OpenAI's official prompts fix GPT-6.2's robotic tone, stalling, and slacking in agents
alexcovo_eth · x · 2026-11-19
After GPT-6.2 launched, users complained it sounded inhuman, got stuck easily, and slacked off. OpenAI's official docs ship targeted prompts for all three:
- Robotic tone: "Rewrite for the user. Don't copy or paraphrase your reasoning... Don't bring shorthand, jargon, or abbreviations from reasoning, tool calls, or instructions to sub-agents into the user-facing reply."
- Self-blocking: Only stop and ask the user in four cases — substantially changed task scope, genuinely stuck, a decision only the user can make, or an irreversible/destructive action. Otherwise keep going; don't wrap up with "suggested next steps / a plan."
Both prompts come straight from OpenAI's official GPT-6 usage documentation and can be dropped into system prompts today.
Related event: OpenAI Shares Prompt Fixes for GPT-6.2's Three Big Complaints(2 posts)→
More from coding & agent
- Ultrafast mode review: great for math proofs, not recommended for coding — gnukeith · 2026-11-19
- Speed up CI for AI-written code: caching, faster runners, test sharding — rseroter · 2026-11-19
- Agentcard raises $8M to let agents buy from 5,000+ stores via Stripe Link — jeff_weinstein · 2026-11-19
- hstack open-sources one-command self-hosting for the Hermes agent — TheMoonMidas · 2026-11-19
- 16 AI models ran produce buying for 90 days — only 3 beat a simple reorder formula — ycombinator · 2026-11-19
- DeepLearning.AI Launches Short Course on Tabular Foundation Models as Agent Tools — allenainie · 2026-11-19