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.

Chris Manning says Yann LeCun sees language as a low bandwidth communication channel compared to vision. But the gap between a chimp and a human wasn’t produced by superior eyes. What took off for humans was language. Not just for communication, but as a cognitive tool.

vitrupo

37,524 subscribers

118,405 views • 4 months ago •via X (Twitter)

Education Science & Technology

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Asalaamu Alaykum, Here's a video I filmed a few months ago on the importance of every Muslim committing to learning the Arabic language, not just as a religious language to recite Quran but as a means of communication with the wider Ummah. I hope you benefit from it insha'Allah.

Asalaamu Alaykum, Here's a video I filmed a few months ago on the importance of every Muslim committing to learning the Arabic language, not just as a religious language to recite Quran but as a means of communication with the wider Ummah. I hope you benefit from it insha'Allah.

Roshan M Salih

45,993 views • 4 months ago

Yann LeCun explains that large language models are trained on about 30 trillion words, representing nearly all public internet text. He says it would take a human over 500,000 years to read that much. But a 4-year-old child sees just as much visual data in their first few years of life. This shows how much richer and more complex real-world experience is compared to reading text. Training on the web is huge but it still doesn’t match what a child learns just by living.

Yann LeCun explains that large language models are trained on about 30 trillion words, representing nearly all public internet text. He says it would take a human over 500,000 years to read that much. But a 4-year-old child sees just as much visual data in their first few years of life. This shows how much richer and more complex real-world experience is compared to reading text. Training on the web is huge but it still doesn’t match what a child learns just by living.

Wes Roth

294,535 views • 8 months ago

Yann LeCun says we're fooled by LLMs because they manipulate language well, and we associate that with intelligence But language fluency doesn't mean underlying intelligence Every generation since the 1950s claimed its technique was the ticket to human-level AI All were wrong. "this generation with LLMs is also wrong"

Yann LeCun says we're fooled by LLMs because they manipulate language well, and we associate that with intelligence But language fluency doesn't mean underlying intelligence Every generation since the 1950s claimed its technique was the ticket to human-level AI All were wrong. "this generation with LLMs is also wrong"

Haider.

625,760 views • 8 months ago

.OmniBridge makes real-time communication possible between American Sign Language users and English speakers. Their vision is to help create a more inclusive world for both Deaf and hearing communities. 🗣💻✌🏽

.OmniBridge makes real-time communication possible between American Sign Language users and English speakers. Their vision is to help create a more inclusive world for both Deaf and hearing communities. 🗣💻✌🏽

Intel

21,070 views • 3 years ago

📁 Yann LeCun says language models do extract meaning, but only at a superficial level. Unlike humans, their intelligence is not grounded in physical reality or common sense. They answer many questions well, but break down when faced with new situations because they do not truly understand the world they describe.

📁 Yann LeCun says language models do extract meaning, but only at a superficial level. Unlike humans, their intelligence is not grounded in physical reality or common sense. They answer many questions well, but break down when faced with new situations because they do not truly understand the world they describe.

Jon Hernandez

49,318 views • 8 months ago

Elon Musk just put an expiration date on human language. Not because talking is slow. Because it’s lossy. Musk: “Our brain spends a lot of effort compressing a complex concept into words.” Language isn’t communication. It’s failed compression. You have a complete thought. You crush it into words. The other person catches fragments and rebuilds a guess. The rest never arrives. Every person who has ever loved you loved a compressed version of you. Not for lack of trying. The channel wouldn’t carry the rest. We don’t communicate. We approximate and hope the gap doesn’t matter. Musk: “You would be able to communicate very quickly and with far more precision.” Neuralink doesn’t improve communication. It replaces it. No compression. No loss. Thought moving at the speed it happens. Not describing the painting. Transmitting the seeing of it. Musk: “You wouldn’t need to talk.” Five to ten years until speech becomes optional. Talking survives for tone, for intimacy, for the pleasure of it. For information, it’s obsolete. A lifetime of memory in a second. Full schematics, not summaries. The entire structure of a thought, intact. Not better communication. Telepathy with an engineering spec. Musk: “Ideally, we are a symbiosis with artificial intelligence.” Symbiosis means the humans who don’t merge at high bandwidth don’t just fall behind. They stop being legible to the intelligence that matters. We’re already cyborgs running terrible hardware. Your phone extends your mind. Your thumbs are the bandwidth. Terabytes of thought throttled down to a few words a second. Neuralink doesn’t optimize that constraint. It detonates it. Five to ten years. Not science fiction. A deployment window. From language as default to neural link as standard. From crushing thoughts into words to sending them whole. From humans using AI to humans running at its speed. Every mind that has ever existed ran alone. That was never the human condition. It was a bandwidth limit. Solitude was never philosophy. It was hardware. Every thought you’ve ever had arrived somewhere smaller than it started. That’s not a flaw in you. It’s the width of the channel every human has ever been issued. The channel is the thing that changes. Not soon. But not in some other lifetime either, in this one, on a clock Musk just put a number on.

Elon Musk just put an expiration date on human language. Not because talking is slow. Because it’s lossy. Musk: “Our brain spends a lot of effort compressing a complex concept into words.” Language isn’t communication. It’s failed compression. You have a complete thought. You crush it into words. The other person catches fragments and rebuilds a guess. The rest never arrives. Every person who has ever loved you loved a compressed version of you. Not for lack of trying. The channel wouldn’t carry the rest. We don’t communicate. We approximate and hope the gap doesn’t matter. Musk: “You would be able to communicate very quickly and with far more precision.” Neuralink doesn’t improve communication. It replaces it. No compression. No loss. Thought moving at the speed it happens. Not describing the painting. Transmitting the seeing of it. Musk: “You wouldn’t need to talk.” Five to ten years until speech becomes optional. Talking survives for tone, for intimacy, for the pleasure of it. For information, it’s obsolete. A lifetime of memory in a second. Full schematics, not summaries. The entire structure of a thought, intact. Not better communication. Telepathy with an engineering spec. Musk: “Ideally, we are a symbiosis with artificial intelligence.” Symbiosis means the humans who don’t merge at high bandwidth don’t just fall behind. They stop being legible to the intelligence that matters. We’re already cyborgs running terrible hardware. Your phone extends your mind. Your thumbs are the bandwidth. Terabytes of thought throttled down to a few words a second. Neuralink doesn’t optimize that constraint. It detonates it. Five to ten years. Not science fiction. A deployment window. From language as default to neural link as standard. From crushing thoughts into words to sending them whole. From humans using AI to humans running at its speed. Every mind that has ever existed ran alone. That was never the human condition. It was a bandwidth limit. Solitude was never philosophy. It was hardware. Every thought you’ve ever had arrived somewhere smaller than it started. That’s not a flaw in you. It’s the width of the channel every human has ever been issued. The channel is the thing that changes. Not soon. But not in some other lifetime either, in this one, on a clock Musk just put a number on.

Dustin

137,425 views • 28 days ago

Have vision-language models achieved human-level compositional reasoning? Our research suggests: not quite yet. We’re excited to present CREPE – a large-scale Compositional REPresentation Evaluation benchmark for vision-language models – as a 🌟highlight🌟at #CVPR2023. 🧵1/7

Have vision-language models achieved human-level compositional reasoning? Our research suggests: not quite yet. We’re excited to present CREPE – a large-scale Compositional REPresentation Evaluation benchmark for vision-language models – as a 🌟highlight🌟at #CVPR2023. 🧵1/7

Zixian Ma@CVPR

21,366 views • 3 years ago

Dwarkesh Patel says we haven't yet achieved AGI for one simple reason: the models can't learn on the job. They can rewrite a transcript or suggest a clip to tweet out, but they won't remember what you liked or why it worked. Without learning loops, they can't evolve from tool to employee. “Why can't we just delegate all language work?”

Dwarkesh Patel says we haven't yet achieved AGI for one simple reason: the models can't learn on the job. They can rewrite a transcript or suggest a clip to tweet out, but they won't remember what you liked or why it worked. Without learning loops, they can't evolve from tool to employee. “Why can't we just delegate all language work?”

vitrupo

85,997 views • 1 year ago

🇺🇸 What if Neuralink’s real goal isn’t healing the broken... but giving ordinary humans superpowers? Stephen Hawking could only get out a few words a minute, a brutal example of how wide the gap can get between what the mind knows and what the body can express. Elon sees that bottleneck as something technology can finally break, not just restoring lost abilities, but expanding what a human being is actually capable of doing. Neuralink / Writer: Val

🇺🇸 What if Neuralink’s real goal isn’t healing the broken... but giving ordinary humans superpowers? Stephen Hawking could only get out a few words a minute, a brutal example of how wide the gap can get between what the mind knows and what the body can express. Elon sees that bottleneck as something technology can finally break, not just restoring lost abilities, but expanding what a human being is actually capable of doing. Neuralink / Writer: Val

Mario Nawfal

86,781 views • 9 days ago

Ever wondered what robots 🤖 could achieve if they could not just see – but also feel and hear? Introducing FuSe: a recipe for finetuning large vision-language-action (VLA) models with heterogeneous sensory data, such as vision, touch, sound, and more. Details in the thread 👇

Ever wondered what robots 🤖 could achieve if they could not just see – but also feel and hear? Introducing FuSe: a recipe for finetuning large vision-language-action (VLA) models with heterogeneous sensory data, such as vision, touch, sound, and more. Details in the thread 👇

Carlo Sferrazza

46,111 views • 1 year ago

📁 Yann LeCun says that AGI does not exist and is a poorly defined concept built around an idealized view of human intelligence. He argues that humans are highly specialized, good at navigating the physical and social world, but weak at many other tasks. We seem general only because we can imagine the problems we are good at. For LeCun, AGI mistakes human limits for universal laws of intelligence.

📁 Yann LeCun says that AGI does not exist and is a poorly defined concept built around an idealized view of human intelligence. He argues that humans are highly specialized, good at navigating the physical and social world, but weak at many other tasks. We seem general only because we can imagine the problems we are good at. For LeCun, AGI mistakes human limits for universal laws of intelligence.

Jon Hernandez

28,104 views • 8 months ago

Yann LeCun says we're never gonna get to human-level intelligence by just training on text AI must learn from high-bandwidth sensory data like video to build true world models Current models look PhD-smart but mostly regurgitate, with no real understanding "even a cat understands the physical world better"

Yann LeCun says we're never gonna get to human-level intelligence by just training on text AI must learn from high-bandwidth sensory data like video to build true world models Current models look PhD-smart but mostly regurgitate, with no real understanding "even a cat understands the physical world better"

Haider.

138,265 views • 10 months ago

Kamala Harris’ low IQ represents the same thing as Joe Biden’s cognitive decline: it’s not a bug, but a *feature* for the people who control them.

Kamala Harris’ low IQ represents the same thing as Joe Biden’s cognitive decline: it’s not a bug, but a feature for the people who control them.

Vivek Ramaswamy

1,392,450 views • 2 years ago

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.

AK

125,353 views • 3 years ago

Maduro “was, of course, a dictator and a thug and responsible for egregious crimes against his own people,” says international human rights lawyer Jared Genser. “But at the same time… what took place was illegal as a matter of international law.”

Maduro “was, of course, a dictator and a thug and responsible for egregious crimes against his own people,” says international human rights lawyer Jared Genser. “But at the same time… what took place was illegal as a matter of international law.”

Christiane Amanpour

36,386 views • 7 months ago

Demis Hassabis says AGI should become a tool before we try to make it conscious. First use it to read the language of the universe, then decide what it means to build something with agency or a mind.

Demis Hassabis says AGI should become a tool before we try to make it conscious. First use it to read the language of the universe, then decide what it means to build something with agency or a mind.

vitrupo

30,462 views • 3 months ago

WATCH: NBC's Tom Costello says we should "collectively" be proud of the Artemis II launch, "not as Americans, not as North Americans or as- but just as humans" TOM COSTELLO: I think it's important and relevant to take a moment and say wow, we should be collectively, not as Americans, not as North Americans or as- but just as humans, proud of the achievement here- that humans have been able to do this. And we're about to go even further than ever before. It's- it's exactly what they had hoped for, setting the stage to eventually put humans on the moon.

WATCH: NBC's Tom Costello says we should "collectively" be proud of the Artemis II launch, "not as Americans, not as North Americans or as- but just as humans" TOM COSTELLO: I think it's important and relevant to take a moment and say wow, we should be collectively, not as Americans, not as North Americans or as- but just as humans, proud of the achievement here- that humans have been able to do this. And we're about to go even further than ever before. It's- it's exactly what they had hoped for, setting the stage to eventually put humans on the moon.

Jorge Bonilla

49,010 views • 4 months ago

A fashion show featuring both robots and humans was held in Seoul On the runway, human models and robots appeared side by side, dressed in identical outfits. The show was designed as a demonstration of a future in which humans and artificial intelligence coexist and interact. Organizers describe the project as a step toward “physical AI,” where robots are not a replacement for humans, but partners in creative and social processes.

A fashion show featuring both robots and humans was held in Seoul On the runway, human models and robots appeared side by side, dressed in identical outfits. The show was designed as a demonstration of a future in which humans and artificial intelligence coexist and interact. Organizers describe the project as a step toward “physical AI,” where robots are not a replacement for humans, but partners in creative and social processes.

NEXTA

22,955 views • 3 months ago

Max Tegmark says the way through the AI black box may be to deploy its outputs instead. A cat can learn, but it cannot export what it learned. A human scientist can. As intelligence rises, more knowledge may become externalized into code and proofs. Deploy what the mind can verify, not the mind itself.

Max Tegmark says the way through the AI black box may be to deploy its outputs instead. A cat can learn, but it cannot export what it learned. A human scientist can. As intelligence rises, more knowledge may become externalized into code and proofs. Deploy what the mind can verify, not the mind itself.

vitrupo

49,717 views • 4 months ago