@mattpocockuk: LLM's are notoriously bad at following negative instructions That's why code review works so well, it turns all the neg…

X AI KOLs Timeline News

Summary

Matt Pocock observes that LLMs struggle with negative instructions, and explains how code review converts those into positive, actionable directives.

LLM's are notoriously bad at following negative instructions That's why code review works so well, it turns all the negative instructions into positive ones:
Original Article
View Cached Full Text

Cached at: 07/31/26, 02:57 PM

LLM’s are notoriously bad at following negative instructions

That’s why code review works so well, it turns all the negative instructions into positive ones:

Matt Pocock (@mattpocockuk): That’s why you do it during review

“Don’t do these footguns” - BAD “Find and eliminate these footguns” - GOOD

Similar Articles

@jakevin7: Sharing a god-tier review prompt methodology. The LLM self-correction survey 'When Can LLMs Actually Correct Their Own Mistakes?' concludes that without reliable external feedback such as test results or tool outputs, a model relying only on self-reflection often cannot steadily correct errors...

X AI KOLs Following

Shared a prompt methodology based on LLM self-correction research, emphasizing that self-checking is limited without external feedback, and recommending progressively enhanced prompting strategies such as adversarial review.

Do as I Say, Not as I Do: Instruction-Induction Conflict in LLMs

arXiv cs.CL

This paper investigates the conflict between instruction-following and pattern completion in LLMs, finding that instruction-following is brittle under induction pressure and varies widely across models, with output diversity being the primary factor for robustness.

Quoting Bryan Cantrill

Simon Willison's Blog

Bryan Cantrill critiques LLMs for lacking the optimization constraint of human laziness, arguing that LLMs will unnecessarily complicate systems rather than improve them, and highlighting how human time limitations drive the development of efficient abstractions.