Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Why Richard Sutton thinks LLMs go against the bitter lesson:

dwarkesh_sp's profile picture

Dwarkesh Patel

202,473 subscribers

67,497 Aufrufe • vor 1 Jahr •via X (Twitter)

Wissenschaft & TechnologieBildung
LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Lots of people think it's obvious that humans learn by imitation, starting with babies imitating their parents. Richard Sutton thinks this is completely wrong. Sutton basically thinks direct imitation plays almost no role in animal learning, and at most a small role in human intelligence. When one animal ends up doing the same thing as another, he thinks this is the result of RL-style trial-and-error, not direct imitation. And if imitation isn't really a kind of learning, then training LLMs to imitate humans isn't really Bitter Lesson–pilled.
1:41

Lots of people think it's obvious that humans learn by imitation, starting with babies imitating their parents. Richard Sutton thinks this is completely wrong. Sutton basically thinks direct imitation plays almost no role in animal learning, and at most a small role in human intelligence. When one animal ends up doing the same thing as another, he thinks this is the result of RL-style trial-and-error, not direct imitation. And if imitation isn't really a kind of learning, then training LLMs to imitate humans isn't really Bitter Lesson–pilled.

Dwarkesh Patel

55,137 Aufrufe • vor 3 Monaten

Why Sutton Stracke thinks her ex wanted the annulment #WWHL #RHOBH
1:21

Why Sutton Stracke thinks her ex wanted the annulment #WWHL #RHOBH

Watch What Happens Live!

21,214 Aufrufe • vor 8 Monaten

.Richard Sutton, father of reinforcement learning, doesn’t think LLMs are bitter-lesson-pilled. My steel man of Richard’s position: we need some new architecture to enable continual (on-the-job) learning. And if we have continual learning, we don't need a special training phase - the agent just learns on-the-fly - like all humans, and indeed, like all animals. This new paradigm will render our current approach with LLMs obsolete. I did my best to represent the view that LLMs will function as the foundation on which this experiential learning can happen. Some sparks flew. 0:00:00 – Are LLMs a dead-end? 0:13:51 – Do humans do imitation learning? 0:23:57 – The Era of Experience 0:34:25 – Current architectures generalize poorly out of distribution 0:42:17 – Surprises in the AI field 0:47:28 – Will The Bitter Lesson still apply after AGI? 0:54:35 – Succession to AI
1:07:08

.Richard Sutton, father of reinforcement learning, doesn’t think LLMs are bitter-lesson-pilled. My steel man of Richard’s position: we need some new architecture to enable continual (on-the-job) learning. And if we have continual learning, we don't need a special training phase - the agent just learns on-the-fly - like all humans, and indeed, like all animals. This new paradigm will render our current approach with LLMs obsolete. I did my best to represent the view that LLMs will function as the foundation on which this experiential learning can happen. Some sparks flew. 0:00:00 – Are LLMs a dead-end? 0:13:51 – Do humans do imitation learning? 0:23:57 – The Era of Experience 0:34:25 – Current architectures generalize poorly out of distribution 0:42:17 – Surprises in the AI field 0:47:28 – Will The Bitter Lesson still apply after AGI? 0:54:35 – Succession to AI

Dwarkesh Patel

3,090,100 Aufrufe • vor 1 Jahr

Bozoma Saint John thinks this is why Sutton Stracke gets invited to The Met Ball every year #RHOBH #WWHL
0:43

Bozoma Saint John thinks this is why Sutton Stracke gets invited to The Met Ball every year #RHOBH #WWHL

Watch What Happens Live!

83,733 Aufrufe • vor 1 Jahr

Sutton talks why she let Avi go: Kyle, Sutton, & Erika discuss Garcelle. #RHOBH Source: Bravo
1:50

Sutton talks why she let Avi go: Kyle, Sutton, & Erika discuss Garcelle. #RHOBH Source: Bravo

jay

35,514 Aufrufe • vor 10 Monaten

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.
1:16

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 Aufrufe • vor 1 Jahr

In 1971 the US government funded a speech AI that used a hand-crafted knowledge graph of 14,000 nodes. It worked. Then hidden Markov models replaced it entirely by learning from data - and outperformed it by orders of magnitude. Richard Sutton called this the bitter lesson. In 2025 Sutton said LLMs are repeating the same mistake. Training on human-generated text is just a more sophisticated version of building human knowledge into the system. AlphaGo Zero - trained with zero human games, only reinforcement learning - discovered strategies experts called "from an alternate dimension." 22 minutes. Bookmark & watch today. The most important question in AI nobody is asking loudly enough.
23:10

In 1971 the US government funded a speech AI that used a hand-crafted knowledge graph of 14,000 nodes. It worked. Then hidden Markov models replaced it entirely by learning from data - and outperformed it by orders of magnitude. Richard Sutton called this the bitter lesson. In 2025 Sutton said LLMs are repeating the same mistake. Training on human-generated text is just a more sophisticated version of building human knowledge into the system. AlphaGo Zero - trained with zero human games, only reinforcement learning - discovered strategies experts called "from an alternate dimension." 22 minutes. Bookmark & watch today. The most important question in AI nobody is asking loudly enough.

Eron

56,969 Aufrufe • vor 1 Monat

"Who are we to decide how the universe is going to evolve?" Richard Sutton argues against controlling AI's long-term future. If we intend to birth "descendants of humanity" reflecting us, why obstruct their evolution?
2:08

"Who are we to decide how the universe is going to evolve?" Richard Sutton argues against controlling AI's long-term future. If we intend to birth "descendants of humanity" reflecting us, why obstruct their evolution?

vitrupo

28,334 Aufrufe • vor 1 Jahr

"Is it Bitter Lesson-pilled?" — the most used, and most misused, phrase in AI right now. Richard Sutton sums it up on the latest Training Data with Khurram Javed , Sonya Huang 🐥, Alfred Lin "Don't be distracted by human knowledge, as AI traditionally has been many times. Instead, focus on learning methods that will scale with computation, like search and like learning."
1:02

"Is it Bitter Lesson-pilled?" — the most used, and most misused, phrase in AI right now. Richard Sutton sums it up on the latest Training Data with Khurram Javed , Sonya Huang 🐥, Alfred Lin "Don't be distracted by human knowledge, as AI traditionally has been many times. Instead, focus on learning methods that will scale with computation, like search and like learning."

Sequoia Capital

28,612 Aufrufe • vor 1 Monat

Genie 2 vs Genie 3. Just 7 months between them. The bitter lesson continues to be bitter.
0:41

Genie 2 vs Genie 3. Just 7 months between them. The bitter lesson continues to be bitter.

Bilawal Sidhu

380,551 Aufrufe • vor 1 Jahr

She taught her husband’s family a bitter lesson after the divorce.
5:15

She taught her husband’s family a bitter lesson after the divorce.

Chidera0402

60,101 Aufrufe • vor 29 Tagen

Seeing Richard Inman go on a rant about why he thinks Restore Britain is racist before accidentally referring to Shabana Mahmood as 'Banana Mahmood' was absolutely hilarious 🤣
0:29

Seeing Richard Inman go on a rant about why he thinks Restore Britain is racist before accidentally referring to Shabana Mahmood as 'Banana Mahmood' was absolutely hilarious 🤣

Basil the Great

17,310 Aufrufe • vor 7 Monaten

⚔️ | "It's Wilder against the world! We've seen it all before." Conor McGilligan thinks Sheff Utd might be starting to put a serious run together. Simon Lillibury Gabriel Sutton
0:44

⚔️ | "It's Wilder against the world! We've seen it all before." Conor McGilligan thinks Sheff Utd might be starting to put a serious run together. Simon Lillibury Gabriel Sutton

Betfred

39,730 Aufrufe • vor 7 Monaten

He regretted Mocking Tyson. All bad boys were taught a bitter lesson by Tyson?? Those who had won against Tyson had felt the heat from Tyson’s punches. #Boxing
2:44

He regretted Mocking Tyson. All bad boys were taught a bitter lesson by Tyson?? Those who had won against Tyson had felt the heat from Tyson’s punches. #Boxing

Mr Vincent🇬🇧🇺🇸

169,038 Aufrufe • vor 1 Monat

Chris Sutton thinks Tottenham will be relegated from the Premier League this season ⚪⬇️
0:48

Chris Sutton thinks Tottenham will be relegated from the Premier League this season ⚪⬇️

Match of the Day

133,296 Aufrufe • vor 5 Monaten

Richard Tice calls on Labour MPs to go against the whip and investigate Keir Starmer. 🚨
5:31

Richard Tice calls on Labour MPs to go against the whip and investigate Keir Starmer. 🚨

Reform UK

114,574 Aufrufe • vor 5 Monaten

Andy Garcia and Richard Gere had such a bitter feud while filming Internal Affairs (1990) that Garcia refused to go to the wrap party. Both later downplayed it. Richard Gere’s take: “Whatever happened on set ultimately resulted in brilliant cinema.”
1:38

Andy Garcia and Richard Gere had such a bitter feud while filming Internal Affairs (1990) that Garcia refused to go to the wrap party. Both later downplayed it. Richard Gere’s take: “Whatever happened on set ultimately resulted in brilliant cinema.”

The Sting

147,725 Aufrufe • vor 2 Monaten

Today we release my favorite episode of Training Data yet: the great Rich Sutton. Richard Sutton wrote the textbook, wrote The Bitter Lesson (and many other on-point essays like "Self-Verification, The Key to AI"), and trained a mafia of talented students who went on to change the AI landscape forever including David Silver, inventor of built AlphaGo. Khurram Javed was Rich's PhD student at Alberta and wrote The Big World Hypothesis. They just left academia to start Oak Lab Their core argument: (1) The Bitter Lesson: the world is massively more complex than any model of it, so anything trained on human-curated data has a ceiling (2) Continual Learning: intelligence is continual by definition, and today's models stop learning the moment they ship. The conversation covers: — what The Bitter Lesson actually says, and what people get wrong — why synthetic data is "just a big mistake," and the Big World Hypothesis behind it — how LLMs are both a positive and a negative example of his own essay — why no animal learns by supervised learning, and what squirrels can do that we can't — the cure for catastrophic forgetting: per-weight step sizes and continual backprop — why the biggest labs can't take a path where performance gets worse before it gets better — a trillion parameters on 20 watts, and the Moore's Law math that makes it plausible — why the endpoint isn't one mind but one design, running as many minds It was both a fun generative idea- and debate-filled conversation, and a surprisingly human one too. Rich, thank you for beating cancer and changing the trajectory of AI. 💙 00:00 Introduction 02:10 An AI winter, a cancer diagnosis, and the move to Alberta 07:07 Writing "The Bitter Lesson," and what people get wrong 09:53 Are LLMs a positive or a negative example of it? 11:03 Synthetic data is "just a big mistake," and the Big World Hypothesis 18:01 AlphaGo, human priors, and why prior knowledge and learning should be friends 22:37 "Their weights never change": do LLM assistants actually learn? 26:09 Babies, squirrels, and why no animal learns by supervised learning 32:02 Rockets, imagination, and where paradigm shifts come from 36:42 The Alberta Plan and its 12 steps 38:53 Catastrophic forgetting and the cure 43:43 Oak's biggest ambition: a self-maintaining mind 47:56 Why the big labs are stuck in a local minimum 49:13 If everything goes right: LLMs, many minds, and hiring The man who pioneered reinforcement learning thinks the rest of the field is weird, and lays it all out in today's episode. Together w/ Alfred Lin Sequoia Capital
53:43

Today we release my favorite episode of Training Data yet: the great Rich Sutton. Richard Sutton wrote the textbook, wrote The Bitter Lesson (and many other on-point essays like "Self-Verification, The Key to AI"), and trained a mafia of talented students who went on to change the AI landscape forever including David Silver, inventor of built AlphaGo. Khurram Javed was Rich's PhD student at Alberta and wrote The Big World Hypothesis. They just left academia to start Oak Lab Their core argument: (1) The Bitter Lesson: the world is massively more complex than any model of it, so anything trained on human-curated data has a ceiling (2) Continual Learning: intelligence is continual by definition, and today's models stop learning the moment they ship. The conversation covers: — what The Bitter Lesson actually says, and what people get wrong — why synthetic data is "just a big mistake," and the Big World Hypothesis behind it — how LLMs are both a positive and a negative example of his own essay — why no animal learns by supervised learning, and what squirrels can do that we can't — the cure for catastrophic forgetting: per-weight step sizes and continual backprop — why the biggest labs can't take a path where performance gets worse before it gets better — a trillion parameters on 20 watts, and the Moore's Law math that makes it plausible — why the endpoint isn't one mind but one design, running as many minds It was both a fun generative idea- and debate-filled conversation, and a surprisingly human one too. Rich, thank you for beating cancer and changing the trajectory of AI. 💙 00:00 Introduction 02:10 An AI winter, a cancer diagnosis, and the move to Alberta 07:07 Writing "The Bitter Lesson," and what people get wrong 09:53 Are LLMs a positive or a negative example of it? 11:03 Synthetic data is "just a big mistake," and the Big World Hypothesis 18:01 AlphaGo, human priors, and why prior knowledge and learning should be friends 22:37 "Their weights never change": do LLM assistants actually learn? 26:09 Babies, squirrels, and why no animal learns by supervised learning 32:02 Rockets, imagination, and where paradigm shifts come from 36:42 The Alberta Plan and its 12 steps 38:53 Catastrophic forgetting and the cure 43:43 Oak's biggest ambition: a self-maintaining mind 47:56 Why the big labs are stuck in a local minimum 49:13 If everything goes right: LLMs, many minds, and hiring The man who pioneered reinforcement learning thinks the rest of the field is weird, and lays it all out in today's episode. Together w/ Alfred Lin Sequoia Capital

Sonya Huang 🐥

119,665 Aufrufe • vor 1 Monat