On Tue, 28 Jul 2026 at 09:17, Simon Richter <[email protected]> wrote: > Hi, > > On 7/27/26 10:55 PM, Andrey Rakhmatullin wrote: > > >> Perhaps such decisions are within the wide margin of appreciation we > >> already afford to maintainers. > > > Definitely, with the appropriate measures applied to those maintainers > > if deemed necessary. > > So your position is "we should not create a rule against AI because that > would be heavy-handed, but come down with the full force of the > apparatus on anyone who does not accept AI"?
https://en.wikipedia.org/wiki/Paradox_of_tolerance seems to apply in full force here. *Assuming* that the community decides that use of particular tooling is acceptable and even beneficial to the community, it becomes quite anti-social for some individuals to actively continue blocking the use of that tooling (by others) with their individual decisions. Consider the developers on the loosing side of the systemd discussion for example. This is different to forcing *you* to *use* AI (in your Debian work). That would clearly be a bad call. But your preferences should not be enforced on others that you collaborate with. So you should not be able to force *others* not to use AI, including by blocking their patches and submissions. In team maintenance scenarios this could be solved by having someone else review those patches instead. Without such option this kind of behaviour would become disruptive and problematic. You could see the exact same pushback if you substitute "LLMs" with "Windows". A maintainer rejecting any and all patches created on Windows would not *really* be widely seen as the paragon of free software virtue that they themselves (presumably) see as. MAybe RMS could get away with it. But the rest of us are not RMS. :) -- Best regards, Aigars Mahinovs
