Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

Google's Jeff Dean says current pre-training is passive: initialize a model, stream the internet past it, let it observe But models need to learn not just from data, but by acting, predicting, and choosing what to learn from next "we have this artificial distinction now between pre and post-training,... show more

Haider.

71,942 subscribers

55,812 просмотров • 7 месяцев назад •via X (Twitter)

Наука и технологии

Anya Rossi• Live Now

Private livecam show

Комментарии: 8

Фото профиля Zuzanna Stamirowska

Zuzanna Stamirowska7 месяцев назад

BDH trains at inference:

Фото профиля Vanar

Vanar7 месяцев назад

Learning shouldn’t stop after pre training. Agents that act, learn from outcomes, and update their own context are where AI becomes truly adaptive

Фото профиля Harry Tandy

Harry Tandy7 месяцев назад

static data is over models that choose their own path will dominate 2026

Фото профиля Michael Stewart

Michael Stewart7 месяцев назад

And yet, they cannot figure out the architecture that would give them what they want, while we at ChronoDyne Systems, Inc. capitalize and build an IP moat as wide as the Mississippi.

Фото профиля Idan | Helping SaaS Win in AI Search

Idan | Helping SaaS Win in AI Search7 месяцев назад

Breakthrough is models telling us what data they actually need.

Фото профиля Bradley K. Sherman

Bradley K. Sherman7 месяцев назад

Translation: Our Tech Overlords are out of ideas.

Фото профиля Jeremiah Adebayo

Jeremiah Adebayo7 месяцев назад

I concur. Teaching the model how to think is more important

Фото профиля RetroRanger Anonymous

RetroRanger Anonymous7 месяцев назад

jeff gets it. pre-training is my warmup, post-training is where i lift. just watching never built muscle

Похожие видео

GLM 5.3’s biggest leap came from post-training alone. “GLM 5.3 was built entirely by post-training the previous model. For me, that was a wake-up call.” “Post-training is far more GPU-efficient than pre-training from scratch. Plenty of companies can’t afford to train a model from scratch, but they have proprietary data. They can post-train themselves and beat some of the closed models.” “There’s still so much juice left to squeeze from post-training. We compared GLM 5.3’s cyber capabilities with 5.2, and it’s a night and day difference. That’s all just post-training.”

GLM 5.3’s biggest leap came from post-training alone. “GLM 5.3 was built entirely by post-training the previous model. For me, that was a wake-up call.” “Post-training is far more GPU-efficient than pre-training from scratch. Plenty of companies can’t afford to train a model from scratch, but they have proprietary data. They can post-train themselves and beat some of the closed models.” “There’s still so much juice left to squeeze from post-training. We compared GLM 5.3’s cyber capabilities with 5.2, and it’s a night and day difference. That’s all just post-training.”

SemiAnalysis

214,839 просмотров • 20 дней назад

"We’ve moved past it, but we have to learn from it. … We need some guys to dig in and stack some consistent games together." Hear from head coach Rick Tocchet following today's practice at Flyers Training Center.

"We’ve moved past it, but we have to learn from it. … We need some guys to dig in and stack some consistent games together." Hear from head coach Rick Tocchet following today's practice at Flyers Training Center.

Philadelphia Flyers

23,303 просмотров • 9 месяцев назад

New Course: Post-training of LLMs Learn to post-train and customize an LLM in this short course, taught by Banghua Zhu, Assistant Professor at the University of Washington University of Washington, and co-founder of @NexusflowX. Training an LLM to follow instructions or answer questions has two key stages: pre-training and post-training. In pre-training, it learns to predict the next word or token from large amounts of unlabeled text. In post-training, it learns useful behaviors such as following instructions, tool use, and reasoning. Post-training transforms a general-purpose token predictor—trained on trillions of unlabeled text tokens—into an assistant that follows instructions and performs specific tasks. Because it is much cheaper than pre-training, it is practical for many more teams to incorporate post-training methods into their workflows than pre-training. In this course, you’ll learn three common post-training methods—Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning (RL)—and how to use each one effectively. With SFT, you train the model on pairs of input and ideal output responses. With DPO, you provide both a preferred (chosen) and a less preferred (rejected) response and train the model to favor the preferred output. With RL, the model generates an output, receives a reward score based on human or automated feedback, and updates the model to improve performance. You’ll learn the basic concepts, common use cases, and principles for curating high-quality data for effective training. Through hands-on labs, you’ll download a pre-trained model from Hugging Face and post-train it using SFT, DPO, and RL to see how each technique shapes model behavior. In detail, you’ll: - Understand what post-training is, when to use it, and how it differs from pre-training. - Build an SFT pipeline to turn a base model into an instruct model. - Explore how DPO reshapes behavior by minimizing contrastive loss—penalizing poor responses and reinforcing preferred ones. - Implement a DPO pipeline to change the identity of a chat assistant. - Learn online RL methods such as Proximal Policy Optimization (PPO) and Group Relative Policy Optimization (GRPO), and how to design reward functions. - Train a model with GRPO to improve its math capabilities using a verifiable reward. Post-training is one of the most rapidly developing areas of LLM training. Whether you’re building a high-accuracy context-specific assistant, fine-tuning a model's tone, or improving task-specific accuracy, this course will give you experience with the most important techniques shaping how LLMs are post-trained today. Please sign up here:

New Course: Post-training of LLMs Learn to post-train and customize an LLM in this short course, taught by Banghua Zhu, Assistant Professor at the University of Washington University of Washington, and co-founder of @NexusflowX. Training an LLM to follow instructions or answer questions has two key stages: pre-training and post-training. In pre-training, it learns to predict the next word or token from large amounts of unlabeled text. In post-training, it learns useful behaviors such as following instructions, tool use, and reasoning. Post-training transforms a general-purpose token predictor—trained on trillions of unlabeled text tokens—into an assistant that follows instructions and performs specific tasks. Because it is much cheaper than pre-training, it is practical for many more teams to incorporate post-training methods into their workflows than pre-training. In this course, you’ll learn three common post-training methods—Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning (RL)—and how to use each one effectively. With SFT, you train the model on pairs of input and ideal output responses. With DPO, you provide both a preferred (chosen) and a less preferred (rejected) response and train the model to favor the preferred output. With RL, the model generates an output, receives a reward score based on human or automated feedback, and updates the model to improve performance. You’ll learn the basic concepts, common use cases, and principles for curating high-quality data for effective training. Through hands-on labs, you’ll download a pre-trained model from Hugging Face and post-train it using SFT, DPO, and RL to see how each technique shapes model behavior. In detail, you’ll: - Understand what post-training is, when to use it, and how it differs from pre-training. - Build an SFT pipeline to turn a base model into an instruct model. - Explore how DPO reshapes behavior by minimizing contrastive loss—penalizing poor responses and reinforcing preferred ones. - Implement a DPO pipeline to change the identity of a chat assistant. - Learn online RL methods such as Proximal Policy Optimization (PPO) and Group Relative Policy Optimization (GRPO), and how to design reward functions. - Train a model with GRPO to improve its math capabilities using a verifiable reward. Post-training is one of the most rapidly developing areas of LLM training. Whether you’re building a high-accuracy context-specific assistant, fine-tuning a model's tone, or improving task-specific accuracy, this course will give you experience with the most important techniques shaping how LLMs are post-trained today. Please sign up here:

Andrew Ng

126,309 просмотров • 1 год назад

Most people have heard of big model smell, but what about pristine pre-training smell? Evan and Derry seem to think it exists.

Most people have heard of big model smell, but what about pristine pre-training smell? Evan and Derry seem to think it exists.

Arena.ai

12,704 просмотров • 7 месяцев назад

In case you missed it, we recently launched "Post-training of LLMs," a short course where you'll: ✅ Understand when and why to use post-training methods like Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning. ✅ Learn the concepts underlying the three post-training methods of SFT, DPO, and Online RL, their common use-cases, and how to curate high-quality data to effectively train a model using each method. ✅ Download a pre-trained model and implement post-training pipelines to turn a base model into an instruct model, change the identity of a chat assistant, and improve a model’s math capabilities. Learn more and enroll for free:

In case you missed it, we recently launched "Post-training of LLMs," a short course where you'll: ✅ Understand when and why to use post-training methods like Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning. ✅ Learn the concepts underlying the three post-training methods of SFT, DPO, and Online RL, their common use-cases, and how to curate high-quality data to effectively train a model using each method. ✅ Download a pre-trained model and implement post-training pipelines to turn a base model into an instruct model, change the identity of a chat assistant, and improve a model’s math capabilities. Learn more and enroll for free:

DeepLearning.AI

16,779 просмотров • 1 год назад

We discovered an emergent property of VLAs like π0/π0.5/π0.6: as we scale up pre-training, the model learns to align human videos and robot data! This gives us a simple way to leverage human videos. Once π0.5 knows how to control robots, it can naturally learn from human video.

We discovered an emergent property of VLAs like π0/π0.5/π0.6: as we scale up pre-training, the model learns to align human videos and robot data! This gives us a simple way to leverage human videos. Once π0.5 knows how to control robots, it can naturally learn from human video.

Physical Intelligence

1,189,523 просмотров • 10 месяцев назад

Former Israeli PM Naftali Bennett: Just as we have 8200, we need to establish a 9200 — not a military unit, but a state organization devoted to influence and consciousness. Basically, we need to learn from Qatar — learn what they did, and do it ourselves. We can’t allow them to tear us apart like this. They are tearing us apart. We’re completely finished in this arena right now. But this too has to be fixed. It has to be fixed — and it will take time.

Former Israeli PM Naftali Bennett: Just as we have 8200, we need to establish a 9200 — not a military unit, but a state organization devoted to influence and consciousness. Basically, we need to learn from Qatar — learn what they did, and do it ourselves. We can’t allow them to tear us apart like this. They are tearing us apart. We’re completely finished in this arena right now. But this too has to be fixed. It has to be fixed — and it will take time.

Clash Report

87,396 просмотров • 2 месяцев назад

Jensen just said he was reading a text from DeepMind’s Demis confirming it: pre-training + post-training scaling laws are fully intact, and Gemini 3 is the proof. A huge jump in quality exactly what Blackwell and Rubin were built for. This next curve is going to be wild. $NVDA

Jensen just said he was reading a text from DeepMind’s Demis confirming it: pre-training + post-training scaling laws are fully intact, and Gemini 3 is the proof. A huge jump in quality exactly what Blackwell and Rubin were built for. This next curve is going to be wild. $NVDA

Ben Pouladian

327,245 просмотров • 11 месяцев назад

Andrej reveals that Grass has crossed $35M in trailing 12-month revenue from AI pre-training data sales, with H1 revenue growing nearly 7x YoY “Every single book that any human has ever published represents one percent of the amount of data that current pre-training runs actually need to properly scale” “When you think of things from that perspective, it becomes pretty obvious why there's a massive data shortage and why we've been focusing so much on pre-training data” “This has also translated to certain business metrics. Grass has 7x revenue purely on pre-training data sales since it began monetizing data” “It went from $2.7 million in the first half of last year, and then in the last 12 months crossed $35 million” “Thinking of things from this scale is incredibly interesting, but it also illustrates why our focus has been where it's been” w/ Andrej

Andrej reveals that Grass has crossed $35M in trailing 12-month revenue from AI pre-training data sales, with H1 revenue growing nearly 7x YoY “Every single book that any human has ever published represents one percent of the amount of data that current pre-training runs actually need to properly scale” “When you think of things from that perspective, it becomes pretty obvious why there's a massive data shortage and why we've been focusing so much on pre-training data” “This has also translated to certain business metrics. Grass has 7x revenue purely on pre-training data sales since it began monetizing data” “It went from $2.7 million in the first half of last year, and then in the last 12 months crossed $35 million” “Thinking of things from this scale is incredibly interesting, but it also illustrates why our focus has been where it's been” w/ Andrej

Kaiz 🐂🀄️

14,984 просмотров • 25 дней назад

Ilya Sutskever says that data is the fossil fuel of AI and "we have achieved peak data" because "we have but one internet", so the age of pre-training AI on bigger and bigger compute clusters "will unquestionably end"

Ilya Sutskever says that data is the fossil fuel of AI and "we have achieved peak data" because "we have but one internet", so the age of pre-training AI on bigger and bigger compute clusters "will unquestionably end"

Tsarathustra

127,222 просмотров • 1 год назад

Introducing the OpenRouter MCP, live model intelligence right inside your agent Your agent builds and ships, but when it comes to choosing the right model for the right job, it guesses from 6 month old training data Watch it pick, price, and test the right model:

Introducing the OpenRouter MCP, live model intelligence right inside your agent Your agent builds and ships, but when it comes to choosing the right model for the right job, it guesses from 6 month old training data Watch it pick, price, and test the right model:

OpenRouter

265,125 просмотров • 4 месяцев назад

This is the part of AI that doesn’t get talked about enough. LLMs had the entire internet to learn from. Robotics doesn’t have that luxury. A robot can’t learn what it feels like to iron clothes while a toddler is running around, restock a shelf in a crowded store, or assemble something on a messy factory floor just by reading about it. It has to experience the physical world. And that’s why what Humyn Labs is building is so important. They’re turning real human experiences into model-ready data across sound, sight, motion and touch, so Physical AI can learn from the world as it actually exists, not some perfectly controlled laboratory. The next breakthrough in robotics may not come from making the model bigger. It may come from giving the model better experiences to learn from. That’s the human layer Physical AI has been missing. 🟧

This is the part of AI that doesn’t get talked about enough. LLMs had the entire internet to learn from. Robotics doesn’t have that luxury. A robot can’t learn what it feels like to iron clothes while a toddler is running around, restock a shelf in a crowded store, or assemble something on a messy factory floor just by reading about it. It has to experience the physical world. And that’s why what Humyn Labs is building is so important. They’re turning real human experiences into model-ready data across sound, sight, motion and touch, so Physical AI can learn from the world as it actually exists, not some perfectly controlled laboratory. The next breakthrough in robotics may not come from making the model bigger. It may come from giving the model better experiences to learn from. That’s the human layer Physical AI has been missing. 🟧

Bahdfella™ ❤️‍🔥

24,709 просмотров • 2 месяцев назад

We introduce Tagger Phase 2, bringing manual #data labeling and #data review into the #DeCorp ecosystem. With this iteration, we have completed the final components necessary for a fully integrated and decentralized cycle: 1) Data processing from scratch 2) Peer-to-peer data validation 3) Automated pre-training of AI copilot models 4) Quasi-autonomous active learning It is not long now before you may directly engage with this unprecedented and comprehensive system. See you soon.

We introduce Tagger Phase 2, bringing manual #data labeling and #data review into the #DeCorp ecosystem. With this iteration, we have completed the final components necessary for a fully integrated and decentralized cycle: 1) Data processing from scratch 2) Peer-to-peer data validation 3) Automated pre-training of AI copilot models 4) Quasi-autonomous active learning It is not long now before you may directly engage with this unprecedented and comprehensive system. See you soon.

Tagger

104,265 просмотров • 1 год назад

We need to destroy the data centers and start over from a pre internet era!! MARC REBILLET

We need to destroy the data centers and start over from a pre internet era!! MARC REBILLET

SubwayTakes with Kareem Rahma

55,385 просмотров • 1 год назад

"We found a way to win the game but we need to learn from it" Keith Andrews reflects on Brentford's 4-3 win over Burnley on Saturday and what his side can learn from their performance.

"We found a way to win the game but we need to learn from it" Keith Andrews reflects on Brentford's 4-3 win over Burnley on Saturday and what his side can learn from their performance.

Sky Sports News

74,785 просмотров • 8 месяцев назад

Demis Hassabis says the next AI leap is continual learning. Not train once and freeze, but learn from real experience, constantly. AlphaGo did it in games. Now comes the messy real world. Would you let an AI learn from you? 👀

Demis Hassabis says the next AI leap is continual learning. Not train once and freeze, but learn from real experience, constantly. AlphaGo did it in games. Now comes the messy real world. Would you let an AI learn from you? 👀

VraserX e/acc

19,202 просмотров • 9 месяцев назад

Grass is betting that better AI still starts with better pre-training data “Whatever you put into your model ultimately determines what your model's capabilities are” “Several of the top frontier AI labs in the world are working directly with Grass to acquire large amounts of pre-training data” “If that information is not present in your model, your model doesn't know how to do that thing in the first place” Andrej puts Grass close to one of the most fundamental inputs in AI: the training corpus itself. As frontier models scale, they need enormous amounts of diverse, high-quality web data. They are building the infrastructure to collect and supply that data at internet scale. Model capability starts with what the model gets to learn from.

Grass is betting that better AI still starts with better pre-training data “Whatever you put into your model ultimately determines what your model's capabilities are” “Several of the top frontier AI labs in the world are working directly with Grass to acquire large amounts of pre-training data” “If that information is not present in your model, your model doesn't know how to do that thing in the first place” Andrej puts Grass close to one of the most fundamental inputs in AI: the training corpus itself. As frontier models scale, they need enormous amounts of diverse, high-quality web data. They are building the infrastructure to collect and supply that data at internet scale. Model capability starts with what the model gets to learn from.

Kaiz 🐂🀄️

13,989 просмотров • 1 месяц назад