Home | Notifications | New Note | Local | Federated | Search | Logout

Note Detail

Reply to @anthropy@mastodon.derg.nz
rubén@ruben@friendship.quest (2026-09-21 18:51:47)
@anthropy@mastodon.derg.nz @relay@relay.toot.io Some of the contradictions could well be “a social effect of llms” vs “the technical capability of the llm”.

Eg. Jobs being taken away = a social effect due to hype. But the output being terrible = a technical one.
---Reply---
Anthropy@anthropy@mastodon.derg.nz (2026-09-21 18:53:49)
@ruben I mean, I think there should be better protections in this era for employees, and in europe who already had good labor laws it seems like it's much less of a problem; Personally I'd say that this is more of a 'vicious abuse of workers' case than something that's truly 'the fault of AI', because those companies would use any other excuse to fire people.

I however do think that technically, they're doing a great job when applied properly, but that's just my two cents.

---Replies---

rubén@ruben@friendship.quest (2026-09-21 19:03:07)
@anthropy@mastodon.derg.nz @relay@relay.toot.io @ruben Yeah. I agree the conversation needs nuance. But I also think those aren’t contradictions!

It’s not just worker protection, plenty of ills:
⁃ No human is reading my resume, I actively need to make it worse for humans so it passes that filter
⁃ Bad bosses (not an llm problem) are using them to further antagonize workers and pretend it’s “hard data”
⁃ They’re eroding the open web that created the conditions for them to exist (both in corpus and skilled individuals)
⁃ I believe they’re eroding communities by reinforcing individualism and a disdain of other disciplines through dunning-kruger.

I worry about those things because they’re
Problems now and feel they’re more urgent than the productivity boost. They’re not new problems but they’re supercharged, and why I’d rather put a harder line on “maybe we shouldn’t go there” rather than wait for it to play out.