Post #4143110
2026-07-27 17:05 UTC
and that's one of the problems with the LLM thing - damn fool machine will not contradict people's plans, but will keep yes-anding, so when they decide to implement something in a bespoke and bullshit manner the fool thing enables their nonsense instead of pointing to any actual resources on how to do the thing correctly and maybe discussing concepts like "hey maybe have a risk analysis from an actual professional on the implications of making this shit public"
Replies (1)
-
@munin@infosec.exchange 2026-07-27 17:41
like, if you look hard and trace back enough, you can sometimes figure out the jumping off point where things when sideways and they went from a 'normal' idea into the depths of candyland's circus peanuts factory and started trying to 'optimize' for that shit, but it's -exhausting- to try to think it through