Ideas on venture studios, AI leverage, and the messy reality of building companies from scratch.
The AI correction is coming. Venture studios were designed for exactly this moment -- disciplined iteration, lean teams, and the ability to commit fully when the signal is clear.
Read article
YC's Spring 2026 RFS calls for AI-powered government fraud investigation. The same detection capabilities built for insurance extend naturally to the largest fraud market on earth.
Read article
2025 was about compute constraints. 2026 is about data. The industries sitting on the most valuable unstructured data are the hardest to crack -- and that is where the opportunity lives.
Read article
a16z says the next billion-dollar AI companies will be built on reliability, not smarter models. We agree -- and we have the scars to prove it.
Read article
Three leading-edge papers from H2 2025 that shape how we build. DINOv3, DeepSeek-V3.2, and why small models beat large ones at agentic tasks.
Read article
When building is fast and cheap, knowing what to build becomes the only advantage.
Read article
Most ideas will not matter. The discipline is testing fast enough to find the ones that do -- then going all in.
Read article
Not AI as a feature. AI as the foundation -- in how we research, build, validate, and grow.
Read article
Stop treating AI agents like junior employees. The teams getting 10x output are designing systems, not assigning tasks.
Read article
The hardest skill in a venture studio isn't building -- it's knowing when to stop. How we run stage gates that actually work.
Read article
Our two cents on where AI is heading next. Contrarian takes on what matters and what's noise in the current wave.
Read PDF
Single-shot founding is a terrible way to find product-market fit. We built a studio because the maths demanded it -- and AI made the leverage irresistible.
Read article