Most startups over-engineer
Seed-stage startup. Two founders. Three users. Architecture: 6 microservices, a message queue, a service mesh, Kubernetes, and a CI/CD pipeline that takes 20 minutes to run.
You don't need that. Postgres + Vercel + a single Next.js app covers 90% of early-stage use cases. One database. One deployment. One codebase to debug at 2 AM.
Microservices solve organizational problems, not technical ones. They make sense when you have 10 teams that need to deploy independently. They don't make sense when you have 2 engineers who share a Slack channel.
The best architecture for a startup is the one that lets you ship a feature in hours, not days. Usually that's a monolith with good separation of concerns. You can always split it later when the organizational pain justifies the operational complexity.
Ship the simplest thing that works. Optimize when you have data, not anxiety.
More in Thoughts
Remote work is a skill
Working across Japan, USA, and India taught me that async communication is everything.
Model size is a vanity metric
Ran Kokoro-82M on my MacBook for TTS. 337MB. Broadcast-quality speech. Zero cost.
AI won't replace engineers
It'll replace engineers who can't use AI. Different thing.