Post #4014899

2026-07-22 14:59 UTC

I love the argument of "any mistake an LLM can make can be done by a human". 1. That is relative. 2. Unlike LLMs people can get held accountable for commiting those mistakes. 3. That's still not a great argument to justify it as a "tool", a tool should not make mistakes, it should do what you expect it to do, and we both know LLM output is not 100% deterministic.

Replies (4)

@ainmosni@social.ainmosni.eu 2026-07-22 15:03

@NullTheFool@mastodon.gamedev.place Also, I'm just going to say that it's incorrect. Humans make different kinds of mistakes, as there is still a certain level of understanding there. Also, we're actually good at intuiting the kinds of mistakes humans make. LLMs don't make mistakes, they extrude the wrong text, it's not a mistake, because they don't have any cognition. And because there is no cognition, the kinds of errors are harder to catch in review, because they happen in ways that are not intuitive to humans. This is also exactly why machine learning can be a great addition to humans in detecting things, because what either detects is very different.

Open ##4014898
@VANTABlack2000@mastodon.gamedev.place 2026-07-22 15:22

@NullTheFool@mastodon.gamedev.place People can LEARN from mistakes. LLMs will just forget in 2 seconds.

Open ##4275486
@loucyx@mastodon.social 2026-07-22 18:30

@NullTheFool@mastodon.gamedev.place I love when they say “but humans make mistakes too”, like that somehow justifies adding more error prone factors to the pipeline. Like yeah, humans make mistakes, but if an LLM adds mistakes ON TOP of the human mistakes, that’s WORSE!

Open ##4275491
@Linebyline@mastoart.social 2026-07-22 21:57

@NullTheFool@mastodon.gamedev.place At the risk of stating the obvious, #2 is very much the point (if there *is* a point beyond sailing into the sunset with investors' money once the bubble pops).

Open ##4278677