Article 01
Read more →Building AI Products That Don’t Fall Apart Under Real Users
Why your prototype works great in dev but collapses in production — and the system patterns that prevent it.

Thoughts on building digital products, engineering culture, and modern tech.
Why your prototype works great in dev but collapses in production — and the system patterns that prevent it.
A brutally honest breakdown of deploying agents into workflows where mistakes actually cost money.
Token streaming, batching, caching, and when you should actually switch models.
What to build first, what to ignore, and how to hit proof-of-value before you run out of energy.
Skeletons, progressive reveals, and UX tricks that hide model delays while keeping users calm.
Why retrieval performance has nothing to do with just picking the right database.