Software is easy to start and hard to finish. I finish.

I'm CTO and co-founder of Blank. We build AI and full-stack products with ambitious companies - teams that need the thing to work in production, in front of real users and real money. The latest is Totebot, an AI agent platform that lets online stores deploy shopping assistants that sell and support customers around the clock.

My view is simple. The model writes the code now. What's scarce is judgment - deciding what to build, knowing what good looks like, and having the nerve to refuse to ship the thing that isn't ready.

So I'm skeptical of anything that asks you to stop thinking. If the problem is genuinely hard and it matters, that's the work I want.