GSoC has always had projects targeting third parties in the past, i don't
see any issue with that as long as they are reasonable projects.

If anything, i think the third party gsoc projects have a generally better
track record than the Wikimedia ones.

--
Bawolff

On Tuesday, 3 February 2026, Sohom Datta via Wikitech-l <
[email protected]> wrote:

> I genuinely hope there has been some misunderstanding here. While I'm all
> for "let's use AI consciously and not put AI on Wikipedia," I do not see
> how using it in a third-party instance that is likely never going to be
> deployed on any Wikimedia project is relevant to the WMF especially in the
> context of Google Summer of Code, where the funding for the project is not
> even coming from the WMF to start with.
>
> Even putting aside the AI issue, I'm also personally not currently a very
> big fan of how this year's event is playing out. The docs were unclear,
> and the "we need microtasks before submission" requirement was not clearly
> communicated. More importantly, the previous year's events had been "chill"
> in that there was never a competition/selection process to get projects in
> time, and whenever we fell short of the requirements as mentors, we were
> given gentle nudges to fix them, not entirely removed from the program.
>
> Regards,
> Sohom Datta
> ---
> Open-source contributor @Wikimedia
>
>
> On Mon, Feb 2, 2026 at 7:31 PM Pine W via Wikitech-l <
> [email protected]> wrote:
>
>> Speaking in general (emphasis: I have not done a deep dive into the
>> information presented in this specific scenario), I'm not a fan of
>> requirements being sprung onto folks (this would include being sprung onto
>> WMF staff) who have designed a project according to specifications that
>> they thought they understood, and then after time has been spent to design
>> a project according to the known specifications, the goalposts are moved,
>> an undocumented expectation is cited as a reason for rejecting a proposal,
>> or a new requirement is implemented, especially if the people planning the
>> project are not given a window of opportunity to align with the revised
>> goalposts or the new understanding of requirements. In general, I believe
>> that surprises during review processes should be minimized, especially
>> after an opportunity for revisions has closed. And I would hope that, if
>> such surprises happen, there will be deep dives into the review procedures
>> to understand what happened and why.
>>
>> Thanks,
>> Pine🌲
>>
>> On Mon, Feb 2, 2026 at 8:29 AM Yaron Koren via Wikitech-l <
>> [email protected]> wrote:
>>
>>> I've been mentoring Google Summer of Code projects for the Wikimedia
>>> Foundation for over 15 years now, and in all that time, I don't recall a
>>> technically viable project suggestion being rejected by the WMF. This year,
>>> six were. [1] Three were rejected for technical reasons (a few hours past
>>> the deadline, no microtasks yet) - which seems harsh, but those are the
>>> rules now, I suppose. That still leaves the following three rejected
>>> projects:
>>>
>>>  - Agentic editing capability for Wanda [2]
>>>  - Improve Commons Android app using privacy-friendly edge AI [3]
>>>  - Querying of structured data for Wanda extension [4]
>>>
>>> Wanda is a MediaWiki extension that provides an AI chatbot, so all of
>>> these are AI-related. I'm involved with the first and third one; the third
>>> one was a last-minute substitute after the first one was rejected - we
>>> thought that perhaps the issue with the first one was that it
>>> related specifically to AI *editing*, so we switched to just AI querying
>>> instead. To no avail, though.
>>>
>>> (And yes, the third one was ostensibly rejected not because it was AI
>>> but because I was a listed mentor for two different projects - but I
>>> offered to replace myself with someone else, both before [5] and after the
>>> rejection, and got no response, so I'm guessing that was not the real
>>> reason.)
>>>
>>> The Wikimedia Foundation is free to set any rules it wants about which
>>> projects to accept and reject. However, if the rule is now that no AI-based
>>> project will be accepted, I think that should be stated publicly, to avoid
>>> wasting potential mentors' time. And it should be clarified whether this
>>> policy applies just to the Google Summer of Code, or also to other
>>> mentorship programs, or even other things like development grants.
>>>
>>> [1] https://phabricator.wikimedia.org/project/board/8423/query/all/
>>> [2] https://phabricator.wikimedia.org/T414281
>>> [3] https://phabricator.wikimedia.org/T414881
>>> [4] https://phabricator.wikimedia.org/T415465
>>> [5] https://phabricator.wikimedia.org/T414617#11523658
>>>
>>> -Yaron
>>>
>>> --
>>> WikiWorks · MediaWiki Consulting · http://wikiworks.com
>>> _______________________________________________
>>> Wikitech-l mailing list -- [email protected]
>>> To unsubscribe send an email to [email protected]
>>> https://lists.wikimedia.org/postorius/lists/wikitech-l.
>>> lists.wikimedia.org/
>>
>> _______________________________________________
>> Wikitech-l mailing list -- [email protected]
>> To unsubscribe send an email to [email protected]
>> https://lists.wikimedia.org/postorius/lists/wikitech-l.
>> lists.wikimedia.org/
>
>
_______________________________________________
Wikitech-l mailing list -- [email protected]
To unsubscribe send an email to [email protected]
https://lists.wikimedia.org/postorius/lists/wikitech-l.lists.wikimedia.org/

Reply via email to