Look at the timeline and it reads like restlessness: Rails, then React Native, then Web3, then agents. Four different worlds in eighteen years.
But it isn’t restlessness. It’s the same move, four times.
- 2005: Rails. Early adopter in Czechia when the ecosystem was three blog posts and a prayer. The lesson: shipping beats waiting for the docs.
- 2015: React Native. Cross-platform before it was obvious. I shipped OSS bridges; one hit 121 stars. Distribution taught me more than the framework did.
- 2022: Web3. Four years deep on a Solana protocol. Rust on-chain, SDKs, infra, then agents. Not a tourist.
- 2025: Agents. The whole studio turned. Now I build the plumbing while the interface layer sorts itself out.
What it costs
Catching a wave early is not free. You pay in missing tooling, in explaining yourself, in things that break because nobody’s hit them yet. Half of what you build gets shelved when the wave doesn’t land the way you thought.
Why keep doing it
Because the appetite is the point. I’d rather explore ten things than perfect one. The early, awkward version of a technology teaches you things the polished version hides. And occasionally you’re standing in the right place when it becomes obvious to everyone else.