Dwarkesh Podcast

We Must Pace the Frontier, On the Navier–Stokes Millennium Prize Problem, and more ›

Covered by2 sites

LessWrong, and more ›

Covers 24 stories

AI researchers debate how close we are to recursive self-improvement

Dwarkesh Podcast1w1 week agoHNContent type: Audio

GPT-5.6, Do you respect 'Vibe Coders'? Can you actually call them devs?, and more ›

Covers 57 stories

Dwarkesh PodcastDwarkesh PatelJerry Han

Pretraining progress is mostly coming from data

Dwarkesh Podcast1w1 week agoHN Covered by510 min read

Covered by5 sites

All 5 sites ›

Covers 4 stories

FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness, Language Models are Unsupervised Multitask Learners, and more ›

Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face

Dwarkesh Podcast2w2 weeks agoHN2 Covered by4045+ min read

newer thread·older thread

Covered by40 sites

LessWrong, jerpint, and more ›

HN2

Covers 38 stories

Security incident disclosure — July 2026, Brief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident, and more ›

The rise and fall of agent civilizations

Dwarkesh Podcast2w2 weeks agoContent type: Audio

The Rise and Fall of Agent Civilizations

Dwarkesh Podcast2w2 weeks agoHN Covered by8315 min read

Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the Incident, OpenAI Hugging Face Incident Technical Report, and more ›

Covered by83 sites

exe blog, The Verge, and more ›

Covers 8 stories

Dylan Patel – Anthropic and OpenAI will have most of world’s compute by 2028

Dwarkesh Podcast3w3 weeks agoHN Covered by645+ min read

Open Router- A unified interface for LLMs, Previewing GPT‑5.6 Sol: a next-generation model, and more ›

Covered by6 sites

beren.io, and more ›

Covers 33 stories

Ryan Greenblatt – Human level AIs might build runaway superintelligences by 2032

Dwarkesh Podcast5w5 weeks agoCovered by845+ min read

Scaling Laws for Neural Language Models, ability of an artificial intelligence to modify itself to further improve its capability, and more ›

Covered by8 sites

Substack, thezvi.wordpress.com, and more ›

Covers 13 stories

Predictions for the Era of Continual Learning

Dwarkesh Podcast6w6 weeks agoHN Covered by65 min read

Covered by6 sites

The Decoder, Mark's Blog, and more ›

Why smarter AI models could drive up compute prices 10x

Dwarkesh Podcast6w6 weeks agoHN

Why compute might get 10x+ more expensive in coming years

Dwarkesh Podcast7w7 weeks agoHN Covered by106 min read

AI Value Capture, Wikipedia, and more ›

Covered by10 sites

Forward Future, and more ›

Covers 4 stories

Adam Brown – A deep but accessible introduction to general relativity

Dwarkesh Podcast10w10 weeks ago45+ min read

keyboard_arrow_right Learn more, TIL that Isaac Newton, the scientist who formulated the laws of motion and gravity, was born on December 25, 1642. While most of the world celebrates Christmas,..., and more ›

Covers 5 stories

The Winning Essays for the Big Questions About AI

Dwarkesh Podcast11w11 weeks agoHN Covered by315 min read

Mechanize, Inc, Can AI companies become profitable?

Covered by3 sites

3 Quarks Daily, and more ›

Covers 2 stories

Grant Sanderson – AI and the future of math

Dwarkesh Podcast11w11 weeks agoHN Covered by245+ min read

An OpenAI model has disproved a central conjecture in discrete geometry, TIL that AI models like ChatGPT don’t understand language like a real mind does. They predict the next word based on patterns in data., and more ›

Covered by2 sites

AGI Friday, and more ›

Covers 11 stories

The next big breakthrough will be AIs learning on the job

Dwarkesh Podcast12w12 weeks agoHN Covered by515 min read

Online Business Banking For Startups & Small Businesses, Gemini Agentic Video Understanding, and more ›

Covered by5 sites

hans.gerwitz.com, and more ›

Covers 3 stories

The data black hole at the center of AI

Dwarkesh Podcast13w13 weeks agoHN7 min read

Anthropic raises $65B in Series H funding at $965B post-money valuation, Open models lag state-of-the-art closed models by 4 months, and more ›

Covers 4 stories

Ada Palmer – Machiavelli is the most misunderstood thinker of all time

Dwarkesh Podcast13w13 weeks agoCovered by345+ min read

Age of Enlightenment, TIL that Machiavelli, prior to being exiled from Florence, was arrested, tortured, and imprisoned for 3 weeks on treason. Not only were the charges false, the o...

Covered by3 sites

DeLong's Grasping Reality, and more ›

Covers 2 stories

The sample efficiency black hole

Dwarkesh Podcast14w14 weeks agoHN Covered by87 min read

Anthropic raises $65B in Series H funding at $965B post-money valuation, Open models lag state-of-the-art closed models by 4 months

Covered by8 sites

All 8 sites ›

Covers 2 stories

Alex Imas and Phil Trammell – What remains scarce after AGI?

Dwarkesh Podcast15w15 weeks agoHN45+ min read

Jevons Paradox, TIL that AI models like ChatGPT don’t understand language like a real mind does. They predict the next word based on patterns in data., and more ›

Covers 13 stories

Reiner Pope – Chip design from the bottom up

Dwarkesh Podcast17w17 weeks agoHN45+ min read

How To Scale Your Model, +12 years of programming, now what?, and more ›

Covers 6 stories