Fixing LLM Word Salad: The 'Deslopify' Prompting Trend

sh_reya · x · 2026-08-04

Developer Sh Reya complains about the atrocious default writing style of LLMs, often filled with slop and unnecessary fluff. To fix this, she added a 'deslopify' command to her writing skill to prompt agents into generating better, more natural responses.

Braden Hancock draws an insightful analogy: ICs now act like managers of AI agents. Just like briefing an executive in a big company, you need to provide AI with highly info-dense, clear, and standalone briefs. Setting the guidance to 'talk to me like I'm a sharp CEO who has no project context' yields particularly effective results.

Original post →

More from coding & agent

coding & agent channel →