Post #4318439
2026-08-01 17:55 UTC
Replies (6)
-
@duckdotexe@mastodon.social 2026-08-01 19:48
@nixCraft@mastodon.social Lol, get fucked. Also, the implication that future models somehow won't have this problem that appears to be inherent to this technology is laughable.
-
@frequencyfault@mastodon.social 2026-08-01 20:06
@nixCraft@mastodon.social it's because of tokenmaxxing. Lol LLMs are being denied valuable new data from humans to learn from and the tech and finance bros can't figure out why their models are collapsing. The models are learning from THEMSELVES and they're stupid.
-
@pa27@mastodon.social 2026-08-01 21:38
@nixCraft@mastodon.social I've seen instances, were on inspection, there's malware in LLM code. At the end of the day, the LLM gets it's stuff from anywhere, but you don't know...
-
@42phi61861@mastodon.social 2026-08-02 04:57
@nixCraft@mastodon.social There is NO other effect possible as ALL remains written by the same Incoigherent human Minds !!! One needs to start at Home First !!! ALL is Cause and Effect !!!
-
@xan@xantronix.social 2026-08-01 18:01
@nixCraft@mastodon.social keep the schadenfreude flowing
-
@doragasu@mastodon.sdf.org 2026-08-01 18:03
@nixCraft@mastodon.social The AI bros: * He did not use a good model. * He did not use the correct plugins. * He must not look at the code. * He did not prompt it correctly (or several variants: did not ask XXX, did not create/update agents.md, did not ask it to "make no mistakes" 😅...) Yet here we are, another project with apparently months of work (and I assume a few hundred dollars at least on tokens/subscriptions) that goes down the drain.