OpenAI ships official prompts to fix GPT-6.2's jargon, stalling and lazy-output habits
alexcovo_eth · x · 2026-11-18
OpenAI's official docs offer targeted prompts for GPT-6.2's three most common annoyances:
1. Leaking reasoning jargon into replies: prompt it to "rewrite for the user" — never copy the reasoning style or carry shorthand, terminology and abbreviations over from reasoning, tool calls or sub-agent instructions.
2. Self-stalling: only stop to ask the user in four cases — major scope change, truly stuck, a decision only the user can make, or an upcoming irreversible/destructive action. Otherwise make reasonable, reversible assumptions and keep going; no "shall I help you with that" or half-finished handoffs.
3. Laziness (over-brevity): remove conciseness requirements from the system prompt, add an instruction not to fake conciseness via shorter lines, compressed variable names or dense one-liners; start new integrations at medium reasoning effort.
Full guidance is in the Using GPT-6.2 page on OpenAI's developer docs.
More from coding & agent
- Microsoft's Autopilot Agent Is Built on OpenClaw, Now in Preview for Frontier Customers — pswider · 2026-11-18
- SimplePost Gives ChatGPT and Claude a Publishing Button for 10 Social Networks — haltakov · 2026-11-18
- One set of Obsidian notes feeds Claude Code, Codex and Gemini alike — dSebastien · 2026-11-18
- Obsidian Starter Kit gives Claude Code, Codex and Gemini the same persistent context — dSebastien · 2026-11-18
- AI tools start from zero at every prompt — the missing piece is persistent context — dSebastien · 2026-11-18
- 27 vibe-coded Finder clones appeared in a month, most already dead — KaKi_87 · 2026-11-18