One claim at a time.
Published before it’s finished.

Working Theory.

A good chunk of what lands in my notes isn’t ready for print. Not wrong, exactly – half-formed. Observations I keep mulling over, theories I can’t quite close, ideas that need someone to push on them before they’d survive a Tuesday radical Briefing deep dive.

Working Theory is where those go. One claim per piece, stated plainly enough that I can be publicly wrong about it. No schedule. It publishes when a claim is ready to be argued with.

The reason is a lesson Jane and I took out of Mozilla: thinking in the open, with people who will actually argue with you, beats thinking alone every single time. So when one of these is wrong, it gets revised here – in public, on the piece, with the date attached. That’s the part that makes it worth doing.


#002 — August 25, 2026

Cheap Intelligence, Expensive Companies

Among companies that have actually put AI into operations, through roughly the end of 2027, the cost of turning AI into operating advantage will be hourglass-shaped in company size – cheap for the very small, affordable for the very large, and worst for firms in the middle (call it fifty to a thousand employees).

#001 — August 11, 2026

revised

The Bill for Agentic AI Lands on You, Not the Labs

The first serious legal exposure from agentic AI is going to land on individual users, not on the AI labs or the makers of AI harnesses – and it will land there precisely because the labs are the only parties equipped to survive it.


The Briefing.

Twice a week. ~10,000 readers.

Two short essays a week on what’s breaking, what’s rebuilding, and what to do about it. Read by 10,000+ founders, operators, and leaders.

Free · No spam · Unsubscribe anytime.