Results for “training”

14,225 results

Transcript mentions 9,657

See all ›

State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI | Lex Fridman Podcast #490

Lex Fridman87 views
1:17the post-training lead at the Allen Institute for AI, and author of the definitive book on Reinforcement Learning from Human Feedback.8:06are using the ideas from DeepSeek. For example, you mentioned Kimi: same architecture, they're training it. And then again, we have this leapfrogging35:55you can add more data post-training. Like, specialize them into, let's say, law, medical models, whatever you

Topics & chapters 3,812

See all ›

AI summaries 756

See all ›

Black John Wick on What MO3 Should Have Done on Highway, Karmelo Anthony, Charleston White Shot At

RealLyfe Productions3 views
Rick Childs, a professional security expert and pilot, discusses the critical importance of professionalism, continuous training, and legal knowledge in security and self-defense.summaryThe conversation also covers the realities of celebrity security, the inadequacies of typical law enforcement training, and the rigorous standards required for aviation safety.summaryTraining vs. Expectation: In high-stress situations, individuals fall to their level of training, not their level of expectation. Civilians with concealed carry permits often lack the practical training to effectively ustakeaway

The REAL Way to Get Big & Strong (No BS) | Dr. Mike Israetel, Roman Mori & Bryan Callen

Bryan Callen23 views
The conversation argues that athletic performance is primarily driven by sport-specific practice and raw strength/muscle size, rather than complex functional training or balance exercises.summaryStrength and conditioning for athletes should focus on making the body 'big and strong' through resistance training, as this provides the physical architecture necessary for sport, whereas coordination is task-specific atakeawayHigh-intensity strength training requires long rest periods (3+ minutes) to ensure each set is performed with maximum intent and quality, preventing the session from devolving into hypertrophy or metabolic conditioning wtakeaway

Greg Brockman

Tetragrammaton with Rick Rubin38 views
The conversation also covers the technical evolution of AI, from pre-training to agentic workflows, and the guest's personal journey from a rural background to co-founding Stripe and OpenAI.summaryTechnical Evolution (Pre-training vs Post-training): Pre-training is described as building a 'Library of Congress' (learning underlying rules from vast data), while post-training is like giving the machine 'taste' (steertakeaway“Pre-training is like the Library of Congress... Post-training is almost taste, giving the machine a sense of which of those books it would like or what to do with that information once it's retrieved.”quote

“Giant F*cking Scam!” Exercise Scientist Exposes Biggest Lies In Health & Nutrition | Mike Israetel

Dalton Fischer Podcast17 views
Mike Israetel emphasizes that consistency in training and nutrition is the primary driver of results, debunking myths about organic food, diet soda, and the necessity of extreme training styles.summaryConsistency is the most critical factor in hypertrophy; training 2-3 times a week for 30 minutes consistently yields better results than sporadic high-volume sessions.takeawayMyo-reps (short rest periods between mini-sets) allow for high volume training in less time with similar hypertrophy outcomes to traditional long-rest sets.takeaway

State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI | Lex Fridman Podcast #490

Lex Fridman87 views
Data Quality over Quantity: The industry is shifting focus from just scaling compute to curating high-quality training data. Synthetic data generation, filtering for educational value, and using specific sources like arXtakeaway“It's like a better base model... you don't let's say have your most complex tasks during pre-training or after pre-training... In an ideal world, you want to do all of them.”quote