AI analysis written by someone who built the systems.

I spent 16 years building technology, the last 11 of them specifically in AI — before AI was the thing everyone claimed to have always been working on.

At Google, I worked on large-scale ML systems at a time when transformer architectures were still research papers. At CoreWeave, I saw the infrastructure side: what it actually costs to train and serve models at scale, which hyperscalers are serious about their AI buildouts, and where the real bottlenecks are.

I started AI Briefly because I kept getting the same question from investor friends: "What should I actually make of this?" — about an earnings call, a new model release, an infrastructure announcement. The information was out there, but translating it into something actionable required a background most investors don't have.

This newsletter is my attempt to close that gap — not by simplifying, but by providing the practitioner context that makes the complexity legible.

Three formats, each serving a different purpose:

• Daily Digest —The five most consequential AI stories, each with the context you need to understand what it means. Sent every weekday morning.
• Deep Dives —Extended analysis of structural trends: infrastructure economics, software stack value capture, competitive dynamics. The pieces you'd share with your IC.
• Paper Trail —Important AI research papers translated into investment implications. No math, but no dumbing down either.

AI Briefly is written for people making consequential decisions about AI: private equity investors evaluating AI companies, underwriters modeling AI-related risks, due diligence analysts who need to go beyond the pitch deck, and public market investors who want to understand what actually drives AI company performance.

It's not for people looking for stock tips or hype. It's for people who want to understand the technology well enough to have an informed view.

Three rules:

1. No hype, no FUD. The AI space generates extraordinary amounts of both. I try to be accurate about what technology can and can't do today, and honest about uncertainty when the future is genuinely unclear.
2. Practitioner context always. The goal isn't to explain AI to a general audience — it's to give investors the engineering context they need to evaluate claims and compare options.
3. Investment implications, not investment advice. AI Briefly is not a financial advisor and does not make buy/sell recommendations. Everything is for informational purposes only.