from Fedora Project Leader Matthew Miller, on behalf of the Fedora Council

First, a personal note! As you may have seen, I was out sick with Covid for a month after getting home from our annual Council face-to-face meeting. It’s not been fun — some respiratory symptoms, but primarily, overwhelming fatigue. Somewhat ironically, the timing suggests that I managed to avoid catching anything at FOSDEM itself (where I wore a mask most of the time), or at the Council meeting, but rather on the plane or in the airport on the way back. Although emergency measures have been lifted, there really is still a pandemic going on. Be careful, everyone, especially when traveling! In any case, I’m back to myself now, and am excited for Fedora’s next big steps.

The Story so Far

So! I’ve been talking about “Strategy 2028” for a while — we started this effort seriously about a year ago. If you’re just joining in, or want a refresher, Fedora Strategy 2028: a topic index for our planning process is a great place to start. I won’t rehash all of that here.

The important thing is: 2023 was kind of a hard year, and although we made some progress, we lost momentum. The Council hackfest helped get things back on track, and we’re moving forward now. We’re not making any fundamental changes, but we are restructuring how we present things — and we’re moving on from theory to practical work.

A New Presentation

Over the last year, we got feedback from many people who felt that the organization into Themes and Focus Areas was overwhelming and confusing. It felt really big, maybe too much, and hard to know where one might fit in.

After some discussion, we decided that we could better present our various objectives as they align with something we’re all already very familiar with: Fedora’s “four foundations”: Freedom, Friends, Features, and First. The objectives we’ve talked about over the past year aren’t going away, but this gives us an easier way to organize and prioritize them over the next few years.

What Else?

Even with a big and ambitious plan, we left some important ideas behind. For example, we want to better tell our Fedora stories (both inside and outside the project), and we want work to improve distribution security. We’re not immediately adding planned effort on these, but our new framing around the Foundations gives us more room to add those things as needed.

Our project has amazing, passionate people who have done amazing things for Fedora and for the world of community-built, free and open source software in general. We should do more to celebrate them, to give everyone the recognition they deserve. We should show new and potential contributors the incredible things we’ve done and the exciting things we’re doing — and what it means to be part of our community.

The recent xz exploit relied on long-term, persistent social engineering within an open source project, along with a very sophisticated technical attack. A long discussion on the devel mailing list shows that we have many ideas for how we can make Fedora more resilient against such attacks as a project, and Fedora Linux safer as an operating system. Initiatives to work on these  improvements may end up under a different umbrella, but will certainly have some interconnection.

As we go through the next five years, don’t be surprised to see us add Community Initiatives focused on these things — and others, as they emerge.

And then, there’s AI

We’ve all heard a lot about various “large language model” chatbots, and about image and even video creation. There is obviously a lot of hype around AI — and inevitably, over-the-top nonsense. We’re not really “just five years away” from actual generalized Artificial Intelligence. This isn’t the start of Skynet from The Terminator movies. Human creativity isn’t being replaced, and I don’t think all programmers will end up as “prompt engineers”.

However, there is something real here.

Advances in accelerated hardware and machine-learning software have unlocked possibilities which were imagined last century, but which were not practical at the time. When the dust settles around the hyperbole, I believe we’ll still be left with something significant, powerful.

In addition to the big showy LLM-based tools for chat and code generation, these advances have brought big jumps for more tailored tasks: for translation, file search, home automation, and especially for accessibility (already a key part of our strategy). For example, open source speech synthesis has long lagged behind proprietary options. Now, what we have in Fedora is not even close to the realism, nuance, and flexibility of AI-generated speech.

Right now, most of all this is proprietary. It’s corporate-owned closed models trained with hidden data, largely running on hardware without open source drivers. If we ignore this, we’re going to be left behind — not just Fedora, but free and open source software entirely. On the other hand, we can take a leadership position, and build a future where AI belongs to all of us.

The Guiding Star for Strategy 2028 is about growing our contributor base. We can make Fedora Linux the best community platform for AI, and in doing so, open a new frontier of contribution and community potential.

This won’t be easy. We have a lot of basic work on platform fundamentals. That’s drivers and tooling, packages and containers, and even new ways of distributing Fedora software. We also need to improve developer experience — for example, it’d be nice to have Podman Desktop as part of Fedora, with easy paths to getting started.

We can use AI/ML as part of making the Fedora Linux OS. New tools could help with package automation and bug triage. They could note anomalies in test results and logs, maybe even help identify potential security issues. We can also create infrastructure-level features for our users. For example, package update descriptions aren’t usually very meaningful. We could automatically generate concise summaries of what’s new in each system update — not just for each package, but highlighting what’s important in the whole set, including upstream change information as well.

At the same time, we need to work with the rest of the free and open source world to create labels for AI technology which aligns with our values. We need polices on what we allow, and on what we encourage. We may even need to fight for legal frameworks friendly to community-built software.

Finally, of course, we can provide models in the OS. This includes accessibility tooling, which can benefit everyone: imagine an all-local voice assistant that doesn’t send your conversations to some big datacenter or try to sell you things while simultaneously selling your personal data. We could also include tools that make it easier to find help, features to simplify system administration tasks, and interfaces to better organize your documents and media — all within your control, all running on a free and open source software stack.

We’re just getting started here (with, for example, Pytorch coming in Fedora Linux 40 1, CPU-only — for now). There will be more exciting things coming soon.

Next Steps

The next post will be the high-level view of Strategy 2028, updated from this new perspective.

The Council’s next immediate step is to identify Executive Sponsors and leaders for each Focus Area under the broad umbrellas of our Four Foundations. Then, we’ll plan and schedule concrete outputs and practical activities in each area. For larger efforts, we expect to launch Community Initiatives, but much of the work will be organized as smaller projects under each focus area banner. Expect more announcements soon as we build this out!