Dev pushes back on AI misinformation: frontier models all depend on prompting, real gap is safety controls

gerardsans · x · 2026-09-25

The author pushes back on sloppy AI journalism with a technical point: frontier AI doesn't work without prompting — agents and multi-agent harnesses all require prompt engineering. What's actually missing, he argues, is monitoring and better safety controls, which fall on developers deploying these systems at the labs. He laments that AI coverage is riddled with misinformation because editors lack this basic understanding.

Original post →

More from AGI Musings

AGI Musings channel →