Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

Fields Medalist Timothy Gowers: "when we solve maths problems ourselves, we understand exactly what we're doing. when we one-shot a maths problem with ChatGPT, we don't — we just stuck it in a machine" How can AI find such long, complex proofs? What algorithm is operating inside the LLM?... show more

Haider.

72,601 subscribers

30,179 görüntüleme • 10 gün önce •via X (Twitter)

Eğitim Bilim & Teknoloji

Anya Rossi• Live Now

Private livecam show

33 Yorum

Jeffrey Dean Hochderffer profil fotoğrafı

Jeffrey Dean Hochderffer10 gün önce

Nothing is preventing them from studying the AI results until they understand it, if they can. Progress cannot wait for them to understand. We don’t understand how 90% of things work.

Sergio Castro profil fotoğrafı

Sergio Castro10 gün önce

LLMs are great at explaining things. Hook it to Blender, ask for explanation videos.

Amy · fireply.ai profil fotoğrafı

Amy · fireply.ai10 gün önce

nobody asked the calculator how it works either

Sean McClure profil fotoğrafı

Sean McClure10 gün önce

No. The fact that mathematicians aren’t getting it tells us a lot.

Freedom_Aint_Free profil fotoğrafı

Freedom_Aint_Free10 gün önce

Wasn't Riverside who said that "one doesn't need to understand digestion in order to appreciate a good dinner!" ?

Rishabh Kumar profil fotoğrafı

Rishabh Kumar10 gün önce

then we ask the machine to explain 😂

Brian Decoded - Alpha Efficiency.™ profil fotoğrafı

Brian Decoded - Alpha Efficiency.™10 gün önce

bro is envious

Logan Gabriel profil fotoğrafı

Logan Gabriel10 gün önce

I don’t care the exact details of how warp drive works - as long as it works - nerds can take hundreds of years to refine our understanding while we go explore the stars

Xia Li profil fotoğrafı

Xia Li10 gün önce

@grok what algorithms had been deployed to generate the candidates solutions and verify via LEAN and feedback verification results and improve and generate new candidates solutions?

Simo | ai-costguard profil fotoğrafı

Simo | ai-costguard10 gün önce

Long proofs aren't a search algorithm. They're learned representations under pressure

William Zhu profil fotoğrafı

William Zhu10 gün önce

New talk from Terence Tao yesterday:

Biscuit profil fotoğrafı

Biscuit10 gün önce

at least in math you can check every step

Ludovicus D. profil fotoğrafı

Ludovicus D.9 gün önce

All these Fields medalists so-called great mathematicians are revealing themselves hopelessly clueless in front of AI. They have no idea about how&why it works. They don't know how to use it. They don't know what to do next. They look pathetic.

Gill profil fotoğrafı

Gill9 gün önce

Verification being trivial while proof discovery feels uninterpretable is classic math.

Wanderer of the Stars profil fotoğrafı

Wanderer of the Stars10 gün önce

It really won’t matter we don’t understand how it did it. The future will be AI building on top of AI without humans in the loop.

Konstantin profil fotoğrafı

Konstantin10 gün önce

When he states his assumptions, it is categorically incorrect to use the word "we." It would be correct to use the pronoun "I." Just because he doesn't understand, it doesn't mean that no one else understands!

Lemurinus profil fotoğrafı

Lemurinus9 gün önce

I am a mathematician and for many years I have been saying that these "geniuses" are the dumbest people in the world. Finally, the world can see it with its own eyes.

BaloonedWallabohe profil fotoğrafı

BaloonedWallabohe9 gün önce

These guys act like humans didn’t use fire before understanding it. Or oxygen. Or gravity.

j363j profil fotoğrafı

j363j9 gün önce

Prompt an AI reasoning model for help with understanding.

Dominik profil fotoğrafı

Dominik9 gün önce

does the machine show its steps, or only the finished proof?

Sly Sylow profil fotoğrafı

Sly Sylow10 gün önce

Tim is the goat 🐐.

TheMCDStudio profil fotoğrafı

TheMCDStudio10 gün önce

Some of the math papers contain math that is SI thinking 'outside the box'. That is the point, SI is discovering new 'alien' maths as solutions. Mathematicians would do well to try to learn it. It will help them grow in math comprehension. Posing the problems is still the job.

Chahat Sharma profil fotoğrafı

Chahat Sharma10 gün önce

does the full talk say which of the summer proofs he went through line by line himself? the clip makes it sound like even he is taking some of them on trust

Steve Scully For US President profil fotoğrafı

Steve Scully For US President10 gün önce

By making shit up and it being proclaimed in an impossible-to-review manner of a huge dump like a lawyer trying to hide evidence in bulk frivolous shit

Jatin Garg profil fotoğrafı

Jatin Garg10 gün önce

@haider1 the harder question is what happens when the machine gives you a proof that looks complete but has a subtle flaw. how do you catch it without understanding each step?

lio khi profil fotoğrafı

lio khi9 gün önce

The same when reading the proof of transcendance of Pi by Lindemann ..

Inverse Gary Marcus ⏫ profil fotoğrafı

Inverse Gary Marcus ⏫9 gün önce

Sounds like a good job for mathematicians!

B profil fotoğrafı

B10 gün önce

Software devs went through the same phase. The gains come when you realize that you should be driving the AI and not the other way around. Until then, he's right. You're just creating slop.

ChrisvG profil fotoğrafı

ChrisvG9 gün önce

People like him will only become more irrelevant as time passes and the models improve even more.

Chey Loveday profil fotoğrafı

Chey Loveday9 gün önce

2 diff Qs: - why did an LLM produce a proof - why does it work? fThe irst remains open. We pin sources/assumptions, replay certificates & check Lean 4; then use interpolation, counterexamples, premise ablation and proof-transfer maps to investigate why proofs work or fail.

夏日遐想 profil fotoğrafı

夏日遐想10 gün önce

大脑如何工作，这个问题也未解决。DNN和大脑都属于复杂结构的范畴。

Alejandro Estay profil fotoğrafı

Alejandro Estay9 gün önce

Is not an algorithm, is just picking strings of may-be-related text and refining it. It cannot understsnd the proof it writes.

CharlesGPhD profil fotoğrafı

CharlesGPhD10 gün önce

More than that, when a LLM tells us "your proof is wrong", it doesn't understand. It's not checking all the reasoning. So it's fumbling. Not quite hallucinating, but not understanding. When you prove something, you have that intricate understanding.

Benzer Videolar

What do we do when we're not doing maths? Maths.

What do we do when we're not doing maths? Maths.

Oxford Mathematics

28,380 görüntüleme • 10 ay önce

🚨GEOFFREY HINTON: WE DON'T KNOW EXACTLY HOW AI WORKS AI Pioneer Geoffrey Hinton: "We have a rough idea of what it's doing, but when it gets complicated, we don't know what's going on, similar to our understanding of the brain." 60 Minutes: "What do you mean we don't know exactly how it works? It was designed by people." Hinton: "We designed the learning algorithm, like the principle of evolution. But when it interacts with data, it creates neural networks we don't fully understand." Source: 60 Minutes, Newsy on IG

🚨GEOFFREY HINTON: WE DON'T KNOW EXACTLY HOW AI WORKS AI Pioneer Geoffrey Hinton: "We have a rough idea of what it's doing, but when it gets complicated, we don't know what's going on, similar to our understanding of the brain." 60 Minutes: "What do you mean we don't know exactly how it works? It was designed by people." Hinton: "We designed the learning algorithm, like the principle of evolution. But when it interacts with data, it creates neural networks we don't fully understand." Source: 60 Minutes, Newsy on IG

Mario Nawfal

502,988 görüntüleme • 2 yıl önce

"What is P vs NP and how does it relate to human knowledge?" Avi Wigderson (Turing Award & Abel Prize winner): "In a very basic way, all the problems we want to solve are really NP problems. A subset of it is the problems we can solve. So the question of P versus NP is whether we can solve all the problems we want to solve, whether we can know everything we want to know. This is basically about the limits of our knowledge. You can think there is no limit to what we may want to know but I claim that essentially in any human endeavor, what people are searching for have the property that when a solution is found, it is easily recognized. NP are all problems that we can honestly say we really want to solve and P are those that we can solve. Whether they are equal is just the very fact that a solution can be easily recognized would imply that it can be efficiently found. That means that we can know everything we ever want to know, it's a fundamental question about human knowledge."

"What is P vs NP and how does it relate to human knowledge?" Avi Wigderson (Turing Award & Abel Prize winner): "In a very basic way, all the problems we want to solve are really NP problems. A subset of it is the problems we can solve. So the question of P versus NP is whether we can solve all the problems we want to solve, whether we can know everything we want to know. This is basically about the limits of our knowledge. You can think there is no limit to what we may want to know but I claim that essentially in any human endeavor, what people are searching for have the property that when a solution is found, it is easily recognized. NP are all problems that we can honestly say we really want to solve and P are those that we can solve. Whether they are equal is just the very fact that a solution can be easily recognized would imply that it can be efficiently found. That means that we can know everything we ever want to know, it's a fundamental question about human knowledge."

Ryan Peterman

20,069 görüntüleme • 4 ay önce

AlphaFold's awesomeness in a nutshell. Many people still don't realize how big the success of generative AI is and only think of ChatGPT (3.5) when they hear “AI”. Generative AI is more, it is a success in many areas in many fields! I am serious when I say that the world will no longer be the same in a few years' time. When we will no longer die from previously incurable diseases. When we can solve problems that we thought were unsolvable.

AlphaFold's awesomeness in a nutshell. Many people still don't realize how big the success of generative AI is and only think of ChatGPT (3.5) when they hear “AI”. Generative AI is more, it is a success in many areas in many fields! I am serious when I say that the world will no longer be the same in a few years' time. When we will no longer die from previously incurable diseases. When we can solve problems that we thought were unsolvable.

Chubby♨️

26,611 görüntüleme • 1 yıl önce

Tedros: "[The WHO] is working... to understand how misinformation and disinformation spreads... and what we can do to counter this problem." "But we must also make sure that when when we seek the trust of others, we are ourselves trustworthy." On a scale of 0-100, how "trustworthy" do you find Tedros and the WHO? 🤔

Tedros: "[The WHO] is working... to understand how misinformation and disinformation spreads... and what we can do to counter this problem." "But we must also make sure that when when we seek the trust of others, we are ourselves trustworthy." On a scale of 0-100, how "trustworthy" do you find Tedros and the WHO? 🤔

Wide Awake Media

86,456 görüntüleme • 1 yıl önce

Working with agents is a collaboration problem. Our tools didn’t even really support working together *before* agents. What happens when we reinvent then around collaboration? If we get the primitives right, we can build workflows that match exactly how we want to work

Working with agents is a collaboration problem. Our tools didn’t even really support working together before agents. What happens when we reinvent then around collaboration? If we get the primitives right, we can build workflows that match exactly how we want to work

Amelia Wattenberger 🪷

16,277 görüntüleme • 8 ay önce

Sometimes, we need to ask ourselves a difficult question: Is religion making us less creative, less innovative, or even less driven to solve our own problems? When you look at what our ancestors achieved with the limited resources available to them, it is worth asking why we seem to be doing less today despite having access to more knowledge and technology. I’m not saying religion is the problem—but it is a question worth asking: are we using faith to inspire our progress, or are we sometimes using it as a substitute for action?

Sometimes, we need to ask ourselves a difficult question: Is religion making us less creative, less innovative, or even less driven to solve our own problems? When you look at what our ancestors achieved with the limited resources available to them, it is worth asking why we seem to be doing less today despite having access to more knowledge and technology. I’m not saying religion is the problem—but it is a question worth asking: are we using faith to inspire our progress, or are we sometimes using it as a substitute for action?

JusticeCrack

27,950 görüntüleme • 22 gün önce

So countering America with tariffs that they admit it hurts us...is a good thing because that somehow shows we defend ourselves? Jean Charest: "We (Canada) are one of the few countries in the world that's actually applied counter-tariffs in the automotive sector and elsewhere. By the way, when we do that, just to be clear, we're very aware, we hurt ourselves. We know that. We know that. it's part of what we need to do to be able to defend ourselves." Huh

So countering America with tariffs that they admit it hurts us...is a good thing because that somehow shows we defend ourselves? Jean Charest: "We (Canada) are one of the few countries in the world that's actually applied counter-tariffs in the automotive sector and elsewhere. By the way, when we do that, just to be clear, we're very aware, we hurt ourselves. We know that. We know that. it's part of what we need to do to be able to defend ourselves." Huh

Kirk Lubimov

29,815 görüntüleme • 2 ay önce

Charlie McAvoy on why the Bruins keep struggling in home playoff games: "It's just not acceptable. I don't know exactly what it is. We can talk a big game about how excited we are. ... We were awesome at home this year, but we didn't do when it mattered the most."

Charlie McAvoy on why the Bruins keep struggling in home playoff games: "It's just not acceptable. I don't know exactly what it is. We can talk a big game about how excited we are. ... We were awesome at home this year, but we didn't do when it mattered the most."

Conor Ryan

61,555 görüntüleme • 5 ay önce

Warren Buffett on how to read annual reports. (1) "We start by looking at the reports of companies that we think we can understand." (2) "Then we see, from that report, whether the management is telling us about the things that we would want to know about if we owned 100% of the company." (3) "When we find a management that does tell us about those things and that is candid ... and talks in language that we can understand, it definitely improves our feeling about investing in such a business." (4) "We want to understand the business better when we get through with the annual report than when we picked it up. That is not difficult for a management to do — if they want to do it." (Berkshire Hathaway AGM || 1998)

Warren Buffett on how to read annual reports. (1) "We start by looking at the reports of companies that we think we can understand." (2) "Then we see, from that report, whether the management is telling us about the things that we would want to know about if we owned 100% of the company." (3) "When we find a management that does tell us about those things and that is candid ... and talks in language that we can understand, it definitely improves our feeling about investing in such a business." (4) "We want to understand the business better when we get through with the annual report than when we picked it up. That is not difficult for a management to do — if they want to do it." (Berkshire Hathaway AGM || 1998)

Kevin Carpenter

12,967 görüntüleme • 16 gün önce

"The business we are in is a performance-based business. It is what it is. When we expect a service, we're going to be disappointed when we don't get that service. I get it. That's why we're paid what we are...to perform at a high level. We run to this smoke. This is the job we've chosen, so this is part of it." #Browns head coach Todd Monken says it's not difficult for him to understand where the fans are coming from with the reaction to Deshaun Watson

"The business we are in is a performance-based business. It is what it is. When we expect a service, we're going to be disappointed when we don't get that service. I get it. That's why we're paid what we are...to perform at a high level. We run to this smoke. This is the job we've chosen, so this is part of it." #Browns head coach Todd Monken says it's not difficult for him to understand where the fans are coming from with the reaction to Deshaun Watson

92.3 The Fan

17,596 görüntüleme • 2 ay önce

We asked David Hsu (founder of Retool) about new AI use cases. "The US has invested $1-1.5T into AI, but revenue is only $20-30B. That's a crappy 3% ROI." "My guess is 80% of that revenue is ChatGPT + cloud. But chat is pretty limited, is it going to grow 30x from where we are now?" "Can we figure out some LLM uses cases beyond just chat? What we think is missing is leveraging LLMs to automate labor." "What we're here to really solve is, can we actually allow LLMs to do things in your business (like approving orders or plans). The next frontier for AI is AI that actually does things." "I think we're almost there. That's what we are working on."

We asked David Hsu (founder of Retool) about new AI use cases. "The US has invested $1-1.5T into AI, but revenue is only $20-30B. That's a crappy 3% ROI." "My guess is 80% of that revenue is ChatGPT + cloud. But chat is pretty limited, is it going to grow 30x from where we are now?" "Can we figure out some LLM uses cases beyond just chat? What we think is missing is leveraging LLMs to automate labor." "What we're here to really solve is, can we actually allow LLMs to do things in your business (like approving orders or plans). The next frontier for AI is AI that actually does things." "I think we're almost there. That's what we are working on."

TBPN

86,230 görüntüleme • 1 yıl önce

"Not ever as long as [we're] in charge. What we do, and why we do it, is grounded in humanity ... no algorithm can replicate [that] ... true effort, inspiration, and perspiration"

"Not ever as long as [we're] in charge. What we do, and why we do it, is grounded in humanity ... no algorithm can replicate [that] ... true effort, inspiration, and perspiration"

Culture Crave 🍿

49,243 görüntüleme • 1 yıl önce

Man asks Ai how long it things we have to “everything actually hits” “What happens when the big one hits? What happens that day? How do we prepare?” “Welcome to survival mode”

Man asks Ai how long it things we have to “everything actually hits” “What happens when the big one hits? What happens that day? How do we prepare?” “Welcome to survival mode”

Concerned Citizen

209,606 görüntüleme • 14 gün önce

Excellent answer from President Trump to a question wondering "why...running a country in South America" is "America first": "We want to surround ourselves with good neighbors. We want to surround ourselves with stability. We want to surround ourselves with energy. We have tremendous energy in that country. It's very important that we protect it. We need that for ourselves. We need that for the world, and we want to make sure we can protect it."

Excellent answer from President Trump to a question wondering "why...running a country in South America" is "America first": "We want to surround ourselves with good neighbors. We want to surround ourselves with stability. We want to surround ourselves with energy. We have tremendous energy in that country. It's very important that we protect it. We need that for ourselves. We need that for the world, and we want to make sure we can protect it."

Curtis Houck

185,634 görüntüleme • 9 ay önce

Netanyahu: There is one challenge in particular—what I call the "Manhattan Project." We will be the first in the world to solve the problem of explosive drones. It is a global problem. But we are going to solve it.

Netanyahu: There is one challenge in particular—what I call the "Manhattan Project." We will be the first in the world to solve the problem of explosive drones. It is a global problem. But we are going to solve it.

Clash Report

22,432 görüntüleme • 4 ay önce

Homan: "We don't racially profile. When we detain somebody and question somebody for a short period of time, we need reasonable suspicion. That's what we're doing. But people say 'The Supreme Court allows you to racial profile.' That's not what the Supreme Court said. They said we're doing things legally."

Aaron Rupar

81,495 görüntüleme • 8 ay önce

As sons, the loss of a father leaves a space that no triumph can ever truly fill. What we can do, however, is carry ourselves in a way that honours their memory, in how we stand, how we lead, and how we endure. Proud of my nephew. Spoke with conviction

As sons, the loss of a father leaves a space that no triumph can ever truly fill. What we can do, however, is carry ourselves in a way that honours their memory, in how we stand, how we lead, and how we endure. Proud of my nephew. Spoke with conviction

DON STRAPZY *SE DØNS*

24,019 görüntüleme • 7 ay önce

.President Donald J. Trump on Iran: "You can just keep doing exactly what we're doing, and take them apart piece by piece. We could do it at a more rapid fashion, which we might do — or we can negotiate with them, which we're also doing right now."

.President Donald J. Trump on Iran: "You can just keep doing exactly what we're doing, and take them apart piece by piece. We could do it at a more rapid fashion, which we might do — or we can negotiate with them, which we're also doing right now."

Rapid Response 47

168,807 görüntüleme • 3 ay önce