Loading video...

Video Failed to Load

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

Bill Gates says scaling AI systems will work for two more iterations and after that the next big frontier is meta-cognition where AI can reason about its tasks

Tsarathustra

82,693 subscribers

160,259 views • 2 years ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

10 Comments

Tsarathustra2 years ago

Source:

David Matta2 years ago

Very interesting! The next frontier in AI 'thinking' is 'thinking about thinking.' However, metacognition would be a useless activity unless supported by cognitive and reflective frameworks with some strategic intent.

Bunagaya2 years ago

sounds to me like gates thinks after two more iterations of scaling, we will start seeing diminishing returns

Michael Chen2 years ago

You can doubt a lot of what @BillGates does. But he is quite good of identifying frontier and trends when it comes to digits 😉 MultiModal MuiltiAgent LLM is the frontier or as he says it “meta-cognition”.

Britton Winterrose2 years ago

If this is what he’s saying, I’m betting it’s what he is hearing.

web3.com2 years ago

Ohad Asor CTO of #Taunet has discovered a solution & designed an Ai that can reason about its tasks. The native token on the network is $AGRS. Please check it out

Elie Messo2 years ago

What's an iteration in this case?

Tsarathustra2 years ago

a "crank"

Peter Voss2 years ago

Two iterations aren't likely to yield even 10-fold increase in practical value. We should be working on the right approach now...

Shawn2 years ago

Continuous learning, persistent memory, ability to make many predictions, ability to build and update mental maps on the fly and planning. It sounds like a lot but we will get there pretty quick. Many good papers have come out lately.

Related Videos

Bill Gates says AI is the biggest productivity advance in our lifetime. “The world will be richer, work less and have more."

Bill Gates says AI is the biggest productivity advance in our lifetime. “The world will be richer, work less and have more."

unusual_whales

929,453 views • 2 years ago

Meet the Copilot Cowork Builders. They’re building new ways to work—where you can hand off complex tasks to AI and stay focused on the parts that need your attention. Learn more about Copilot Cowork, now available through the Frontier program:

Meet the Copilot Cowork Builders. They’re building new ways to work—where you can hand off complex tasks to AI and stay focused on the parts that need your attention. Learn more about Copilot Cowork, now available through the Frontier program:

Microsoft 365

14,323 views • 2 months ago

AI demand is exploding, but scaling it isn’t just about more compute. At NVIDIA GTC, Samsung Semiconductor US President Paul Cho shares where the real innovation is happening. Samsung’s HBM4 shows that the next phase of AI performance will be driven by memory, integration, and system design. And the companies solving those layers will define what AI can actually do.

AI demand is exploding, but scaling it isn’t just about more compute. At NVIDIA GTC, Samsung Semiconductor US President Paul Cho shares where the real innovation is happening. Samsung’s HBM4 shows that the next phase of AI performance will be driven by memory, integration, and system design. And the companies solving those layers will define what AI can actually do.

Six Five Media

1,208,573 views • 3 months ago

Richard Sutton says that the current dominance of LLMs is a "momentary fixation" the real breakthroughs will come from scaling computation, not building AI systems based on how humans think they work. LLMs will not be the leading edge of AI for more than another decade, perhaps even half a decade.

Richard Sutton says that the current dominance of LLMs is a "momentary fixation" the real breakthroughs will come from scaling computation, not building AI systems based on how humans think they work. LLMs will not be the leading edge of AI for more than another decade, perhaps even half a decade.

Haider.

137,573 views • 1 year ago

“Is there a chance that AI will be the death of democracy?" "It’s more likely that AI saves democracy.” @TylerCowen says AI is good at "spotting misinformation" and can improve the way we work. @HugoRifkind | #TimesRadio

“Is there a chance that AI will be the death of democracy?" "It’s more likely that AI saves democracy.” @TylerCowen says AI is good at "spotting misinformation" and can improve the way we work. @HugoRifkind | #TimesRadio

Times Radio

13,152 views • 1 year ago

Ray Kurzweil says AI models will exceed human verbal ability in a year or two and the next step is large event models that enable robots to perform physical tasks

Ray Kurzweil says AI models will exceed human verbal ability in a year or two and the next step is large event models that enable robots to perform physical tasks

Tsarathustra

71,501 views • 2 years ago

Holy: OpenAI says it’s reallocating compute and talent toward its next generation of models, echoing the internal shift that preceded GPT-3. The focus now is on “automated researchers” and agent-based systems that can execute complex tasks end-to-end. It’s a signal that the lab sees these systems as the next major step beyond current AI tools. Thats why they sacrificed Sora. For automated reserachers.

Holy: OpenAI says it’s reallocating compute and talent toward its next generation of models, echoing the internal shift that preceded GPT-3. The focus now is on “automated researchers” and agent-based systems that can execute complex tasks end-to-end. It’s a signal that the lab sees these systems as the next major step beyond current AI tools. Thats why they sacrificed Sora. For automated reserachers.

Chubby♨️

31,819 views • 3 months ago

Yann LeCun says the real world is far more complex than the world of language LLMs can accumulate knowledge, but they fail with high-dimensional, continuous, noisy sensory data "the next revolution is physical AI" Systems that can truly plan, reason, and understand the physical environment

Yann LeCun says the real world is far more complex than the world of language LLMs can accumulate knowledge, but they fail with high-dimensional, continuous, noisy sensory data "the next revolution is physical AI" Systems that can truly plan, reason, and understand the physical environment

Haider.

66,616 views • 5 months ago

Bill Gates announces at the World Economic Forum that his artificial intelligence is being rolled out into the healthcare system. He reveals a new partnership with OpenAI aimed at integrating AI directly into health systems worldwide. Gates says he wants AI healthcare linked across systems, likely combining digital health records, biometric ID, digital payments, and mass data sharing.

Bill Gates announces at the World Economic Forum that his artificial intelligence is being rolled out into the healthcare system. He reveals a new partnership with OpenAI aimed at integrating AI directly into health systems worldwide. Gates says he wants AI healthcare linked across systems, likely combining digital health records, biometric ID, digital payments, and mass data sharing.

Shadow of Ezra

558,611 views • 6 months ago

Facebook $META CEO Mark Zuckerberg just announced the release of its newest version of Meta AI "We believe Meta AI is now the most intelligent AI assistant that you can freely use"

Facebook $META CEO Mark Zuckerberg just announced the release of its newest version of Meta AI "We believe Meta AI is now the most intelligent AI assistant that you can freely use"

Evan

519,009 views • 2 years ago

Yann leCun — Current AI systems lack real-world understanding, reasoning, and memory. However, within the next decade or two, AI will surpass human intelligence, gaining common sense, and safety guardrails to stay under control while completing tasks.

Yann leCun — Current AI systems lack real-world understanding, reasoning, and memory. However, within the next decade or two, AI will surpass human intelligence, gaining common sense, and safety guardrails to stay under control while completing tasks.

Haider.

81,286 views • 1 year ago

Jensen Huang says technology has now reached a positive feedback loop where AI is designing new AI and is now advancing at the pace of "Moore's Law squared", meaning that the progress we will see in the next year or two will be "spectacular and surprising"

Jensen Huang says technology has now reached a positive feedback loop where AI is designing new AI and is now advancing at the pace of "Moore's Law squared", meaning that the progress we will see in the next year or two will be "spectacular and surprising"

Tsarathustra

405,902 views • 1 year ago

SALESFORCE: AI NOW DOES 30–50% OF THE WORK - AND NO, WE’RE NOT HIRING MORE CODERS CEO Marc Benioff says today’s CEOs will be the last to manage fully human workforces, as AI takes over more routine tasks: "I think that all of us have to get our head around this idea that AI can do things that before we were doing, and we can move on to do higher value work. AI is not 100% accurate. In our own work, we're doing hundreds of thousands of conversations with our customers, and we have about 93% accuracy." Source: vitrupo

SALESFORCE: AI NOW DOES 30–50% OF THE WORK - AND NO, WE’RE NOT HIRING MORE CODERS CEO Marc Benioff says today’s CEOs will be the last to manage fully human workforces, as AI takes over more routine tasks: "I think that all of us have to get our head around this idea that AI can do things that before we were doing, and we can move on to do higher value work. AI is not 100% accurate. In our own work, we're doing hundreds of thousands of conversations with our customers, and we have about 93% accuracy." Source: vitrupo

Mario Nawfal

221,479 views • 1 year ago

Anthropic co-founder Jack Clark says the AI frontier is about to cross a much stranger line than coding assistants. "My prediction is by the end of 2028, it's more likely than not that we have an AI system where you would be able to say to it, 'Make a better version of yourself.' And it just goes off and does that completely autonomously." The reason: models have moved from basic coding into "the production of science itself" and the "unglamorous but fundamental engineering work" behind large-scale R&D. "An AI system that can build itself [has] big implications, and we'd like to start thinking about it now." That is the actual agent threshold: not demos, not chat UX, but AI becoming part of the loop that improves AI.

Anthropic co-founder Jack Clark says the AI frontier is about to cross a much stranger line than coding assistants. "My prediction is by the end of 2028, it's more likely than not that we have an AI system where you would be able to say to it, 'Make a better version of yourself.' And it just goes off and does that completely autonomously." The reason: models have moved from basic coding into "the production of science itself" and the "unglamorous but fundamental engineering work" behind large-scale R&D. "An AI system that can build itself [has] big implications, and we'd like to start thinking about it now." That is the actual agent threshold: not demos, not chat UX, but AI becoming part of the loop that improves AI.

Karl Mehta

208,701 views • 23 days ago

Watch Devin Fusion work: The sidekick model handles fetching information and skimming sources. The frontier model makes a plan for implementation, then delegates well-scoped sub-tasks to the sidekick with its own cached context.

Watch Devin Fusion work: The sidekick model handles fetching information and skimming sources. The frontier model makes a plan for implementation, then delegates well-scoped sub-tasks to the sidekick with its own cached context.

Cognition

23,943 views • 20 days ago

Greg Brockman says the long-term future is clear: AI will run on its own computer, manage fleets of agents, and solve tasks in parallel. Brockman imagines a workflow where you review your agent’s work in the morning, guiding it like a partner, but models are not yet smart enough for that. He explains that today’s reality is AI helping directly in the terminal or editor, making coding workflows familiar but more efficient.

Greg Brockman says the long-term future is clear: AI will run on its own computer, manage fleets of agents, and solve tasks in parallel. Brockman imagines a workflow where you review your agent’s work in the morning, guiding it like a partner, but models are not yet smart enough for that. He explains that today’s reality is AI helping directly in the terminal or editor, making coding workflows familiar but more efficient.

Wes Roth

39,230 views • 10 months ago