The AI market today is bacon in a hot skillet. Everything is sizzling, moving, & changing at an incredible pace. We’re all watching it closely. Market share is fluid because no one yet knows what AI can do & the second we think have grasped it, models improve. The Nvidia chip

Great conversation with Harry, Jason & Rory on where AI, coding, and venture capital are heading. Thanks for having me on!

We're back! @HarryStebbings @rodriscoll me + special guest @ttunguz!! 🧵Cursor just crossed $1B ARR + raised $2.3B at $29.3B valuation with ~ 300 employees. And got to scale with just 30. Stripe's doing tender offers at all-time highs instead of going public. The entire

Most of the hard problems in AI aren't what you think.

This was a ton of fun! Great to be with you all to talk about the fastest moving spaces in AI

Two revelations this week have shaken the narrative in AI : Nvidia’s earnings & this tweet about Gemini. The AI industry spent 2025 convinced that pre-training scaling laws had hit a wall. Models weren’t improving just from adding more compute during training. Then Gemini 3

When's the last time you evaluated your evaluator? The LLM tooling market is getting noisy, and differentiation is fading fast. How does it feel using eval tools inside a real workflow, under real constraints, with real debugging needs? During his internship, @coledumansk19

> be me, go to SF > never did a hackathon before > find the context engineering one from @Theoryvc > claude code yolo with Mixedbread > won > have a new PC

@ttunguz Open source is essential, helps innovation. Agreed.

FUTURES TRADERS: Get 40% off all evals, no activation fees, end-of-day drawdown in our live-market PRO+ accounts…and still daily PRO payouts!

70% of production AI teams use open source models. 72.5% connect agents to databases, not chat interfaces. This is what 375 technical builders actually ship - & it looks nothing like Twitter AI. 70% of teams use open source models in some capacity. 48% describe their strategy as

"Our business, which today is an end-user tools business, will become essentially an infrastructure business in support of agents doing work." — Satya Nadella Microsoft talking about their view where the value accrues : agent infrastructure https://t.co/1TaxJpOeCB

Thanks to everyone who came out to @Theoryvc's America's Next Top Modeler yesterday! We had a great turnout, not to mention the 70+ submissions after the event ended 👀.

"Context is the Defensible Moat: A generic LLM is a commodity; your business context is what makes your application defensible, because domain knowledge is the most valuable asset." - Yuanzheng Zhu, Instacart Couldn't agree more. Customizing AI systems with the internal

It’s getting started! #antmTheoryHack Hosted by @Theoryvc

What took a 3000-node Dremel cluster can now run on a single DuckDB instance. Modern hardware has fundamentally changed the equation. During his talk at this year's Big Data LDN, MotherDuck CEO Jordan Tigani demoed how a single large DuckDB instance on MotherDuck can run

We're hiring our first product designer at @AlliumLabs! If you love turning complex, data-rich systems into clear, intuitive experiences, come help us build our end-to-end UX and design system from the ground up. Shoot me a dm if you want to learn more.

📣 Data Plumbers of New York! come visit us in NYC and we’ll hook you up with a hat! (or if you’re a former data plumber who now uses Allium 🪠) featuring our goated eng Shank, we are always hiring!

Every week brings a new agent framework or model. It's exciting, but founders need to know: what’s actually being used in production? We asked 413 senior builders how they’re adopting AI today, where adoption breaks, and where the real gaps are. Interactive findings ↓

Ten months ago, DeepSeek collapsed AI training costs by 90% using distillation - transferring knowledge from larger models to smaller ones at a fraction of the cost. Distillation works like a tutor training a student : a large model teaches a smaller one. As we’ve shifted from