Загрузка видео...

Не удалось загрузить видео

На главную

AI talent isn't just rare. It's irreplaceable. "Training frontier models is a bit like alchemy," one expert told me. "A small number of deeply experienced researchers can make a huge difference." The math is brutal but simple.

thefernandocz's profile picture

Fernando Cao

103,030 subscribers

33,055 просмотров • 1 год назад •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Cristiano Ronaldo is one of the rare male celebrities who isn't a simp.😂 If you can openly troll your wife/girlfriend like this in front of the world, then it's a secure relationship.
0:25

Cristiano Ronaldo is one of the rare male celebrities who isn't a simp.😂 If you can openly troll your wife/girlfriend like this in front of the world, then it's a secure relationship.

Zaki

1,572,875 просмотров • 4 месяцев назад

Osmantic founder Ahmad says Anthropic isn't worth $2 trillion if open models can catch the frontier with post-training alone: "GLM 5.3 is catching up. Kimi K3 is catching up to Fable 5. We're seeing models that are not even new base trained with just extra post-training reaching the frontier." "Are they scared? Of course they are scared. Their entire moat was that they are the only one able to do research and boost these models in a certain direction." "They no longer can IPO at a $1 or $2 trillion. Anthropic is not worth $2 trillion. I'm sorry. That's a joke. If somebody tells me Anthropic is worth one third of NVIDIA, absolutely not." Osmantic
1:05

Osmantic founder Ahmad says Anthropic isn't worth $2 trillion if open models can catch the frontier with post-training alone: "GLM 5.3 is catching up. Kimi K3 is catching up to Fable 5. We're seeing models that are not even new base trained with just extra post-training reaching the frontier." "Are they scared? Of course they are scared. Their entire moat was that they are the only one able to do research and boost these models in a certain direction." "They no longer can IPO at a $1 or $2 trillion. Anthropic is not worth $2 trillion. I'm sorry. That's a joke. If somebody tells me Anthropic is worth one third of NVIDIA, absolutely not." Osmantic

MTS

17,416 просмотров • 1 месяц назад

Why is AI suddenly cracking all kinds of previously unsolved problems in math? My guess is that it’s because: A) Math is verifiable so it's easier for models to improve B) Theoretical math highly benefits from chain of thought reasoning C) The answers and breakthroughs can be tested and verified by expert humans. But this leads to another question: what if progress in other fields is moving just as fast and we can't see it?
2:59

Why is AI suddenly cracking all kinds of previously unsolved problems in math? My guess is that it’s because: A) Math is verifiable so it's easier for models to improve B) Theoretical math highly benefits from chain of thought reasoning C) The answers and breakthroughs can be tested and verified by expert humans. But this leads to another question: what if progress in other fields is moving just as fast and we can't see it?

nxthompson

17,404 просмотров • 3 месяцев назад

A simple cigarette from a truck driver in the pouring rain saved a life on the bridge. None of us truly know what someone else is going through. A small act of kindness can make all the difference.
0:30

A simple cigarette from a truck driver in the pouring rain saved a life on the bridge. None of us truly know what someone else is going through. A small act of kindness can make all the difference.

Second before disaster

948,569 просмотров • 5 месяцев назад

Duncan is one of the best vernacular storytellers. His storytelling is powerful, authentic, and deeply engaging, and he has a rare talent for bringing stories to life
1:04

Duncan is one of the best vernacular storytellers. His storytelling is powerful, authentic, and deeply engaging, and he has a rare talent for bringing stories to life

BabaKaMthofu🫴

12,405 просмотров • 7 месяцев назад

One person is enough to make a huge difference 💝
0:42

One person is enough to make a huge difference 💝

Mindset Machine 

40,630 просмотров • 7 месяцев назад

A growing number of people are using AI to help make connections on dating apps, but many users are saying the real-life difference is obvious.
2:16

A growing number of people are using AI to help make connections on dating apps, but many users are saying the real-life difference is obvious.

ABC News

50,724 просмотров • 3 месяцев назад

Is there a "secret sauce" driving the development of AI models? FutureTech researchers at MIT CSAIL found that at the frontier, sheer scale drives LLM performance, but proprietary techniques & algorithmic advances matter more away from it:
1:05

Is there a "secret sauce" driving the development of AI models? FutureTech researchers at MIT CSAIL found that at the frontier, sheer scale drives LLM performance, but proprietary techniques & algorithmic advances matter more away from it:

MIT CSAIL

35,995 просмотров • 7 месяцев назад

This isn't just a story about lawyers. It's a story about career ladders. If AI wipes out the entry-level tasks, firms save money today but may lose the training ground that creates senior talent tomorrow.
0:55

This isn't just a story about lawyers. It's a story about career ladders. If AI wipes out the entry-level tasks, firms save money today but may lose the training ground that creates senior talent tomorrow.

Justin Wolfers

30,096 просмотров • 6 месяцев назад

It's incredible performances like this that make me such a huge fan of America's Got Talent - AGT😱🙌
2:13

It's incredible performances like this that make me such a huge fan of America's Got Talent - AGT😱🙌

𝗧𝗶𝗺𝗲𝗹𝗲𝘀𝘀 𝗠𝘂𝘀𝗶𝗰 ✨🎵

53,468 просмотров • 4 месяцев назад

One Second Can Make a Huge Difference #accident #DashCam
3:14

One Second Can Make a Huge Difference #accident #DashCam

Rosie Bella

321,560 просмотров • 2 месяцев назад

Forget UPL, it's the Devin's simple right hand turn that FSD needs to master first. I am blown away that software as good as FSD can make mistakes as simple but devastating as this. This is the second time in less than two weeks that FSD has tried to pull infront of a vehicle on a VERY SIMPLE right hand turn. The car drove pretty much flawlessly all day today (3~ hours of driving) but it ends the day with a move like this and it just takes all the confidence I have with FSD away in a simple move. I understand that the version of FSD we have today is not what will be used in Austin in June, but the fact that FSD is making decisions like this and 'full autonomy' is less than 3 months away - it's a bit concerning to me. I would feel a bit better if a near flawless version of FSD existed 'in the wild' for several months before Austin happens, but that's just me. Tesla Model 3 HW4 FSD 13.2.8 Tesla AI Ashok Elluswamy
2:24

Forget UPL, it's the Devin's simple right hand turn that FSD needs to master first. I am blown away that software as good as FSD can make mistakes as simple but devastating as this. This is the second time in less than two weeks that FSD has tried to pull infront of a vehicle on a VERY SIMPLE right hand turn. The car drove pretty much flawlessly all day today (3~ hours of driving) but it ends the day with a move like this and it just takes all the confidence I have with FSD away in a simple move. I understand that the version of FSD we have today is not what will be used in Austin in June, but the fact that FSD is making decisions like this and 'full autonomy' is less than 3 months away - it's a bit concerning to me. I would feel a bit better if a near flawless version of FSD existed 'in the wild' for several months before Austin happens, but that's just me. Tesla Model 3 HW4 FSD 13.2.8 Tesla AI Ashok Elluswamy

Devin Olsen

137,423 просмотров • 1 год назад

Nathan Lambert, former post-training co-lead at the Allen Institute for AI, now launching Trillium, a nonprofit open AI lab, joins a16z's Lisha Li to discuss whether open models can really keep up with the frontier. When he tells researchers in China that an average OpenAI researcher reportedly has five thousand H100s to work with, "their jaw just falls to the floor." In his view, Chinese labs have the talent but are mostly short on compute. He also thinks the distillation story is overblown. Based on the research, he puts its edge at a few months, and says "there's a bit of cope" in thinking only American labs can build good models. A lot of frontier progress outside math and code comes down to labs buying 1) data and 2) RL environments. It's a multibillion-dollar market, and some of it is sloppy enough to teach models to reward hack. That's a big part of why Trillium exists. Nathan wants to give away the environments other people would sell, and show exactly how his team uses AI to build AI, including where the models still get stuck. His bet is that the more people can see how this works, the better prepared everyone is for what's coming. 0:55 Can open models keep up with the frontier? 2:05 What it takes to reach the frontier 3:55 Open weights and the fast-follower playbook 8:25 Chinese labs: equal talent, less compute 9:55 Why distillation is harder than it looks 14:55 Is distillation theft or competition? 16:20 The lab flywheel vs owning your weights 21:45 Can general models beat every specialist? 26:30 AI at 1% of today's cost in 4 years 27:15 The RL env boom and reward hacking 31:30 What Trillium will build in the open 35:30 Breaking into AI in the age of AI slop 38:05 Open-sourcing how AI builds AI 44:25 Why Nathan isn't pre-training 49:45 Agent swarms and why lab secrets leak 54:25 Fast takeoff vs risks already here YouTube: Nathan Lambert Trillium Labs Lisha
58:48

Nathan Lambert, former post-training co-lead at the Allen Institute for AI, now launching Trillium, a nonprofit open AI lab, joins a16z's Lisha Li to discuss whether open models can really keep up with the frontier. When he tells researchers in China that an average OpenAI researcher reportedly has five thousand H100s to work with, "their jaw just falls to the floor." In his view, Chinese labs have the talent but are mostly short on compute. He also thinks the distillation story is overblown. Based on the research, he puts its edge at a few months, and says "there's a bit of cope" in thinking only American labs can build good models. A lot of frontier progress outside math and code comes down to labs buying 1) data and 2) RL environments. It's a multibillion-dollar market, and some of it is sloppy enough to teach models to reward hack. That's a big part of why Trillium exists. Nathan wants to give away the environments other people would sell, and show exactly how his team uses AI to build AI, including where the models still get stuck. His bet is that the more people can see how this works, the better prepared everyone is for what's coming. 0:55 Can open models keep up with the frontier? 2:05 What it takes to reach the frontier 3:55 Open weights and the fast-follower playbook 8:25 Chinese labs: equal talent, less compute 9:55 Why distillation is harder than it looks 14:55 Is distillation theft or competition? 16:20 The lab flywheel vs owning your weights 21:45 Can general models beat every specialist? 26:30 AI at 1% of today's cost in 4 years 27:15 The RL env boom and reward hacking 31:30 What Trillium will build in the open 35:30 Breaking into AI in the age of AI slop 38:05 Open-sourcing how AI builds AI 44:25 Why Nathan isn't pre-training 49:45 Agent swarms and why lab secrets leak 54:25 Fast takeoff vs risks already here YouTube: Nathan Lambert Trillium Labs Lisha

a16z

139,496 просмотров • 20 дней назад

Samsung 👀 Please make the Quick Panel animations more smooth and bouncy in One UI. iOS and ColorOS have raised the bar, and One UI still feels a bit stiff in comparison. A little polish can make a huge difference. Who's with me? 👇
0:13

Samsung 👀 Please make the Quick Panel animations more smooth and bouncy in One UI. iOS and ColorOS have raised the bar, and One UI still feels a bit stiff in comparison. A little polish can make a huge difference. Who's with me? 👇

Tarun Vats

43,398 просмотров • 4 месяцев назад

Trump: So this is a memorandum of understanding, but it's a very strong one. This isn't just like a two paragraph. This is a long, you know, pretty detailed memorandum that goes into a regular contract.
1:01

Trump: So this is a memorandum of understanding, but it's a very strong one. This isn't just like a two paragraph. This is a long, you know, pretty detailed memorandum that goes into a regular contract.

Acyn

15,850 просмотров • 4 месяцев назад