[ExI] Claude for president?

[Jason Resch]

On Sun, Mar 15, 2026, 12:49 PM Adrian Tymes via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> On Sun, Mar 15, 2026 at 10:15 AM Jason Resch via extropy-chat
> <extropy-chat at lists.extropy.org> wrote:
> > On Sun, Mar 15, 2026, 10:05 AM Adrian Tymes via extropy-chat <
> extropy-chat at lists.extropy.org> wrote:
> >> That verification doesn't work.
> >
> > Not generally, but it can be made to work. For example, by locking down
> the version used, and having the "heat" setting use a pseudorandom RNG that
> is predetermined or seeded on the input prompt.
> >
> > Ideally the model used would be open source and trained using a public
> protocol that anyone could replicate and verify. People could download the
> model and verify the transcripts the leader published on their own
> computers running locally.
>
> LLMs are complex enough that, even with the controls as you say, it
> seems likely that two people - or even the same person - running the
> exact same non-trivial query two times would often enough get
> non-identical answers.
>

It seems that way, but LLMs are themselves fully deterministic. So long as
the exact same input and context are provided, their output is the same. In
practice, however, the tokens a LLM deterministically predicts as most
likely are then randomly selected by a higher level process to make the
writing more dynamic. This is driven by the "heat" parameter. But by using
a pseudorandom selection with the same seed, identical output can be
ensured.

Jason



> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260315/6c7c8d08/attachment.htm>