oh well, it was good while it lasted
post
Damn⦠these were some of the OSS projects I was the most excited about. uv in particular is brilliant. Hope it breaks out again once OpenAI collapses.
Not a fan of this news. At all.
oh no
The tools are open source. Someone needs to fork them before it's too late.
This does suck, but there's time. I doubt the actual legal hand over has happened yet, and OpenAI is (likely, I don't actually know) an absolute mess, HR/structure-wise, internally. If the aquisitions I've been a part of in the past are any indication, there's a 1-2 year ramp before things really start to go to shit within Astral (assuming there isn't an immediate massive exodous of talent). Or I'm totally wrong!
Is there any example of such move that does not ends with enshittyfication?
Well fuck. I was eyeing uv, but I guess um sticking with pip.
Don't gaslight yourself. uv is open source and will not kill your grandmother. They just happon to be funded via Openai for now under their umbrella.
It's not gaslighting yourself to be concerned when a company as morally bankrupt as OpenAI becomes the steward of these tools.
Yeah it is gaslighting yourself. I get Openai isn't great, but uv is open source.
Explain to me in great detail what exactly you think is going to happen here.
Also Bun, which was acquired by anthropic, is doing just fine as an open source project.
OpenAI could direct anyone working at Astral to insert whatever they want into uv, ruff or ty, without any regard to the needs or desires of the wider community. Or just drop working on it entirely.
Sure, these tools are open source and can be forked. But to do that successfully you'll need to replace everything that Astral currently provides these projects: expertise of core contributors, the time they have to actually work on it, and organisational skills to manage these projects.
This isn't something you can do overnight.
OpenAI could direct anyone working at Astral to insert whatever they want
This would be suicidal for the project if it ignores the community. the project would instantly risk forking and OpenAI losing the ability to influence the ecosystem direction.
to your second point on needing contributors and infrastructor after a fork, while this is true, uv specifically has seen MASSIVE adoption both for hobbyists AND corporate interests. it has greatly simplified the overall python packaging experience. I would not be shocked if other corps began supporting a community fork if the main project went off the rails.
Bun, another open source project, has been doing just fine under Anthropic nearly a year later. Until there is more concrete info on the direction this will take, the fear is MASSIVELY overblown. uv is quickly starting to hit critical mass adoption and may be too important for it to end up in a dumpster.
but why would the vibe-coding future that Sam wants to shove up our arses need these tools?
Fuck shit asshole what thr fuck motherfucker why??????? I was starting to like uv :'(
Have zero exposure to Astral. If they were bought out by OpenAI tomorrow, would not notice the difference.
If i were a Rust coder, would have interest. Since i'm not, would be at the mercy of the maintainers.
Saw uv and ruff as a liability. Now? What is uv and ruff. Never heard of it.
Look forward to future discussions on requirement files again. Reintroduce everyone to the -r and -c options. How to put requirements into a hierarchy. How to deal with multiple venv. And how to manage requirement files across multiple packages.
i highly recommend this excellent thread.
uv has a CLAUDE.md since January. Sooo, back to pip it is?
There is poetry for package management. Apparently uv is substantially faster at solving package dependencies although poetry is more feature rich. (I've only used poetry, so I know it is adequate, but I have had times I've sat there for minutes or even tens of minutes while it worked through installing all the right versions of all the right libraries.)
yeah the dep resolution and instance management being snappy as hell are big reasons why people went for uv, but I suspect this will now get reevaluated (it shouldβve been earlier)
pdm is what you want afaik btw, poetry is outclassed for a while now
(and that I even had to type that second bit is a big bit of my gripe with python in the last decade)
Is switching to pdm as easy as installing it, then initializing pdm and remaking the pyproject.toml by adding all the dependencies with pdm? (If I'm understanding right, this basically seems like the same workflow as poetry?)
If it's any consolation, Astral was apparently already funded by our favorite Silicon Valley crypto-fascist and egghead cosplayer, Marc Andressen.
From https://astral.sh/blog/openai
Second, to our investors, especially Casey Aylward from Accel, who led our Seed and Series A, and Jennifer Li from Andreessen Horowitz, who led our Series B. As a first-time, technical, solo founder, you showed far more belief in me than I ever showed in myself, and I will never forget that.
Hmm maybe give pixi a try. Though not all packages are on conda and I think for pypi it also just uses uv.
all 43 comments