Post

What are the reasons people think OSS will not survive AI?

I've seen two arguments: (1) With code being so cheap to produce, people will recreate things instead of building on top of community infrastructure. (2) High volume of slop AI PRs cause OSS to be more locked down as far as who can contribute and cause ecosystem breakdown that way.

1. bespoke bad copies of good software predate slop. the bar is lower, but the pain will be the same. it didn't stop oss then, i don't see it doing so now 2. oss was already underwater in bugs and feature asks. at least these come with maybe fixes, & ai triage might help separate wheat from chaff.

I’m not worried about (1) for projects with ecosystems. Small libraries and apps may be affected. (2) does worry me. We’ll likely go through the same cycle schools face with LLM cheating. There’ll be a way to measure where a PR sits on the human–vibe-coded scale, and maintainers will set their bar.