New Unsupervised Learning with Google AI researchers @NoamShazeer and @jack_w_rae on: - Scaling test-time compute - The power of the Mom eval - The pace of Open Source / DeepSeek - How AI research is where chemistry was in the 15th century - Reactions to Ilya on how

New Unsupervised Learning with @oliviergodement, the Head of Product for Business Products at OpenAI. We discussed lessons from his work with 200+ enterprises deploying AI across use cases from customer support to drug discovery. Some favorite parts: - How startups should

Very fun to get to break some news on Vital Signs with @CMSinnovates Director @AbeSutton and Chief AI/Technology Officer @JacobShiff. We talked about everything related to the exciting new ACCESS model, which opens up the Medicare fee-for-service market to digital health

Diane (Head of Product for Research at @Anthropic) discussed how her perspective on AI progress has shifted, countering recent skepticism about agent timelines. Q: What's one thing you've changed your mind on in AI in the last year? "I actually think we are closer to

ChinaTalk's @jordanschneider came on Unsupervised Learning this week in an episode 15+ years in the making. Jordan is my go-to expert on all things China x AI. His hugely popular newsletter and podcast cover China's technology ecosystem, geopolitics, and economic policy. And

New Unsupervised Learning with Anthropic’s first PM, Dianne Na Penn on everything Opus 4.5 including: - Computer use reality: where it works today and what Opus 4.5 unlocks - How Anthropic builds models - How Anthropic uses Opus 4.5 internally - The evolution of

Our model can now learn from its own experience with RL! Our new π*0.6 model can more than double throughput over a base model trained without RL, and can perform real-world tasks: making espresso drinks, folding diverse laundry, and assembling boxes. More in the thread below.

AI is moving in dog years. I sat down with @Redpoint MD @jacobeffron to talk models vs agents, vertical AI in healthcare and legal, why “T2D3 is dead,” and where the real defensibility will come from in AI apps and tooling. The link to TFR and Episode 496 is in our bio.

We just raised our $75M Series C led by @stepstonegroup, making @ScribeHow a $1.3B unicorn. But that’s not the part I’m most proud of. The funding isn’t the story — what we DO with it is. I set out to change how the world actually works. I started Scribe with one belief:

@SutterHealth Check out the full episode: YouTube: https://t.co/tm3p343rgN Spotify: https://t.co/Mz7AjKHXak Apple:

On our latest Vital Signs, I sat down with Warner Thomas, CEO of @SutterHealth, one of the largest health systems in America. We explored how major health systems are navigating AI adoption, and the evolving healthcare landscape. We hit on: - Sutter's ambulatory expansion -

@billpeeb Check out the full episode on Unsupervised Learning⬇️ Spotify: https://t.co/Z18c2ang3f Apple: https://t.co/juKtEajipC YouTube:

One of my favorite parts of talking with the Sora team: @billpeeb on how this is just the start of exponential improvements in video models and his prediction that they'll drive novel discoveries in classical physics by early 2028

Was really special to do the first long form interview with the Sora team on Unsupervised Learning. I sat down with @billpeeb, @turtlesoupy and @rohanjamin to discuss building the first big AI-native social network, getting such a massive amount of users to create and continuing

$150M Series C. $1.8B valuation. One mission: to empower every lawyer with world-class AI. We build AI with the people who use it. Every success at Legora starts with collaboration: co-developing, iterating, and scaling technology hand in hand with our customers.

Legora is a special company. They're growing insanely fast with customers that absolutely love the product. A huge congratulations to the team on this milestone. Check out their open roles ( https://t.co/C8xgbWilUQ) to join this 🚀

Hit on a bunch of fun topics in our second "AI Roundup" episode with @_RobToews (AI investor at Radical) and @arimorcos (Co-Founder/CEO of Datology): - Reactions to Karpathy's claim on agents being overhyped - Are we in a bubble? - Do you need to be in SF to build an AI

New Vital Signs with @nikillinit and Viren Shetty, Vice Chairman of @NarayanaHealth on: - How extreme standardization and physician specialization drive down costs - Lessons for the rest of the world - Where AI actually works Some great insights on building

🎧Listen Here: Spotify: https://t.co/4NQT3E9Gzh Apple: https://t.co/ZCzBSKieSh YouTube:

With AI talent wars, lots of RL Environment companies being started, high foundation model valuations, and Google's comeback, it's been a wild couple of months in AI. We're introducing a new format at Unsupervised Learning discussing these topical questions alongside our

This was quite a surprising part of my interview with Tri on Unsupervised Learning and reinforces that many of the people who benefit most from models are at the cutting edges of their fields