Hi Hugo,
Hugo Buddelmeijer <[email protected]> writes:
On 5/6/26 03:33, Ian Eure wrote:
Hi Hugo,
Hugo Buddelmeijer via "Development of GNU Guix and the GNU
System
distribution." <[email protected]> writes:
On 3/5/26 22:09, Ian Eure wrote:
Hi Untrusem,
Untrusem <[email protected]> writes:
So that's why I wanted a way to know when a guix package
updates or
gets added. It should explicitly mention that LLMs is being
used in
the software to let the users know that the software they
are using is
slop or not.
I completely agree with you: users should have the freedom to
run
software which doesn’t contain LLM output.
The freedom yes, but that freedom does not necessarily oblige
others.
Oblige others to what?
Your claim seems obviously true at face value: "users should
have the
freedom to run software which doesn’t contain LLM output". As
in, not
having such freedom would entail people forcing others to run
software
containing LLM output.
I think both LLM lovers and haters on this mailing list would
agree
that people should not ever be forced to run any software at all
whatsoever (irrespective of any LLM use).
I agree, the software a person runs is (or should be) their
decision.
So I was inferring that you meant something more: that people
should
actively help users with achieving that freedom, and that is not
necessarily true. It might be true, but that is a much
different
claim.
What I proposed is completely neutral: people declare whether the
package uses LLM output. If you dislike LLM output, you can take
steps to minimize it. If you like LLM output, you can take steps
to maximize it. Obligation to declare has zero effect on
obligation to use (or not).
I'd like software written by people who don't fly (as I see
that as a
bigger existential risk than LLMs), but I'm not going to ask
people to
disclose whether they fly. (And I wouldn't hold it against
them
personally if they do fly.)
Whether I fly on a plane or not has no effect on the
copyrightability of my code; use of an LLM may.
I inferred from your statement that "users should have the
freedom to
run software that doesn't contain LLM output" that your
aversion
against LLM's was mostly about something else than copyright;
that
you have ethical concerns.
I have multiple concerns with it, as it’s bad in multiple ways.
I wouldn’t hold it against someone if they flew on a plane, but
if I
told them I had reasons for not flying, and they argued with me
about those reasons, demanding proof to satisfy them that my
objections were legitimate, said that all travel would happen
via
plane, and made sure to mention often how great first class is,
I
would absolutely hold *that* against them. This last sentence
is
not about planes at all.
In that sentence, you do not ask anyone to do anything. In
particular, you have not asked anyone to stop flying, and you
have not
asked anyone to disclose whether they fly. Are you sure you
have the
same sentiment about LLM use?
If someone would make a P.R. to Guix with LLM generated output,
would
you ask them to disclose that? Would you ask (or entice) them
to stop
doing that? Or would you let it be? If you do want to restrict
others, then it is only fair to be asked to substantiate your
objections.
I don't even care about the P.R.s, or the copyright, I care
about the
essence. I'm trying to figure out what the core objections are,
because this is very powerful technology, so we should
understand it
well.
The writings of the free software movement are excellent and
those of
the electronic frontier foundation are too. They are very
specific in
their complaints and their remedies. They are very convincing.
Is
there really nothing like that for the AI debate?
(And if there isn't, and we are not allowed to ask people, then
how do
we get anywhere?)
Thank you for the vivid illustration of the exact point I was
making.
-- Ian
