Post #2932836
2026-03-14 20:02 UTC
@will@www.librepunk.club @gloriouscow@oldbytes.space @volpeon@icy.wyvern.rip That still has the potential for you getting a model that is trained with heavy biases, manifesting existing oppressive structures. It could still lead to AI psychosis. Just two name two potential issues. Running something locally doesn't fix who trained it and many other problems.
Replies (1)
-
@will@www.librepunk.club 2026-03-14 20:32
@sophie@floss.social @gloriouscow@oldbytes.space @volpeon@icy.wyvern.rip Certainly, but I think these problems can be solved and (my point is that) it's doomerist to think they can't.