There are a ton of great nominees for the packaging council and I think it is hard to go wrong with your votes; regardless of who ends up serving, the council is going to be a dramatic improvement for Python packaging.
The Astral team (especially those of us working on uv) have created a short list of people we’re particularly excited to see serve on the council. We’re proud to endorse William Woodruff, Pradyun Gedam, Ralf Gommers, and Donald Stufft.
My feeling is that pulling your weight as an organisation that’s well-respected in an open election like this is not appropriate - especially when there is a very obvious conflict of interest. It’s gratuitous and divisive for no good reason. The people you’ve endorsed have obviously done a lot of amazing work in the packaging space and deserve their plaudits, but that doesn’t mean you should exert your influence on their behalf.
I’m annoyed that I felt compelled to comment on this at all.
I’m sorry to hear that. There is no ill intent here and our only interest is effective governance and progress in Python packaging.
We were originally just going to make it clear that William was representing the entire team, since we only chose to put forth one candidate, but we were particularly excited about a few other people and ended up wanting to say so
My primary concern here is precedent, during this, or next election, will we now get company endorsements from Meta, NVIDIA, Anaconda, etc.? Do we want that?
A secondary concern I have is branding optics. Does Astral exist as a company any more? Or is this actually OpenAI’s endorsements? Obviously “Astral” has significantly more cache among the Python packaging community than OpenAI does.
I think both concerns would have been significantly alleviated had this been a personal blog post, or at least signed off personally, say by Charlie or Zanie.
Would those endorsements sway your vote? Will making a political endorsement (and this is political!) be enough of a benefit to those organizations to be worth the cost in goodwill of needing to make an opinionated statement?
I don’t see any reason to reprimand Astral for the act of making an endorsement. That said, I do place very little value in endorsement from such for-profit organizations.
These are not OpenAI’s endorsements, these were chosen by and signed off by people on the Astral team, i.e., Charlie, William, Ee, Konstantin, and myself. We wrote the post. I felt it would be more disingenuous for us each to post something separately when we’re clearly a team of people working together.
Those endorsements might not convince people who are involved in the packaging community, as they probably have their picks already, but let’s not forget that the Council is elected by a much broader group of PSF members - who might be swayed by an endorsement by the people making the “cool” tools.
Some vague “goodwill” (which corporations, especially in the LLM industry, don’t care about at all) is a fair price to pay to get your people on the council.
Let’s assume for the sake of argument that Astral putting forth their
own candidate, and endorsing their own candidate, is reasonable. Your
team’s endorsement statement for William did clearly disclaim the
conflict of interest in endorsing your own representative, so people
reading it can surely figure out what that means (i.e. liking what
Astral does == voting for Astral to be on the council).
I suppose it’s a bit more colorful to endorse 4 out of 5 available
seats. It is difficult to see how that doesn’t constitute “choosing a
council”, even if that wasn’t your intention. You are literally telling
people “Astral believes these 4 people should be 4 out of the 5 people
on the council”.
One need not believe this was done out of the desire to subvert the
process, in order to believe that its effect is to subvert the process.
(I’m not an election nerd, so I don’t really know how one would go about
trying to protect against deck-stacking, but the current election design
appears to prioritize sufficient votes for “popular groups of people who
would also be okay to have win” over strongly felt votes for a slate of
diverse candidates. So, throwing Astral’s weight into the election to
sway people towards voting for a specific 4 candidates would, if a
statistically significant group of voters read and reacted positively
towards the Astral endorsements, result in those 4 people being
surprisingly difficult to dislodge regardless of who else those voters
also voted for.)
What’s the point of an endorsement if not to sway some people’s vote?
If you’re publishing a statement saying “I endorse A and B”, then obviously you hope that’ll encourage more people to vote A and B. This is what endorsements are for.
If on the other hand you merely want to make your own vote public, you can do that after the election. And avoid using the term “endorse”.
Thanks for the reminder. It was not obvious from the original statement, and not everyone is privy to the latest news in the wild west of generative AI and its tech barons.
Of course that is the point of an endorsement. My question is whether it will work. Will endorsements sway enough votes to be worth the backlash received for making an endorsement at all?
Thanks, that’s helpful to know, both in my concerns and evaluating how to weight the post in my own choices.
For my primary concern about setting a precedent, it only needs people to think it might work, or have the appearance of having worked, for other companies to decide to join in. There’s no requirement, or way to know, if it actually swayed anyone.
Personally, I woke up this morning thinking “I should write my own blog post endorsements”, alas I am in an exception busy period of my life and haven’t had the time I wanted to engage in this process, at such a time it’s helpful to have endorsements though.
FYI, I’m also not completely against the idea that lots of people/companies do write endorsement blogs, I think it’s worth discussing if we do want that or not. I am concerned about the amount of extra material to keep up with, and if it might cause factions.
Writing for myself, I didn’t run these thoughts by anybody.
I think it is essentially proper for Astral to have made an endorsement. You can fairly say that I have an interest in believing that (given that I’m endorsed), but (1) any reasonable individual could have already inferred that they endorse me given that I used my corporate affiliation, and (2) I believe that it is proper in general for people (and teams and companies) to issue public endorsements on matters they care about.
I think we should want a consultative electoral system; we benefit from more information about the community’s opinions, and each of us remains fully empowered to vote according to our conscience. I would welcome more endorsements of other candidates in the field from parties who have an interest in the future of Python packaging.
(On a personal note: I find it surprising that several people in this thread have implied that my agency is diminished/that my role in this election is that of a proxy for a team or large company. I would like to believe that, in my time working on Python packaging, I have never given anybody any cause to believe that I have ever put a company’s interests over the interests of this community. I would also like to believe that Python, as a community, has had enough experience over the years to understand that humans sit for elected bodies.)
I don’t think this is an accurate way to look at the matter. It is
reasonable to assume that you are affiliated with Astral because Astral
cares about (many of) the same things you do, therefore people who want
to vote for “people that care about the same things Astral cares about”
should obviously vote for someone on the Astral team. That’s a good
thing! You are a subject matter expert on the things that concern people
who are interested in the same things that Astral are subject matter
experts on.
I cannot really see anyone in this thread implying that that means you
would act against the interests of things you’re not interested in.
There’s a general concern that putting a political endorsement by people
with a very large mindshare will result in people voting by endorsement
rather than by their own beliefs about the candidates. This is true
regardless of whether one endorsed candidate is “worse” than a
not-endorsed candidate, or whether one not-endorsed candidate is “even
better than” an endorsed one.
Personally, I don’t think that endorsement is a problem either way, what
bothers me is the number of them. It is better known as “slate voting”.
I think this framing risks treating voters who find Astral’s reasoning persuasive as instruments of Astral, rather than as electors exercising their own judgment and agency.
An endorsement is information about what another person or group believes, and deciding whether or how much to trust that information is part of how voters form their own judgment.
We endorsed four independent candidates because those were the candidates we were particularly excited about. They aren’t running together, didn’t agree to a common platform, and weren’t coordinated as a ticket by us. Calling that “slate voting” seems to imply more coordination than actually exists.
I don’t think corporate endorsements belong in open source elections. If the candidates are not strong enough by their own participation in the ecosystem, then it may just be that they are not the right canidates for the ecosystem.
The packaging council will have to deal with standards that are meant to be for interoperability and should be diverse in perspective.
Proper governance already has to worry about things like capture by employees of a single corporation, and those concerned about those sorts of issues are reasonable in being concerned about astral trying to put their thumb on the scale for 4 out of 5 slots.
While I disagree with the premise that corporate endorsements are likely to have a negative effect on elections, it feels a bit absurd to frame a team of people who work full-time on open source tools for Python as corporate, nor are our endorsements in any way related to the for-profit entity that employs us.
Not every participant in the ecosystem can run for the election, that’s why endorsements exist.
FWIW, I’d find this kind of endorsement behavior equally distasteful coming from a group with no corporate affiliations. I want people to make decisions on what they think, not whoever some influential players favour.
I don’t really feel strongly one way or the other about endorsements (corporate or otherwise), I don’t really think they have much effect (note I didn’t bother to try to get any endorsements on my official nomination ).
This logic seems flawed though? Or at least inconsistently applied. The election system in play allows official endorsements that show up on the candidate page, if the standard is that candidates should stand on their own, absent any endorsements, then presumably we should eliminate the ability to have endorsements at all?
Personally I suspect seeing a well known name like “Brett Cannon” has endorsed a given candidate on the official nominee page is going to have far more impact than a blog post by some company . This effect (to whatever extent is exists) likely would end up amplified the more endorsements someone might collect.
IOW, I don’t really put much stock in the value of an endorsement, but if we as a community think they’re bad, then the ones that show up on the official nominee page that gets linked to by the election software itself seem way more likely to influence the results than some blog post.
As a note for context, this is where we would have preferred to place endorsements by individual members of our team, but the endorsement process was not clear to everyone and when we asked if they could be added after the nominations were posted, we were told no.
).