2026-04-27 23:14 UTC
@Amgine@mamot.fr @mayintoronto@beige.party other: question is a basically an impossible hypothetical. The benefits of unethical training are so many orders of magnitude better than anything possible ethical training that it just isn't going to happen
Replies (1)
-
@Amgine@mamot.fr 2026-04-28 04:07
@Funcan@mastodon.social @mayintoronto@beige.party And no one will write their own kernel, or create a crowd-sourced encyclopedia, or be able to fly an electric airplane. I have given up saying things are impossible. There are plenty of LLM projects which have ethical and curated sources. E.g. whatever is driving LibreTranslate these days. I think the better question to ask is: when is using LLM the best choice? I expect it is very rarely. Improvement means simplifying, reducing materials, less energy.