If Codex’s core quality is anything to go by, it’s time to create a community fork of UV

Maybe they are being acquired to improve the quality of Codex.

That's the thing. To me that says that as soon as cash becomes tight at OpenAI, the Astral staff will no longer get to work on Python tooling anymore, namely uv, etc.

Tale as old as time in SV, why we keep trusting venture capital to be the community's stewards I have no idea.

We need public investment in open source, in the form of grants, not more private partnerships that somehow always seem to hurt the community.

> why we keep trusting venture capital to be the community's stewards I have no idea.

They bought the trust.

> we keep trusting venture capital to be the community's stewards

OpenAI isn't a VC. It's VC-backed. But so is Astral.

what do you mean "trusting" or "hurting the community"? i don't think uv has damaged anything yet. i'll use a tool from whoever if the risk profile is acceptable. given the level of quality in uv already, it seems very low risk to adopt no matter who the authors are, because it's open source, easy to use old version, and if they really go off the deep end, i expect the python community as a whole will maintain a slow-moving but stable fork.

i'd love there to be infinite public free money we could spend on Making Good Software but at least in the US, there is vanishingly small public free money available, while there's huge sums of private free money even in post-ZIRP era. If some VCs want to fund a team to write great open source software the rest of us get for free, i say "great thanks!"

At least it’s in rust.

Unlike those react-game-engine guys over at Claude

The priorities of the tooling will change to help agents instead of human users directly. That's all that's happening.

Eh, if it turns out to be too bad I guess I’ll just end up switching back to pipenv, which is the closest thing to uv (especially due to the automatic Python version management, but not as fast).

Does pipenv download and install prebuilt interpreters when managing Python versions? Last I used it it relied on pyenv to do a local build, which is incredibly finicky on heterogenous fleets of computers.

I would much rather use pipenv, if it only had the speed of uv.

Every interface kenneth reitz originally designed was fantastic to learn and use. I wish the influx of all these non-pythonistas changing the language over the last 10 years or so would go back and learn from his stuff.

Pipenv is a pile of shite

so’s your face

People would just make pipenv fast? There are some new tools that can help with that..