Ask HN: What's your uv exit strategy?

I'm migrating more projects to uv because it's genuinely excellent - faster than anything else, solves real problems (even edge cases like venv + wxPython + macOS), just works.

But I'm feeling the lock-in accumulate. Each project adds uv-specific configs, CI assumes uv behavior, team gets used to uv workflows. The GitHub Action is convenient, so we use it. The resolver is better, so we depend on it.

We've watched this movie before. Great developer tool becomes indispensable, then business realities hit. Even Google dropped "don't be evil." The enshittification pattern is well-documented: be good to users until they're locked in, then squeeze. Not saying Astral will - they seem genuinely focused on developer experience.

But that's what everyone says in the first few years.

What's your approach here? Are you building abstraction layers? Keeping alternative workflows tested? Just accepting that you'll deal with migration if/when needed?

I keep adopting uv because it's the right technical choice, but I'm uneasy about having no real fallback plan if things change direction in 2-3 years. The better the tool, the deeper the eventual lock-in.