I attended Red Hat Summit 2026 in Atlanta last week from May 11-13. As always, the keynotes and product announcements were interesting, but the real value of the event happened in the "hallway track." The 1:1 conversations over coffee or meals provide the blunt truth about how people actually feel, especially regarding heated community topics like the proposed Fedora AI Developer Desktop.
Over breakfast one morning, I asked three different people a simple question. If you could change one thing about Fedora governance or the project right now, what would it be? I received three entirely different answers. Under the Chatham House Rule, I want to share those concerns and what I believe they mean for our immediate future.
1. The AI Developer Desktop: The Weight of the "-1" and the Demand for Consensus 🔗
The first piece of feedback came from a FESCo member who expressed deep exhaustion over the recent Fedora AI Developer Desktop proposal. The FESCo member highlighted a structural imbalance: the feeling that appointed roles on the Fedora Council wield too much executive authority, leaving elected Council representatives feeling powerless to stand up against what may appear like thorny, political pushes.
This is not just an abstract fear. I heard from past and present Council members who privately expressed reservations about a proposal, yet felt immense pressure to vote "yes" publicly. A strong desire for unanimous consent inadvertently coerces people into agreement, even when core concerns remain unaddressed.
This dynamic is exactly why I cast a -1 vote on the Fedora AI Developer Desktop initiative.
If you look at the Fedora Council charter, we operate on a consensus decision-making model designed to foster collaboration and seek agreement.
A -1 vote does not act as a hostile veto or a permanent denial.
By definition, it is a mechanism that "immediately halts the process and requires discussion."
It triggers a mandatory cooling-off period.
When we move too fast, risking the alienation of our core engineering and kernel experts, we must utilize the tools our charter gives us to buy time, address specific concerns, and ensure actual, rather than theatrical, consensus.
Burning bridges in any community is far easier than rebuilding them.
2. The Vision Void Surrounding the AI Developer Desktop 🔗
The second insight came from a long-time Fedora contributor. This contributor works for a major cloud hyperscaler, but their heart always prioritizes the community. The feedback was simple: our messaging is muddled, and we fail to tell a coherent story about what we produce.
This symptom points to a larger issue: we currently lack a cohesive, strategic vision. We build more deliverables than ever, yet new users find it increasingly confusing to know where to begin. It reminds me of the pre-2014 era of Fedora, before Matthew Miller led the Fedora.next strategy. The creation of Fedora Editions gave us a crisp, defined story about our work and impact. Today, in 2026, perhaps we find ourselves lost in the weeds again.
Our current Fedora Project Leader, Jef Spaleta, privately shared a vision that he sees as critical for the future of the project. However, a vision cannot effectively guide a community if it remains private and lacks community input. The project would benefit immensely if Jef shared that vision publicly. I offer this as an open, collaborative invitation to put those ideas out into the open so the community can read them, critique them, and help shape them. We cannot align on a strategy we have not read.
3. Trademarks, Transparency, and an AI Developer Desktop Remix 🔗
The final piece of feedback came from an EPEL maintainer with deep community roots, currently employed by Red Hat. He expressed frustration over Red Hat using the Fedora trademark in major product announcements without clear community visibility.
Effective Fedora governance requires modernized trademark guidelines and a commitment to transparency. I propose creating a public trademark ledger. If a trademark authorization requires an embargo for a corporate announcement, we can accommodate that. It would be unwise for the future of Fedora to turn down opportunities to work together with partners on announcements that require this initial secrecy. However, any such embargo must carry a strict expiration date, after which we publish the authorization to the public ledger.
This brings us back to the Fedora AI Developer Desktop. We need clear branding boundaries, and we may already have the perfect tool for this: the Fedora Remix.
The Fedora AI Developer Desktop could be an excellent candidate for a new Fedora Remix. Whether it operates as a formal Community Initiative or not, the Fedora Remix model gives the team driving the work the liberty to take risks, try new ideas, and include necessary proprietary bits (like Nvidia CUDA) without forcing them to follow the strict, high bar of official Fedora deliverables right out of the gate. Taking this path prevents the alienation of our core contributor base while still allowing innovation to happen. If it succeeds and sustains a community, we can always formalize it later.
Moving Forward from Red Hat Summit 2026 🔗
First, I am grateful for the candid, honest feedback given to me when I asked an open-ended question to these three community members. I emphasized each person to be honest and to think big, if they could really change anything but only a single thing. I admired the thoughtfulness each person gave to their answer, even if we all acknowledged most of these challenges did not have any "easy fix". This goes to say, Fedora remains strong because our contributors care enough to have these hard conversations. So, let’s use the tools we have (e.g., charter-driven cooling-off periods, transparent public visions, and the Remix model) to build a future that respects the Four Foundations that Fedora is built on.