Thinking in public.
Notes on startups, products, AI, hardware and things I learned by getting them wrong. Some of these are drafts. Some are still just titles I'm circling.
Why technically strong products still fail
A working system is not the same as a wanted one. Notes on the gap between building something good and building something people reach for.
Building AI for the physical world
What changes when a model has to perceive an environment instead of a text box: latency, sensors, and the tyranny of real light.
What selling to schools taught me about distribution
Procurement, trust and accountability decide institutional adoption long before the product does.
Distribution is part of the product
Why I now treat how a thing spreads as a design problem, not a phase that starts after launch.
Things I've changed my mind about while building
A running list. Building has been the fastest way I've found to be wrong quickly and cheaply.
The case for intelligence beyond screens
Why the next interesting computing systems may not live inside another app or browser tab.
Essays are being written as the ideas earn it. Check back, or reach out if you want to talk about any of them before they're published.