OSes
Furores are fermenting in the forums
Both Ubuntu and Fedora have made it official: support is coming soon
for running local generative AI instances.
An epic and still-growing thread in the Fedora forums states one of
the goals for the next version: the Fedora AI
Developer Desktop Objective. It is causing some discontent, and at
least one Fedora contributor, SUSE’s
Fernando Mancera, has resigned.
Fedora Project Lead Jef Spaleta, who took
over the role from Matthew Miller a year ago, remains resolute, saying:
I have zero evidence in front of me that users are being driven away
from Fedora because of AI.
As far as Red Hat’s community distribution goes, while this may be
controversial, this should not be a big shock. In October last year,
The Register reported that the Fedora
council approved a policy allowing AI-assisted contributions, and
anyone following the IBM subsidiary’s movements will already know that
last June’s RHEL 10 release includes access to an LLM-based online
helper chatbot: we tried
it out when the product was released. We also reported on the
managers of Red Hat’s Global Engineering department being notably
keen on the use of AI just last month.
Since Red Hat has other offerings for slow-moving stable server OSes
– and arguably because Debian, Ubuntu, and their many derivatives have
the stable-desktop-distro space nicely covered already – Fedora has a
strong focus on providing a distro for developers, and Spaleta’s
announcement makes this clear. The goal is:
to build a thriving community around AI technologies by focusing on
three key areas: equipping developers with the necessary platforms,
libraries, and frameworks; ensuring users experience painless deployment
and usage of AI applications; and establishing a space to showcase the
work being done on Fedora, connecting developers with a wider
audience.
He also spells out what it doesn’t want to do:
Non-goals:
The system image will not be pre-configured with applications that
inspect or monitor how users interact with the system or otherwise place
user privacy at risk.
Tools and applications included in the AI Desktop will not be
pre-configured to connect to remote AI services.
AI tools will not be added to Fedora’s existing system images,
Editions, etc, by the AI Desktop initiative.
In other words, tools for developers, not for end-users, with a strong
emphasis on models that run locally, and which preserve the user’s privacy. It’s also worth pointing out that Fedora has had an AI-Assisted
Contributions Policy in place for six months, and earlier this
month, Fedora
community architect Justin Wheeler explained in some detail Why
the Fedora AI-Assisted Contributions Policy Matters for Open
Source.
Our impression is that the Fedora team feels that it needs to keep
Fedora relevant for growing interest in LLM-bot assisted tooling, and
that it can address concerns from hardcore FOSS types by ensuring that
this means local models, built according to FOSS-respecting terms,
deployed in privacy-respecting ways.
Fedora is not alone in this, though. There are also ructions across the border in Ubuntuland. Right after the release of the Canonical’s new LTS version, Ubuntu 26.04 Resolute Raccoon, Canonical’s veep of engineering Jon Seager laid out the future of AI in Ubuntu.
We interviewed
Seager last year during the 25.10 Ubuntu Summit, and back in January this
year, he published his views on Developing
with AI on Ubuntu. Now the plans are firming up.
Like Fedora, there’s a strong focus on local models and confidential,
privacy-first deployments – and ensuring that the OS and the tools
support GPU acceleration from the big hardware players in that space.
However, unlike Red Hat, Canonical isn’t pushing its developers towards
these tools. In what we see as a veiled jab, Seager’s announcement
says:
We are not setting shallow metrics on token usage, or percentages of
code written with AI, but rather incentivising engineers to experiment
and understand where AI tools add value.
Initially, the focus is on users instead:
AI features in Ubuntu features will come in two forms: first as a
means of enhancing existing OS functionality with AI models in the
background, and latterly in the form of “AI native” features and
workflows for those who want them.
As Fernando Marcela’s exit shows, an emphasis on what could be termed
FOSS-friendly AI – open models, privacy-centric, local execution and so
on – is not enough to placate those who are really
strongly averse to these tools. The Reg FOSS desk counts
himself firmly in this camp.
Back in January, we reported on the rise,
fall, and resurrection of OpenSlopware, a list of FOSS projects
which contain LLM-generated code, integrate LLMs, or even show the
traces of the use of LLM agents. Soon, it seems inevitable that Fedora
and Ubuntu will both feature here.
Resistance, though, is also rising. Stop Slopware tries to help explain
why and how to avoid it, and there’s also The No-AI Software
Directory for projects that have explicit LLM-free policies,
whether they’re FOSS or not.
Bootnote
It amuses us to note that both the Ubuntu and Fedora forums use the
same software, called Discourse. (It’s a sort of web
forum as designed by people who have heard of mailing lists,
but don’t know how to use them and find the idea of bottom-posting
confusing.) Some could interpret this shared adoption as a sign of
underlying similarities between the two projects. ®
