Today’s LLMs are painfully slow and expensive. They are autoregressive and spit out words sequentially. One. At. A. Time. Our dLLMs generate text in parallel, delivering answers up to 10X faster. Now we’ve raised $50M to scale them. Full story from Russell Brandom in

Congratulations Stefano Ermon and the Inception team on their latest Mercury refresh. Best-in-class performance and up to 10× faster latency — a huge step forward for diffusion LLMs. Proud investor :)

Excited to finally announce Subtle Computing! We're building a new class of voice interfaces to make voice computing work where it's never worked before. Our first product arrives early 2026. A step towards a future where we look up at one another, not down at our phones.

Meet Pomelli, our latest Google Labs experiment It's for busy SMBs who need help with marketing campaigns Launched 9 days ago, there are 3 reasons why the early SMBs are loving it…

Enjoyed my conversation with Amjad Masad and Erik Torenberg!

Another cool new benchmark testing software engineering vs. coding: codeclash.ai Initial findings for LLM coding agents: - Significantly behind human programmers - Poor at multi-round iterative improvement - Rampant codebase sprawl/tech debt Lots left to do!

Ilya Sutskever "Three lines of math can prove all of supervised learning" (4:33) "I have not seen an exposition of unsupervised learning that I found satisfying" (7:50) Optimization objective has little relation to the actual objective you care about! youtube.com/watch?v=AKMuA_…

We grew from zero to $100M ARR and 70M users in <2.5 years, profitably it cost us millions in experiments to learn what worked 1100-word post on every growth hack that got us here I'll cover: 1. How to launch a feature in one day 2. How to find top 0.01% talent and keep them

.Satya Nadella gave me and Dylan Patel an exclusive tour of Fairwater 2, the most powerful AI datacenter in the world. We then chatted through Satya's vision for Microsoft in a world with AGI. 0:00:00 - Fairwater 2 0:04:15 - Business models for AGI 0:13:42 - Copilot 0:20:56 -

Join us for LiveKit DevDay, happening Nov 18 in San Francisco. Attend the keynote (3–4 PM) virtually or in person to see new product launches and hear the latest on voice AI. If you can’t make the keynote, LiveKit DevDay is an open house! Stop by anytime from 2–7 PM for

It was fun hearing about all the frontier work everyone is doing at AGI House x NFX: AI Founder Pitch Night, Harvard i-lab. Everyone was super pumped, from high school students to PhDs, an entire range with fascinating viewpoints. #AI #AGI #NFX #Harvard #MIT

What's next for AI agents? Our AI Engineer Sam Stowers 🌎 💪 🌹 spoke with AGI House at the self-evolving agent hackathon about the critical themes he's watching: novel memory, continuous learning, and the growing importance of RL. Watch the full interview below!

Have a bunch of handwritten notes from class or work? No problem. Now, you can add images as sources into NotebookLM and start transforming them into quizzes, flashcards, podcasts, and more!

The culture at Google Labs is unreal. I pitched an idea. Coworker: "Cool, let's build it." <immediately opens laptop>. We were testing it in under an hour. This bias for action is everything.

Sharing an interesting recent conversation on AI's impact on the economy. AI has been compared to various historical precedents: electricity, industrial revolution, etc., I think the strongest analogy is that of AI as a new computing paradigm (Software 2.0) because both are

Everyone wants to BUILD now… Level of prior success doesn’t matter in this wave. See it all the time, second time founders, experienced execs, etc getting off the bench to do something new Being in the arena when it matters is highly motivating, Carpe Diem vibes

We’ve been intensely cooking Gemini 3 for a while now, and we’re so excited and proud to share the results with you all. Of course it tops the leaderboards, including lmarena.ai, HLE, GPQA etc, but beyond the benchmarks it’s been by far my favourite model to use for its style and

Gemini 3 models from Google Google DeepMind have made a significant 2X SOTA jump on ARC-AGI-2 (Semi-Private Eval) Gemini 3 Pro: 31.11%, $0.81/task Gemini 3 Deep Think (Preview): 45.14%, $77.16/task

I played with Gemini 3 yesterday via early access. Few thoughts - First I usually urge caution with public benchmarks because imo they can be quite possible to game. It comes down to discipline and self-restraint of the team (who is meanwhile strongly incentivized otherwise) to

Gone are the days of Google being a gen AI laggard! Developers and founders are raving about Gemini 3: one coding assistant founder told us that Gemini 3 scored 17 percentage points higher than Claude Sonnet 4.5 in their internal benchmarks: theinformation.com/articles/googl…