Yes, I said so specifically. Council members have to make decisions on proposals in the interest of the project. Deciding not to do a thing is reasonable. But if we start making technical decisions for political reasons, I think that would be bad for the project’s reputation.
Do you believe that machine learning is inherently unethical?
I don’t think ML (/AI) systems are inherently unethical, and if we want ethical systems to be developed, I don’t think we can reasonably expect to have any say in their development if we categorically abstain from building or supporting those systems.
If we do not make ethical and political decisions around what type of technology we develop and deploy we may well end up in a system of feudal techno-anarchy.
Politics, and the role of the Council and governance across all systems is the legitimate role of reigning in dangerous and damaging forces, and of promoting positive and constructive ones.
I have zero evidence in front of me that users are being driven away from Fedora because of AI. If you want to bring that language into discussion, please show me the metrics. I’m looking around at the distribution space.. and I’m not seeing distributions being punished by providing a coherent AI developer desktop environment. We have to take care in how we as a project uses AI yes.. but providing AI desktop to developers.. the only controversial part of that is the NV hardware enablement.
I strongly disagree.
Fedora is supposed to be out in front in the conversations… even the controversial things. We aren’t going to have influence in the conversation if we aren’t in them. What we are talking about here is a desktop environment aimed at developers, that is well within scope of Fedora’s target use case, in fact its probably Fedora’s primary use case. The only technically interesting part of this is the NV hardware enablement (which may end up being a remix regardless of user focus). Gaining philosophical distance in the space of developer desktops is how Fedora becomes irrelevant… because like it or not… the people starting their software develepment journey right now are using these tools. You’re worried about turning away existing contributors because we are making a space for new outputs. I’m desperately worried about being compelling enough to attract the type of person I was when I committed to contributing my volunteer time to this project 20+ years ago. I keep asking myself If I were 20 years old again… right now in this moment..and I was going to pick a linux desktop as my daily driver… would I pick Fedora? Would having an AI desktop offering to try inside Fedora matter? I can’t honestly say it would not.
I’m going to put my FPL hat on real tight for this next comment.
I’m not here to side-step problems. I’m here to lean into them and get to a better future. I think some people are generally concerned about the ethical use of AI and are loudly pushing back on a future they feel is coming for them that is outside of their control. Saying no to everything AI related in the places where they have agency to do so feels like a way to take a stand and take back control. I think that’s exactly the wrong thing to do.
I am genuinely concerned about the ethical use of AI. I am genuinely concerned about the hype bubble. But i don’t think this technology goes away, even after the bubble burst. The best possible fiuture I can see involves the Fedora community being part of the conversation around ethical use of this technology and that doesn’t mean we sit on the sidelines and ask downstreams to do the work we refuse to do to provide a developer oriented desktop experience because we can’t bring outselves to do it. Punting on developers is how this project becomes irrelevant. And its definitely not in line with the supporting text of the First foundation, nor inline with what I understand Fedora’s core mission to be.
The people who are going to get us to the better AI future are the people at the start of their journey and see value in the technology and Fedora needs to be influencing those people so they take the technology into an ethical direction that is most congruent with our shared ideals. The skeptics among us, myself included, aren’t the people with the ability to drag the best future out of the darkness. We at best are safeguarding the past and the present from the worst possible future. But that’s not enough. We have to make a space to be in conversation with the people who can get us to the best possible future, we have to give those people a chance to build and fail and build again, dragging the best achievable future forward. One of the leat risky ways we do that is by providing a space to build a self-sustaining developer desktop experience that they can contribute to inside the bounds of our project space and that puts us in conversation with the people who see value where we can’t and we can get to make progress towards the best possible future available to all of us.
You personally don’t have to agree to the goal of this output. This output will be successful or not based entirely on whether it attracts enough people to be long term sustainable. That doesn’t have to be you or anyone in particular. There is value in constructive skeptics that influence the direction, not by applying stop energy, but by challenging those who are attracted to contribute to make progress on the shared ethical concerns. By working on the underpinning technology for other outputs inside the bounds of the same project, you gain trust and have cultural influence in a way that you would otherwise not. And that will matter in the long term.
As the Fedora Project Leader, I am absolutely not concerned about the reputational damage to this project that comes with setting up an entirely new output attractive to developers who want to make use of Ai tools. If this were proposing re-orienting Workstation, which is a mature project output, I would be pushing back as well as a skeptic on the value of the AI stuff. That approach would be materially changing the character of an existing output that users rely on..and that’s probably not appropriate and would result in watching users drop out. This approach deliberately sets up a separate output to avoid that. This will either be attractive and grow or it will not.. based entirely on individual interest. This approach is not disruptive to existing outputs nor asking for users or volunteer contributors relying on those outputs to be coopted into putting work into an output that do not want to contribute to.
I do appreciate your sincere response. But you also somewhat missed my point:
This whole effort could still happen even if it doesn’t become an official “objective” - through the same, existing mechanisms (Change Proposals, Spin / Remix creation process, etc.). We don’t have a mechanism (other than the legal banhammer) to actually prevent somebody from working on something they’re interested in (and I do think that this is mostly a good thing). I only think that it would be better for the project not to additionally promote a technology that is - at least in its current form - deeply problematic (in a number of ways that is getting difficult to keep count of) by making something so close to it one of the project “objectives”.
I am cognizant of the fact that this might just be me being an AI Luddite - over the past few years, I have become allergic to “AI” getting attached to basically anything (from phones to floating data centres in the sky) - it has often been a good indicator of things beeing en ified.
I’m not on the Council, my voice won’t matter here in the end - so feel free to feed my feedback into the /dev/null shredder. But it didn’t feel right to say nothing when I am seriously concerned about the increasing amount of AI Kool-Aid consumption. /part
First. I don’t reach for the luddite language when talking about the concerns. We are technologists talking about the harms being done in the productization of a technology.
There are real concerns about how capitalists have coopted this technology and is driving it to market in an unsustainable way. They’ve pulled a fast one on society and we are all catching up. But we’re gonna have to articulate and promote a vision for this technology along what should have been its more natural sustainable arc from cool research into useful tool.. if capitalism wasn’t entirely off the rails.
But back to specifics…I’ve been reading up on the intent of Editions. And my reading of the intent as expressed by Matthew’s series of magazine articles in conjuction with Fedora.Next and all the referenced links in the Fedora.Next wiki page including the board proposal.. none of what I have reviewed so far talks about specific intent to promote editions in a particular way.
The intent of Editions that i read, and I’m paraphrasing a bit here, is to allow for possibly radically innovatiive experiences that target a specific use case. Editions get a access to a self-governing Workgroup structure and this structure is probably makes sense out of the existing governance available, assuming it catches and finds its community.
I think the idea that all editions have to be promoted in a specific way is a bit of a process mirage. Editions need to be promoted in an effective wave. And it appears that Council ultimately has authority on how to promote editions via the website properties. It shouldn’t be concluded that this will be promoted in the same way as Workstation and KDE Plasma editions. But that’s a discussion we have as a viable working group is coming together. What does effective promotion for this output look like considering the state of the technology? How do we carve out the effective promotion that articles a call to action to participate towards an ethical future for this technology?
And on that note, I’ve already had some discussions about needing a dedicated communication channel separate from Fedora Magazine and the community blog where we are effectively communicating about AI goings on in the context of the Fedora Project.
I do not think that promoting an AI Developer Edition as a Council Initiative is sending the signal at all that we are interested in ethics. It is sending the signal that we are happy to engage and contribute to the status quo and immediate development of AI as it is.
This need for a special space would again highlight how we value AI considerably, and that it is special and worthy. Perhaps instead you could write an article for Fedora Magazine on how you envision an ethical technological future that does contain AI?
Use the existing channels with existing readerships to make Ethical AI part of Fedora, not something superior or different.
And as a note, I don’t think an AI Developer Spin needs a Council Initiative. Being an Initiative means it is a core focus for Fedora to produce that item and make sure it is successful. As an Initiative, the message is that AI, as it is, is something we need to promote.
Hmm..
If there weren’t some interesting hardware enablement asks that need to be resolved here, this probably wouldn’t rise to the level of a council initiative. I definitely don’t want the technical work to be unsuccessful.
But I’ll go further than that. I think outreach into the open source AI developer community is long term strategic. And I will be promoting particpation in Fedora to them, because I think the right people who can put the technology on to a better ethical trajectory are there and we need to be in conversation with them. This output maybe the best way I can grab their attention and have a space for them to work with us and get to the best achievable future.
I’ve taken a week to let my thoughts settle regarding this post, and now I’m ready to reply more carefully.
I agree that we should use more data in our decision-making process. However, this should apply consistently to both positions. Otherwise, we risk holding different arguments to different standards.
As I remarked on my previous reply, this is mixing two different things. Please, let’s avoid the package-deal fallacy here. We can provide an AI developer desktop without making this a Fedora objective. We already do this for multiple things, e.g Fedora KDE Plasma Desktop is widely used but AFAIU, it is not a Fedora Objective to provide/support a Fedora KDE spin to users/developers.
Not sure if this is your intention, but this reads as prioritizing relevance and adoption over ethical considerations. Since Fedora is a community-driven project rather than a corporation, it seems like an appropriate space to bring in those ethical and philosophical aspects.
I don’t think anyone in this discussion is rejecting everything AI-related. Others and I just raised concerns about the reputation impact of making AI a Fedora Objective.
I don’t agree that success would be driven entirely by organic user interest. Adoption of technologies is often influenced by broader ecosystem pressures such as corporate investment.
I’ve spent some time thinking about this paragraph, and I’m struggling with it. As Fedora Project Leader, expressing no concern about potential reputational impact could be interpreted as dismissing an issue that some contributors clearly care about. Even if the change is non-disruptive in a technical sense, it can still influence how the project is perceived and what it signals about Fedora’s priorities. I think that concern deserves more consideration.
I did warn everyone on my first official day, standing on stage at Flock last year, that there were hard conversations coming. AI is going to be several of those hard conversations.
Here is why I’m not currently concerned about reputation damage in May 2026. Last month, the Linux kernel development community adopted an AI code generation policy that explicitly allows for the use of AI code generation. It compares well with the text of the Fedora policy, though they have more details in what should be in the Assisted-by tag to help with data collecting about which models are being used. As a project we’ll probably want to learn from their policy so we can do our own data collection concerning model use as we start seeing assisted by tag use in our project contributions.
I noticed you framed this concern differently than Fabio did. I was replying to Fabio’s specific concern about reputational damage. You’ve stated a wider concern about whether I’ve fully considered reputational impact. I feel like I also need to speak to that reframing. I have considered the impact, both the downside and the upside. I actually think there is net positive reputational impact in this approach. We are creating a space to be in conversation with the segment of the open source contributor ecosystem that values this technology as a development tool. This will help us gain influence with a segment of the community we aren’t well-equipped to reach at present. I know it’s a delicate situation, but I think we are threading the needle here. We avoid putting the AI tooling front and center with the users who don’t want it, but we build an output where it’s the entire focus, so we can get that segment of the community on board with working with us to bring this technology into a better and more ethical future.
As I noted in the second reply to Fabio, there is nuance in how the output from this initiative should be promoted. I’ll reiterate, don’t assume that this will be promoted in the same way Workstation or KDE Plasma Editions are promoted if this ends up being an Edition. Even if this ends up being a spin, I will end up needing to promote the output to specific audiences, because I view it as an important strategic bridge into a segment of the open source community that we have to be in partnership with in order to avoid the worst possible future.
Let me paint you a picture of the future… what I see 2028 looking like.
I’m looking at a future where it’s conceivable that the majority of commits into the Linux kernel upstream will have AI assisted-by tags in a couple of years. There are some people out there who think we’ll hit that mark much sooner than that, but I’m thinking 2 years. And where the Linux kernel goes, so does the bulk of the critical path ecosystem that makes the OS work. Yes, sure, there will be some projects where AI is not allowed at all. But I fully expect the bulk of the critical path components we build an operating system from will have AI-assisted code in it and assisted development will be normalized as an open source linux ecosystem development activity upstream of this project.
What’s the best possible version of that future?
It is an open question whether any of those commits will be done with open AI models that run on hardware contributors control — or if all will be by the large centralized AI models via APIs and pay-as-you-go token use. I believe that Fedora can help make sure it’s not the worst of those two futures.
In this future, users and contributors who have deeply held philosophical concerns about AI are going to tie themselves up in knots, knots of the Gordian variety. Are those people going to fork the Linux kernel and build an entire ecosystem devoid of AI assisted development? Are they going to jump to one of the BSDs that has taken a hardline stance? I don’t know. But I am pretty sure there is no AI-less future for the development ecosystem on which the Fedora Project currently depends. I think it’ll be hypocritical for any Linux distribution in 2 years to take a hard line stance on AI as a development tool. And I don’t see Fedora jumping to the NetBSD kernel.
Like all Gordian knots, it requires bold action. We have to do something. The question becomes what actions are impactful and achievable?
So what is actually achievable in the future? I think Fedora needs to be in the conversation about what exactly AI assisted development looks like broadly for the entire linux development ecosystem. We need to be pushing towards better, more ethical model use by offering a developer environment that, as a long term objective, includes making it easy to work with local models. We need to help the local models compete with the large centralized models for developer workloads.
Will small, more-open models be able to compete? I don’t know. But this space is under active development.. and as the big models seem to be hitting the laws of diminishing returns now. Maybe there is a chance the smaller models can catch up. The people that probably have the best idea about that are working in the AI space right now, and I need to build a bridge to them. This objective is one of those possible bridges. I’d like to not burn it, before we build it.
If we aren’t there trying to push towards the future with the most ethical AI development workflows that are achievable… we risk getting the worst possible future by default. I don’t think there is a future available devoid of AI-assisted development as a normalized workflow in the critical upstream ecosystem that we depend on. We can have the most impact by trying to figure out how to provide the most ethical version of this technology and put it into the hands of developers to choose to use.
This isn’t fun. This is a hard conversation. I’m not dismissing anyone’s concerns. In fact I probably agree with most of the concerns anyone could think to express to me about AI use. I live in the region of the US, Northern Virginia, that has the highest density of data centers. Both power and water usage are real concerns that I’m directly impacted by now that I no longer live in Alaska. There is arguably a drop in the water table regionally because of these things. The path we are on isn’t sustainable. Telling people to stop using this technology isn’t going to work. Offering people a more ethical version, which they can actively contribute to making better, may help.
I’m not dismissing the concerns. If I’m doing anything, I’m disagreeing about what to do with those feelings. I don’t think we can just ignore the technology or wish it away. We may be able to address some of the concerns via government legislation. But I think, as technologists, it’s incumbent on us to provide a version of this non-deterministic technology that is better in the ways we care about, so that our version is what open source developers choose to use when they choose to do assisted development.
That may mean getting our hands a little dirty at the start, so we can build the bridge into the segment of the open source community that sees value in the technology. This bridge should allow us work to make incremental progress towards an achievable ethical future. We do so knowing there is a gap between what is achievable and what is best as seen from where we stand right now. The gap may be too wide for some. And if you’re in that group, all I ask is you give us the space to walk towards what is achievable. By doing so, we might see the gap close because each step changes our vantage point.
Fine. I do not think we can move this forward in a community way. I am withdrawing all my activities in Fedora project starting right now. The present situation in Fedora is clearly not for me.
I will move out with ethics and philosophy to another place so you will have more space.
Can I ask you to stay? I like hearing your view-points and having people in the community that have a basic understanding of philosophy.
My background is in ethics, metaphysics and community development.
While I don’t think Jef has put the ethics of AI out there yet, the ethics of the Fedora community are sound. We are surrounded by techno-fascists and we need to work together with our disagreements.
Jef is totally right in saying that AI is here to stay. The Fedora community cannot influence that. What we can do is keep inclusion, friendliness, being excellent to each other and a secure base operating system to form a bastion of sanity.
I have looked so many places for perfection, for the ethical society, I search for the cup of life. And it is with Fedora, here I will stay and work with one of the best teams I have found.
Thanks for the kind words but I keep my decision. I don’t think there is a community-driven way to move this forward for now. Time will tell and maybe at some point I will be back. For now, I consider I should leave.
This seems like a common sentiment so I want to address it.
In short:
1: Those two things are not in conflict
2: Relevant and adoption are absolutely things we should prioritize
3: While some applications of AI/ML are unethical, boycotting the entire field is (IMO) neither rational nor effective
Regarding ethical considerations:
Fundamentally ML development is the study of the tasks that a computer can repeat, through pattern recognition. Mechanical repetition of tasks is the purpose of a computer. It is no more rational to argue that AI is unethical than to argue that computing is unethical.
Arguing that AI is inherently unethical ignores beneficial applications of the technology. Computer vision has many ethical applications, including assistance for the blind. Voice recognition has many ethical applications, including assistance for the deaf. Input via eye tracking provides assistance for those with mobility limitations. All of these are active fields of research and development using AI. reject the idea that we should abstain from AI development due to ethical concerns, and I think we have a responsibility to support the development of assistive technologies, which are largely built on AI foundations.
None of that is to deny that unethical applications also exist. They certainly do. But they won’t go away because we choose to abstain.
…which brings me to relevance and adoption.
One of my core and fundamental beliefs is that a community must attempt to grow, or it will vanish. Communities are composed of human beings, and the participation of human beings is necessarily temporary. Community members will leave when they get bored, or when the priorities in their lives change, or for limitless other reasons. Attrition is inevitable. In order to be sustainable, a community must actively engage in community building.
I’ve been told that people approach Fedora representatives at conferences and tell us that they would like to use Fedora but that it’s difficult, and that scares the crap out of me, because it tells me that there are technical barriers to community building. It tells me that outreach and philosophy are not enough to make Fedora sustainable. It’s one thing if people choose not to use Fedora because we are not ideologically aligned, but that it’s DIFFICULT? That’s very bad. If we’re having difficulty growing our community with technology we don’t want, then we’re also going to have difficulty growing our community with technology we do want.
I’m not supporting this work because I think relevance is more important than ethics. I don’t think that at all. I would be opposed to growing the community through unethical practices or unethical development. But I don’t think that the things we’re proposing are unethical, so I don’t think these two concerns are in conflict.
I don’t have all the answers. I think the people working on the open source medium and small models have more of the answers. I think some of the AI ethics and safety people may have some of the answers.
We are careening into the worst possible future. I’ve tried standing outside in the rain and screaming for it to stop…that didn’t do anything.
Assuming AI assisted development is going to be normalized in the open source ecosystem. The worst possible future is for everyone that chooses to do ai assisted development is using one of the big models. That’s unsustainable and I believe it puts way too much power into the hands of what you would likely refer to as techno-fascists. If there is any way out of the worst possible future for development workflows specifically, it would involve small models running on local hardware… models we control* on hardware we control*
to the extent we can control either.
For some that means things like one of the new AMD laptops, for others that probably means NVidia hardware. For others that probably means vendored accelerator hardware offered by one of the cloud providers where they ssh into… anywhere where Local Language Models ( the other LLM) can usefully run.
For what it’s worth, as context for how discussions are going in other FOSS projects about these things - I think this post on the KDE Discuss forums sums up a good perspective, and a generally similar conclusion to Jef’s here:
With ollama 0.20.7 in rawhide and pi-coding-agent (bz2464801) fedora will be self contained to do all local agentic work. I use a framework desktop. ROCm works, vulkan should. Give it a try on whatever gaming gpu you have. I am using now to help with the next package.
Yep. We’d have to check back with Gordon for specifics, but my understanding is ROCm is intended to be in scope of baseline hardware coverage. Credit where credit is due. The open source hardware vendor support for ROCm to my understanding is great.
An AI desktop spin targetting just ROCm would be well within the scope of the existing spin approval process. And if this were a year ago, I probably would have agreed that the NV enablement wasn’t important enough to catapult this from a spin into an initiative. Because its clear that the uplifting of this into an initiative comes with introspection about how we all feel about the use case of AI assisted development. If I wanted to just sneak this output in via established community process, it would just be a spin and we could avoid the hard conversation entirely…for a while.
But I’m concerned enough about what I envision as normalized use of AI assisted development in the open source ecosystem, that I think we need to do the NV enablement work, so we can create as large a pool of people as possible who want to work the question of whether the local models on local hardware can serve in assisted development workflows.
I really don’t want yall in the ROCm corner of the ecosystem to feel like I’m ignoring you or playing favorites with vendors. I’m genuinely concerned about the state of things, and I want to create the biggest possible community of people who can work the problem space of the utility of local models so we can maybe avoid the worst possible future available to us. Hence the NV hardware enablement inside the scope of an initiative.
I believe a point of friction is the defining of rules and examples for the AI Developer Desktop objective. We need to get into some more details, even if none of us have all the answers right now; I believe there are concepts and principles we can all try to find agreement on ahead of time, if this endeavor is to suit a broader community driven purpose.
First we need to ask some questions:
Is this desktop supposed to eventually “integrate” with AI models and surrounding software?
Or is it intended just as a stable development platform for use and experimentation of ai workloads?
What are the acceptable use cases for ai on the platform?
Is there is a onger term plan to create Fedora ai as separate software?
Will it be hosted on flathub as a flatpak to serve as an example of ethical ai for everybody to use and contribute to?
What rules will be in place to ensure safety, security, and FOSS compliance from start to finish?
How do we ensure proper attribution for contributers for AI outputs?
And where do we draw the line for content restrictions? Where will the training data come from?
What are the minimum and recommended hardware requirements going to be aiming for?
These are just a few that come to mind. Really we need everyone to chime in with their own questions and expectations for this objective, so that we can bring some detail and clarity to the foreground. If certain aspects remain too vague it breeds distrust, that should be avoided. Let’s figure out how the nuts and bolts fit together for an ethical ai platform to be built, and perhaps an ethical ai model down the line.
And as always the goal should be to create something that the Fedora Community can be proud of.
I would love for you to stay as well, I deff think that while the topic may be rough we -as in a community- still have the right mindset to do so.
Overall, from my side, I would love to have some more clarifications, as it’s mentioned about the thread. AI is here to stay? I think we all agree on that and that we’ll have to learn to co-exist with it.
That said, there’s a lot of concerns on there. Let’s say we accept this, will that mean that Fedora will shift to be an AI-first linux distribution? While a big topic, I think this goes way beyond to just ‘making Fedora AI-capable by default’.
What does this mean? Integrating nvidia/ollama/be my guest and be able to somehow run LLMs natively? What about the model themselves? Do we aim to support open weights models? open source? Do we even care about the dataset the models have been trained into?
I guess that’s also part of the ethical side @ffmancera mentions there. Besides that, AI is just so much mode than LLMs, does this ‘objective’ cover this or aims to? @jspaleta I’ve always valued Fedora as as a distro for both the community and the bleeding edge, and I think whatever we do should take both sides into consideration.
ified.