GIMP at LGM2026

Libre Arts
1

Yesterday at LGM2026, GIMP’s maintainer Jehan Pagès gave a talk about GIMP as a community project and made several interesting announcements.

This post is going to be quite a bit critical, so before I proceed, here is the obligatory disclaimer: I used to be a GIMP contributor, significantly decreased my participation in late 2021, and officially bowed out in late 2022.

I’m omitting a lot of things I found less interesting to discuss. You can watch the whole thing here and make your own judgments.

Past organization attempts

During the talk, Jehan mentioned that the project first considered creating some form of a non-profit in 1999 but never acted on that idea. Then Jehan himself nearly created a non-profit for GIMP in 2023, but instead went on to sign a fiscal sponsorship agreement with GNOME Foundation in June 2024.

Actually, scratch that. He did not exactly mention those last two bits of info: they were listed on the slide but weren’t discussed at all.

There’s no explanation why Jehan decided against creating the non-profit. Neither is there an explanation what the fiscal sponsorship agreement means.

I do know one thing the agreement involved — transferring all BTC to GNOME Foundation for GNOME Foundation to immediately sell it and put the fiat money on GIMP’s account (they’ve been managing GIMP’s money since dawn of times). But I know it from tracking the transactions and reading one contributor’s comments on Reddit, not from any sort of official announcement, because the only official announcement was very inspecific and hidden in release notes of v3.0RC1.

By my rough calculations from a year ago, there should be around 2 mil USD on that account after selling all BTC. So what have they been doing with that since early last year?

Committee

First off, in 2024, the team created a formal committee. It’s currently comprised of 9 people:

Let’s break it down.

  • Akkana Peck created some really useful plugins for GIMP back in the day and wrote several extremely good books for beginner GIMP users.
  • Aryeom Han is a digital artist and animator, she’s 1/2 of Zemarmot animation project that relies on GIMP for production.
  • Jehan Pagès has been the most active developer in the past decade or so and the current GIMP maintainer since 2021.
  • Liam Quin has been a great help in supporting users for over two decades now and, since my departure, managing the social media presence (at least partially).
  • Michael Natterer is former project maintainer and former lead developer. His involvement severely decreased in the past years after he had to take over family business.
  • Michael Schumacher is the guy who has been managing funds on GIMP’s end for a very long time and sorting out various organizational things like LGM participation.
  • Pat David was an active GIMP educator in 2010s, created the current GIMP’s website, and has been maintaining it ever since. He’s been keeping a relatively low profile in the project, you probably know him more as the creator of Pixls.us.
  • Simon Budig used to be one of the most active contributors in the past (see this interview from November 2025). He’s been mostly patching things here and there every once in a while for the past 10-15 years and mainly doing user support.
  • Ville Pätsi is another former active developer. To the best of my knowledge, he’s still around but mostly in the user support role.

So that should give you some idea who the committee members are. Personally, I think the committee’s composition is alright. But what do they do?

Jehan specifically said:

The GIMP committee is only for the funds. It was very important to me that it’s not about leading the software, it’s about managing the funds to support the software. We don’t decide what features go in, it’s not our role and it should not be.

I’ve seen projects like FreeCAD adopt this kind of role separation and still get a lot of heat from a vocal minority, so I can perfectly understand this sentiment. (Full disclosure: in case of FreeCAD, I’m very much biased.) So yes, this planned detachment from making software decisions makes sense to me personally and, hopefully, to you too.

The committee actually has a dedicated project on GNOME’s Gitlab server where they explain a few things. They even use the issue tracker for decision-making in the open.

The caveat is that while the committee has been around since approx. June 2024, it was never properly introduced. It got two minor mentions on the main website in release notes (v3.0RC2 and v3.2.2), and two mentions on developer portal (1, 2). Imagine a committee managing approx. $2,000,0000 of a FOSS project’s money and being nothing but a footnote on that very project’s website.

So, about the things they do…

Grants

One of the committee’s decisions was to try issuing grants. So far, there have been two test grants issued to Jehan himself and to Øyvind Kolås, the lead GEGL developer of 20+ years.

Jehan did not specify what the grants were issued for, how much any of them was, whether there was any formal proposal or agreement or any grant review and approval process whatsoever.

Incidentally, there is a description of the grant awarding procedure. I found it after asking someone who knew where to look. You now know about it too, likely after reading it here in an arcane blog post at the edge of interwebz.

Jehan also mentioned that they are waiting for something to be unblocked from the administrative side to continue with the second round of grants. The slide specifically says “Currently waiting to be allowed to continue”.

Taken literally, this could mean that the committee doesn’t have full control over the funds and depends on the GNOME Foundation to take action.

Again, there is no public information about the relationship between GIMP and GNOME Foundation. We are on the speculation territory here.

No-AI policy

Jehan mentioned that the team has adopted a strict No-AI policy. This is a principle thing, and as you probably know from this study, FOSS projects are all over the spectrum on the use of AI.

To the best of my knowledge, the formal part of this policy is only mentioned in a Gitlab template updated in March 2025 and on the Internship programs page at the developers portal.

The GSoC thing

During the talk, Jehan mentioned that the project’s participation has been very successful, and 9 out of 10 students from 2022-2025 program installments sticked with the project. He went on to attribute this success to the team looking at people more than at proposals.

This is a mostly sensible approach, except his data is all wrong. Let me explain.

The project indeed had 10 slots from GSoC between 2022 and 2025. However they were filled by 8 different people: both CmykStudent and Idriss Fekir participated twice.

Here is a quick breakdown.

  • CmykStudent successfully completed both 2022 and 2023 projects and went on to become GSoC mentor in the organization in 2024 and is still a very active contributor. Most recently, he’s been working on the CMYK color mode and a shape drawing tool.

  • Idriss Fekir did two good GSoC projects dedicated to the text tool, his involvement extended somewhat after the 2024 project, but he hasn’t been active since October 2025. He had 8 commits to the project over the last 12 months.

  • Shubham did one decent project in 2023, layer autoexpansion support, and stopped at that.

  • Cheesequake worked on modernizing the GtkTreeView widget and continued contributing past his GSoC project until about 7 months ago.

  • Varun Samaga B L made 15 commits to GEGL as part of his GSoC project in 2024 and dropped out.

  • Shivam Shekhar Soy completed his part of working on the new extensions website last year and left the project after the program’s end.

  • Gabriele Barbero worked on the text tool and continues contributing. His latest merged commit from March this year fixed the text tool’s overlay positioning when the view is rotated.

  • Ondřej Míchal implemented GEGL filter explorer and keeps contributing. His most recent patches touch widgets, flatpak builds, and controllers UI in Preferences.

To sum it up, the statement that 9 out of 10 people sticked with the project is incorrect. There were 8 GSoC students between 2022 and 2025. 5 people continued contributing past their respective GSoC projects, 3 of them are currently active.

That is still a really good result, plus, as far as I can tell, all 10 GSoC projects in the past years have been successful. I have no idea why this exaggeration was needed when they had a perfectly solid case already.

Summing it up

So what have we learned from this LGM talk and this quick research?

  • The project has had a committee since at least the first half of 2024. It only makes decisions on spending the available funds.
  • GIMP has been in a fiscal sponsorship agreement with GNOME Foundation at least since June 2024, with undisclosed terms.
  • The team has been trying to start actually spending money on development, but all we know about it is that two active developers got the grants.
  • The second round of grants is blocked on the administrative side, whoever they are, for whatever reason.
  • There is no formal explanation what the grant program covers, who is eligible, what the program’s budget is, etc. But there is a short formal description of the process, and you can apply for a grant when the grant program resumes.
  • The project has had a strict No-AI policy since at least March 2025.
  • You have to pay attention to GIMP’s announcements of bugfix releases and release candidates. This is where they hide important news announcements about changes in the project management.

In conclusion

I hope that while I’m very salty about this talk, you will appreciate that I’m also doing my best to stay objective.

Despite the unfortunate GSoC blunder during the talk and the continuous lack of transparency, the team is finally making the right organizational changes, if at a glacial pace. Let’s hope to see more in the coming months/years.

This is a companion discussion topic for the original entry at https://librearts.org/2026/04/gimp-at-lgm2026/
3

I think I can provide some clarification on this point, since it turns out that I was involved in at least one of Jehan’s attempts to establish a foundation (it was in 2023-2024).

I am a researcher at the CNRS (the largest research institute in France), and Jehan contacted me to see if it was possible to create a foundation hosted by the CNRS (which has its own foundation and therefore offers to host “sub-foundations,” for example for open-source projects, and to provide appropriate management services).
To apply for this type of foundation creation, the project(s) involved must have a clear link to research activity, and given the software I develop (G’MIC) and my position in a CNRS research lab, we were indeed able to consider setting up a foundation.

To make a long story short, things didn’t go very well between Jehan and me, because we had very different visions of what such a foundation should aim to achieve. For my part, I wanted the foundation to focus on free software for image processing and graphic design, with the idea of later incorporating other free software projects in the same field, who might need to be represented by a structure like a foundation (for instance, to ease the management of donations).

As for Jehan, I got the impression that he wanted it to be more of a “GIMP Foundation” that would be focused on GIMP but ready to help smaller projects (including G’MIC).
The foundation’s board he proposed was more or less similar to the one mentioned in @prokoudine 's article (plus me), and so it was clearly “GIMP-centric,” which didn’t sit well with me at all.

Another major point of disagreement centered on issues related to AI: As a researcher in image processing and analysis, I was clearly opposed to any negative discourse surrounding AI: From an algorithmic standpoint, these are currently the most effective methods and the ones most widely used in research today. If you’re up to date, you can’t just do without it.

It’s true that these days, people tend to confuse the algorithms behind AI (primarily how neural networks are trained and inferred) with the data theft we all know about, which is carried out by the GAFAM companies to build their huge training datasets.
And from the emails with members of the board we exchanged, I got the distinct impression that this confusion was here, and that most of the people on the GIMP development team (not just Jehan) were very critical of AI and that it was completely off-limits to even mention it or use it to quickly create prototype illustrations.

Having a foundation hosted by the CNRS, related to research and libre software for image analysis, processing and synthesis, and not being able to conduct research on AI (or at least not mention it too loudly). How can I say it…

Quite frankly, I think their messaging about AI is very simplistic: Thinking that digital image creation will be able to do without AI in the future is just shooting oneself in the foot.
This rhetoric might appeal to a few anti-AI extremists who don’t understand the difference between neural networks and training data, but it won’t hold up in the long run.
Yes, there are issues with stealing images from the Internet to train generative neural networks, but that’s mainly the GAFAM fault, and it’s clearly a distinct problem from what AI really is.

This isn’t the first time I’ve heard or read quite extreme rhetoric from some of them, so I’m not particularly surprised.
But in the end, I was actually quite relieved that this joint foundation project wasn’t successful. Given such differences in opinion, tensions within the committee would have been inevitable.

Finally, I’m happy for them if they’ve managed to reach an agreement with the GNOME Foundation to manage their budget and pay developer’s work. GIMP remains a great piece of software, and some of the developers (including Jehan) are really skilled, and deserve financial support.

Anyway, I hope they’ll have the opportunity to reconsider their stance on the use of AI in the future. It’s entirely possible to design AI systems that are useful to artists and trained in a responsible and ethical manner (In fact, that is one of my current research topics.)

3 Likes
4

Thanks, @David_Tschumperle, that’s really insightful!

The no-AI policy is more about AI-contributed patches, as far as it’s possible to figure out from very little information available. I have no first-hand knowledge if they oppose all AI.

IMO, things like background removal, scaling images up, inpainting, etc. are perfectly valid use cases for neural networks. Personally, I would have no problem if they embraced that but rejected prompt-based GenAI. I know Ell was planning to work on actually useful AI-based features, but he hasn’t been around for years.

1 Like
5

During our email exchanges regarding the creation of the foundation, I suggested the idea of a mascot for the foundation (who would not be Wilber) which was a small octopus whose tentacles each held a box featuring the logo of a different open-source graphics software/library (at least GIMP, GEGL and G’MIC to start with), with the idea that the foundation could bring all these programs together under one single entity.

To create a prototype of this mascot, I had used a generative AI (I am absolutely no artist, so I wouldn’t have been able to draw such a mascot on my own) and sent it by mail.

Jehan made it very clear to me that sending such an AI-generated image to the foundation’s mailing list was a very serious matter. From that point on, I realized it wasn’t going to be possible to work with them!

So I think their “anti-AI” policy goes a bit further than just considering AI-generated patches. Then again, given how much time has passed, they may have changed their minds a bit…

6

@prokoudine Hi! To add some more details, the challenge with the grants (as I understand) is that the GNOME Foundation has gone through several leadership changes since the agreement was made. When that happens, we become understandably a lower priority while they get settled in, and then it seems like processes have to be restarted because the prior person did not leave behind all the documentation.

Progress is happening though! For the first time (as I understand) Jehan was able to allocate money to pay for people’s LGM travel rather than wait for reimbursement, and it sounds like we’re very close to the developer grant program being signed off on.

I do a lot of the writing for the news posts now, so if you have any suggestions about how to make that kind of information more transparent (especially when we’re in the “wait and see” mode), I’d appreciate it! Another volunteer is working on a news post to show funding and allocation information, though I don’t know when that’ll be ready.

@David_Tschumperle Hello! There’s some variety of opinons on AI usage among the developers. I think we’re all generally opposed to the “use people’s work for training without permission” kind of generative AI, but other forms are okay. I work with machine learning in the research lab I’m part of, and I’m interested in incorporating more of that into GIMP - for instance, I’ve proposed working on in-painting for GIMP 3.6, and that has received a good response.

3 Likes
7

I’m glad to hear that! Things are finally starting to go in the right direction.

2 Likes
8

Sure, I’ll chat with you privately

1 Like
9

Recently i’ve increased the posting i’ve been doing, but you’re right, we didn’t make announcements around the grants. We did post (for example) about going to #libreGraphicsMeeting.

Posting about problems with other people & organisations has to be done with a lot of care, of course, even if the problems are simply that staff changed. Regarding not allowed, it’s mostly about meeting US legal auditing requirements, which is not a big deal in that we have not done wrong (as far as anyone knows at least :slight_smile: ) but in that there are a lot of steps and it takes a long time, and as mentioned, we had to start the process again several times.

But i do think we should post more about positive things, and you will see more of that in future. There will also, i think, be more detailed financial reports, but again there’s still some stuff being sorted out, such as donations to GIMP made via GNOME. But this is a requirement for a non-profit so it will happen.

2 Likes
10

This was done once or twice in the past, but it was always a very manual and time-consuming process. Even handling reimbursements afterwards was time-consuming, and that contributed to (albeit not singularly caused) a burnout I had in 2017/2018).

As a I result, I didn’t really want much to do with organizing in the following years.

But with things are looking out now, this has turned out radically from a very manual and menial task to something that only requires managerial oversight at some crucial places. I am pretty sure we can write a short post-LGM report about that.

2 Likes