ChatGPT silently drops context in long chats; here is the fix
Ok_Negotiation_2587 · reddit · 2026-09-01
ChatGPT does not warn when a conversation exceeds its context window; instead, the oldest messages silently drop out. The model continues answering confidently based on the remaining context, unaware of what it forgot, which can lead to it contradicting earlier decisions or constraints.
Symptoms:
- Re-asking questions answered at the start.
- Contradicting early decisions (names, budgets, constraints).
- Summaries missing the first third of the conversation.
The Fix:
When a chat gets long, prompt the model to summarize all decisions, goals, constraints, and open items as a briefing for a new chat. Open a fresh chat, paste the briefing, and continue there. Do this proactively before symptoms appear to ensure no critical details are lost.
More from Apps
- Made by Hand turns game crafting into a real woodworking job, one saw cut at a time — WellsLucasSanto · 2026-10-18
- Dev builds Lex Fridman podcast archive site with 506 episodes and Chinese AI summaries, open-sourcing soon — vista8 · 2026-10-18
- Practitioners say Google AI Overviews is quietly getting very good as an answer engine — ricklamers · 2026-10-18
- Claude skill generates hand-drawn interactive diagrams that teach kids how things work — CatAstro_Piyush · 2026-10-18
- Meta wearables VP runs his AI agent across 3 phones and 23 pairs of glasses — nikola_mr64990 · 2026-10-18
- Indie dev's Qiaomu Obsidian plugin suite tops 20,000 downloads, Agent plugin surging — vista8 · 2026-10-18