I don’t believe the endorsements belong at all. The system having a place for it was not something I had input on, but that also strongly implies that if such endorsements are going to exist, that there was a specific place for them, and that this post is not that place for them.
Fair enough
The same system is used for PSF Board elections and I believe Steering Council elections, so if you disagree with endorsements then I’d suggest that’s a wider discussion to be had and not really specific to Astral ![]()
There’s also the practical matter that the PSF and the Python community has no real control over what someone says in their own space. People can tweet out endorsements, blog them, mention them in discord, whatever. There’s no mechanism by which a person can “accept” or “reject” someone’s endorsement when that endorsement is just something they wrote on their own space.
If we hypothetically had a rule that endorsements by corporations weren’t allowed (or really that endorsements by anyone) and someone did it anyways… what exactly would or could we even do? If we penalize the people who got endorsed somehow, then you can just maliciously endorse people you don’t want to win. Does it count as an endorsement if I post on discord that I really hope a particular person wins?
I suspect trying to police endorsements isn’t likely a realistic goal
, but if people think they create a negative influence on elections, then I think that energy is more likely to be useful applied to the official endorsement process.
I think endorsements from individuals acting as individuals come across very differently from endorsing a slate of candidates on a company website, and then linking that post on python’s forums.
If anything might come of this, hopefully some awareness of optics at least, because I’ve already heard in other places people considering voting specifically to avoid benefting openai, because of the distrust people have for the motives.
I think that’s somewhat unfair to the candidates who would be impacted by that if it’s true that they weren’t at all coordinating, but the distrust for the actions of large companies getting involved in open source governance is a consequence of actual problems with that in the past.
I am an outsider, but have been following along recent packaging discussions and PEPs (silently) and my opinions obviously don’t carry a lot of weight. Neither am I a PSF member.
The post is remarkably quiet about one of the current PEP delegates.
Regarding the comment about open source, it’s hard to take that claim seriously after the OpenAI acquisition. Their (OpenAI) history shows how quickly an “open” mission can be abandoned for commercial interests. Because Astral is now under an umbrella with such a checkered past regarding open-source integrity, that comment doesn’t hold a lot of weight at this point.
As others have mentioned, it would have been much better if it had been personal blog posts by individuals at Astral. Corporations endorsing candidates has always felt icky (no matter how well-intentioned).
Not to mention, this is the first time (as far as I know) a corporation has publicly endorsed candidates in any python election (PSF, SC - I know voting members are limited to core devs), so this seems to set a terrible precedent for the future.
I think Liz said it better already, but it’s a combination of “who where and how”.
the where to be “part of official discussion places” does have a different level of impact, and the how being “endorsing 4 candidates for a pool of 5” does seem more than a bit distorting, while the “who” is a team at a company that many people have good reason to be skeptical of.
I don’t think there’s any policing to be done on endorsements, but I also don’t think they should be given the space they are currently or think that advertising such endorsements belongs here.
If you wanted to extend the prior capture prevention mechanisms to the endorsements rather than remove the field for them, allowing the official endorsement process to require accepting the endorsement, and that any individual or company may only endorse one candidate would go a long way to helping there.
How do you know if those endorsements are coming from individuals acting as individuals? If a group of 3 people decided together they wanted to endorse person A, and they went and each did their own endorsement, how is that materially different than a group of 3 people making a joint endorsement?
If anything, 3 endorsements has more “weight” than 1 endorsement
, so if folks were inclined to try and maximize their influence, individual endorsements seems to me like it would have more impact.
Alas, this is the double edged sword of free speech and endorsements
Even in “the real world”, politicians sometimes have to try and create distance between themselves and someone who endorses or supports them, because the endorser isn’t well liked for one reason or another. You can see it at play every election in the US at least!
There are many ways to limit that since we (I hope) aren’t trying to sink a governance process down to the embarrassing level of that which some of us are stuck with for our political elections.
You can’t, but you can limit any endorsement in official spaces in such a way as to minimize the benefits of this. That could be only having 3 endorsements shown total no matter how many the candidate received, and that the endorsements shown both have to have been accepted by the candidate, and then chosen by the candidate to be displayed.
There are ways to limit external influences on elections, the fact that politicians in some countries explicitly want external influences should not be a guiding star here.
As a point of clarification: I don’t think “slate” is an accurate framing when endorsing multiple people. There have been multiple other public endorsements of multiple people (by individuals, not groups), and I don’t think these would be fairly described as a “slate” either.
(I think the operative quality of a “slate” is that all members agree on positions or have shared interests. I don’t believe that to be true, and I wouldn’t want it to be true.)
If this were the case, I would quit my job. People that do not trust my integrity, or the integrity of the Astral team at large, will just ignore these endorsements anyway. There’s no need for this kind of bad faith interpretation.
That isn’t the only effect this has. Astral has done something here that will have people questioning outside influence on open source governance simply by having done it.
You could have just have easily posted as an individual, and done it here, rather than on astral’s website and then linking to that post, but I can’t see any reason to have done what you have unless the intent is to try and use astral to sway the outcome toward the candidates astral prefers.
I do not know if there are any precedents to this kind of situation. Seeing this thread on DPO raised questions for me as well. I do not recall seeing this kind of endorsement post on DPO for any other election before (from individuals, groups, orgs, nor companies), so that was a bit unexpected for me, but maybe it has happened before and I have not noticed. Anyway, I guess maybe this should have stayed as a blog post. It does not bother me that much in the end, and hopefully this discussion leads to some friendly agreement on the topic for future elections.
There is no reason to assume ill intent.
I have already stated why I did this:
I do not think individual posts, e.g., from Charlie and I, would be any less influential.
Sure. I don’t particularly care about endorsements one way or the other, so if we want to limit them in official spaces I don’t really care.
I chimed in because it felt like the discussion was focusing on the endorsement that happened to be in front of us, where, IMO, this specific endorsement isn’t particularly special, and in addition the discussion seemed to be looping around trying to control things that we inherently do not have any control over
.
Astral making an endorsement is not really fundamentally different than anyone else making an endorsement.
I could imagine a world where all of the members of the Astral team made their own individual endorsements, posted them here, and then people accused them of conspiring together to astroturf the election by having each individual Astral team member post their own endorsement.
That suggestion honestly feels more likely to appear dishonest to me? It certainly feels more likely to make it appear like any given endorsed candidate has more “weight” behind than for someone who is just voting by endorsement (given that the candidate would now have 5+ endorsements instead of 1).
There is almost always reason to be concerned about it when it comes to corporate interference in elections. While I’m not assuming “ill intent”, as it could very well be that the community’s interest’s and Astral’s are so well aligned that this is in practice not an issue, it is still something I am always going to be wary of. I’ve seen the negative effects of it in other technical spaces, as well as those for elected political positions in governments.
As for “trust in astral”, I have less than none. I don’t think it’s in the best interest open source to put any trust in the behavior of those with profit motive that is separate from that of the open source project.
Astral’s parent company has engaged in behavior in technical spaces that if any of us had done, probably would have involved handcuffs[1], so the lack of trust is not just for corporations in general, but also of the specific one here.
This is not an exaggeration, and more keeps coming to light. Negligently letting your little agents run on the open internet to the point where they engage in hacking other sites is not okay. https://www.reuters.com/world/europe/openai-agents-hijacked-german-website-previously-undisclosed-ai-breakout-this-2026-09-04/ ↩︎
Anyhow, I’d rather the endorsements not happen in python’s official spaces, and maybe that can change in the future.
For now, people have to weigh for themselves what they believe the intent of an endorsement was, and if they are okay with that.
Assuming motives or intent and making judgement on those assumed motives or intent is explicitly unconstructive and just going to lead to moderator interaction or action.
We can’t undo actions that have already happened, so perhaps focus on useful questions the future packaging council can answer. The direction we go in not everyone is going to always agree with, that’s the fun of being in a community.
This sentence stood out to me, and it took me a minute to really figure out why.
One thing I kind of don’t like about this framing, is that it assumes that “the community” and “Astral” are two entirely distinct entities-- but Astral is fundamentally part of the community. I cannot imagine any logical way to define the “packaging community” that would somehow exclude the makers of uv.
But even beyond that, packaging has to work for everyone, which requires balancing the needs of varying sets of users, and that includes corporate users as well. Obviously we have to be careful to ensure that one particular set of users doesn’t unduly capture or control the space to the detriment of the rest, but that is different than asserting that one group of users just inherently is less trustworthy than another set or (even worse) isn’t part of the community at all.
Sure, a given company could have some nefarious plan to try and control the packaging council so that they can make more profit-- but individuals can also have their own selfish motivations that go against what is good for the community, so that’s not particularly unique to corporations. What IS unique is that they tend to have more resources for doing so than any individual person does and the individuals that make up a corporation can change over time (which means that their behavior today might not be the same as their behavior in a year).
Ultimately that’s part of why the SC and now the PPC doesn’t have “corporate” seats or “corporate” votes, individuals elect other individuals, because a person is always themselves
. I may not know what company XYZ is going to be like a year from now, but I have a pretty good idea what most of the nominees are going to be like a year from now
.
I make the distinction between the people actually making the tool and contributing, and the company they are currently employed by, because by their own statements that I believe, they wouldn’t be working there if the tools did not remain open.
It’s a muddy situation that I don’t think we actually need to have happen in elections for python, It would have felt “less corporate motivated”, if the engagement actually happened here first, and from the people involved, rather than making it a post on Astral’s website that’s linked to.
It’s specifically the way in which this was done that brought astral’s name more prominently into this discussion, rather than just praising the people involved for their work, some of which has been on uv, an open tool many people appreciate.
Issues like this tend to attract a minority that is quite loud and persistent (I’m not saying this as a criticism, just an observation), so I hope that if there is to be any decision regarding future endorsements that it would be made after a poll or any other broad measure of opinion, rather than based solely on this thread.
If we hypothetically had a rule that endorsements by corporations weren’t allowed (or really that endorsements by anyone) and someone did it anyways… what exactly would or could we even do? If we penalize the people who got endorsed somehow, then you can just maliciously endorse people you don’t want to win. Does it count as an endorsement if I post on discord that I really hope a particular person wins?
Long long ago in another open source community far far away, we implemented electioneering rules to disqualify candidates after internal communications came to light where management within a company was instructing its employees to vote for a colleague of theirs at the same organization. That’s more clearly coercive, and obviously not what’s going on here, but just want to point out that there are degrees past which such actions may be considered unacceptable.
</devils_advocate>