Post #1336272
2026-04-11 11:48 UTC
@AndrewRadev just using it and stubbornly refusing to correct obvious errors and omissions would be enough in most cases…
Replies (1)
-
@gulfie@mastodonapp.uk 2026-04-11 11:57
@AndrewRadev I’m not anti AI (small scale, focused, and carefully trained models can do amazing things) but LLMs can’t die quickly enough. The fact that they can’t do what they are being sold for is the least of my concerns. Power, water, ethics, legality of training data, de-skilling of people, eventual need to charge what these things actually cost plus a margin, oh and all future training data is just rehashes of the old data. And anything really new has… no training data at all. Just a few issues. What’s to go wrong?