most people imagine. We are already inside the curve.

Elon Musk just said the AI community is misunderstanding the math of superintelligence by two orders of magnitude. Not slightly off. Not directionally wrong. A hundred times off. Musk: “Most people in the AI community don’t yet understand. The intelligence density potential is vastly greater than what we’re currently experiencing.” Everyone is focused on the hardware race. Bigger data centers. More GPUs. Nuclear power plants built to feed the compute. That’s half the equation. Musk: “I think we’re off by two orders of magnitude in terms of intelligence density per gigabyte. That’s just algorithmic improvement. Same computer.” Read that carefully. Not more hardware. Not more energy. Not more capital. The same machine. A hundred times smarter. Through software alone. That’s before the hardware improvements compound on top of it. Musk: “And the computers are getting better. That’s why I think it is a 10x improvement per year type thing. 1,000 percent.” A thousand percent compounding annual growth rate in raw intelligence. A system that becomes 10x more capable every twelve months doesn’t follow a linear curve. It doesn’t follow an exponential curve that human intuition can track. It follows a curve that human intuition cannot simulate at all. In year one it’s 10x smarter. In year two it’s 100x. In year three it’s 1,000x. At that point, the gap between that system and a human brain is wider than the gap between a human brain and a calculator. This is the math the public isn’t running. The models aren’t just getting better. They are compounding on themselves at a rate that makes every previous technology curve look flat. Musk: “The intelligence density potential is vastly greater than what we’re currently experiencing.” We aren’t approaching superintelligence on the timeline most people imagine. We are already inside the curve.

Dustin

1,553,512 просмотров • 4 месяцев назад

Goodmayes, east London. In the latest grim dispatch from the front lines of demographic replacement, footage captures the raw chaos unfolding on the High Road after a “man” burst out of a mobile phone shop brandishing a freshly snatched Samsung device. The suspect another product of the “transformed & enriched streets of England, had been aggressively approaching multiple people in the vicinity, including a visibly distressed woman, before or during the theft. Bystanders, overwhelmingly foreign faces in the blurred melee scattered in panic as the thief tried to make his getaway before being apprehended by the police. The shop’s “bossman” taking matters into his own hands, gave chase and collared the culprit further down the road. Just a proprietor defending his patch in a postcode where phone snatchings and street predation have become the new normal. Even the thief’s choice of a Samsung drew mockery London’s imported phone bandits apparently hold out for iPhones these days, tossing lesser models like yesterday’s curry. Goodmayes and neighbouring Ilford, once solid English working-class territory, have been colonised by waves of South Asian, African, and other non-native inflows. What the BBC euphemistically calls “diversity” looks, on the ground, like eroded trust, women unable to walk without harassment, and natives dodging the fallout of parallel societies that refuse to assimilate. this isn’t random crime; it’s the visible symptom of mass immigration without integration, open borders, and an establishment that prioritises imported “communities” over the safety and cohesion of the English people who built these isles. we’re importing not just people, but incompatible values, lower social trust, and rising disorder. Native Londoners – the actual English – foot the bill in stolen property, fearful daughters, and streets that feel increasingly foreign. The shopkeeper’s pursuit is a microcosm of what’s coming: when the state abdicates, the locals will act. put England first. Our people, our streets, our country. No more surrender.

Goodmayes, east London. In the latest grim dispatch from the front lines of demographic replacement, footage captures the raw chaos unfolding on the High Road after a “man” burst out of a mobile phone shop brandishing a freshly snatched Samsung device. The suspect another product of the “transformed & enriched streets of England, had been aggressively approaching multiple people in the vicinity, including a visibly distressed woman, before or during the theft. Bystanders, overwhelmingly foreign faces in the blurred melee scattered in panic as the thief tried to make his getaway before being apprehended by the police. The shop’s “bossman” taking matters into his own hands, gave chase and collared the culprit further down the road. Just a proprietor defending his patch in a postcode where phone snatchings and street predation have become the new normal. Even the thief’s choice of a Samsung drew mockery London’s imported phone bandits apparently hold out for iPhones these days, tossing lesser models like yesterday’s curry. Goodmayes and neighbouring Ilford, once solid English working-class territory, have been colonised by waves of South Asian, African, and other non-native inflows. What the BBC euphemistically calls “diversity” looks, on the ground, like eroded trust, women unable to walk without harassment, and natives dodging the fallout of parallel societies that refuse to assimilate. this isn’t random crime; it’s the visible symptom of mass immigration without integration, open borders, and an establishment that prioritises imported “communities” over the safety and cohesion of the English people who built these isles. we’re importing not just people, but incompatible values, lower social trust, and rising disorder. Native Londoners – the actual English – foot the bill in stolen property, fearful daughters, and streets that feel increasingly foreign. The shopkeeper’s pursuit is a microcosm of what’s coming: when the state abdicates, the locals will act. put England first. Our people, our streets, our country. No more surrender.

Lewis.B.Rendell Official

36,670 просмотров • 2 месяцев назад

We're approaching the dawn of medical superintelligence - the moment when affordable, world-class medical knowledge and support is at your fingertips whenever you need it. I think people are still underestimating how profound this transformation is going to be. Today we're announcing Copilot Health, enabling users to connect all their EHR records and wearable data in a secure, private health space that Copilot can analyze and reason about to provide personalized insights and proactive nudges. You choose what information to connect - from hospital lab results to your fitness tracker - and Copilot Health applies medical intelligence to surface easy to understand, personalized insights that you can actually act on. It's a dedicated space to bring your personal health data together in a single profile, including: - Activity, sleep, and vital trends from 50+ wearable devices, including Apple Health, Oura, Fitbit and many more - Health records from 50,000+ U.S. hospital and health systems, including visit summaries, medications, and test results - Comprehensive lab test results from Function Copilot Health enables people to arrive at their appointment with the right questions and the right context to make the time they have with doctors really count. Your data is always your data, and you are always in full control. Your data won't be used to train our AI models, and you can disconnect sources at any time. Our Copilot Health responses are also grounded in information from credible health organizations like Harvard Health, as well as real-time US provider directories to find the right real-world care. Copilot Health is launching first in the US to adults over 18, but we ultimately want to make this service available to the billions of people around the world who struggle to access reliable medical advice. Please give it a go and sign up to join the early Copilot Health community and help shape what comes next. More on the MAI blog:

Mustafa Suleyman

166,870 просмотров • 4 месяцев назад