[ Blog ]

Ideas, shipped.
Not just written.

Notes on engineering, product, and design from the team that builds the case studies you just read.

Why a good MVP still takes 10 weeks, not 10 days

AI tools compressed the wrong part of the timeline. Here's what actually determines how fast you can responsibly ship a first version.

A checkout flow can be perfectly usable and still feel untrustworthy. The difference isn't in the UX. It's in what the interface chooses to show you.

Most AI features die within a month of launch. The ones that survive share a pattern that has nothing to do with model quality.

Sometimes the right technical answer is the one that costs you the contract. This is one of those times, and what happened after.

It's not the debt itself that kills velocity. It's the compounding decision fatigue of working around it every single day.

Most of what we got wrong in early hiring wasn't about skill. It was about what we optimized for in the interview itself.