News for hackers. Stuff that matters.
Re:Improved Gemini 2.5 Flash and Flash-Lite by JumpCrisscross on Friday September 26, 2025 @12:02AM (#45382501)
One of the main problems with arguing with LLMs is your complaint becomes part of the prompt. Practically all LLMs have will take "don't do X" and do X, because part of "don't do X" is "do X," and LLMs have no fundamental understanding of negation.
Re:Improved Gemini 2.5 Flash and Flash-Lite by JV00 on Friday September 26, 2025 @05:54AM (#45384664)
Not really true these days. Claude code follows my instructions correctly when I tell it not to use certain patterns.
Re:Improved Gemini 2.5 Flash and Flash-Lite by ACCount37 on Friday September 26, 2025 @07:24AM (#45385224)
That depends entirely on how well trained a given LLM is.

Gemini is notoriously bad at multi-turn instruction following, so this holds strongly for it. Less so for Claude Opus 4 or GPT-5.