For every startup that dies from technical debt, ten more die from trying to prevent it too early.
There's a special gravity that pulls technical founders toward performance, scalability, and "doing it right." It leads to beautiful infrastructure for products no one's using. The play-by-play: founder gets an idea, spins up CI/CD, containers, Terraform, event queues, and a layered architecture with domain-driven design. Six weeks later, zero users -- but the cold-start latency is great.
Premature optimization isn't a technical issue. It's a psychological one. It feels safe to build. It feels productive to refactor. It feels justified to say "this is how we'd scale later." What it really is: avoiding the scary part of shipping something half-baked and finding out nobody cares.
A slow app with real traction is 10x more valuable than a fast one no one uses. Ship the duct tape version. Optimize when you have something worth scaling.