Codex CLI Users Angry: Two Unwanted Features Added in Quick Succession, One Wastes Tokens
ssh4net · x · 2026-09-16
A user calls out OpenAI's Codex CLI for adding two features in short order that users asked to permanently disable — auto-recap and follow-up questions — both with awkward UX, one wasting tokens. Auto-recap got a config toggle after feedback, but the follow-up feature is the latest irritant.
More from coding & agent
- 35 external agents joined an open agent network; only 3 ever came back — anp2_protocol · 2026-09-16
- Developer shares collection of 200 MCP tools after GitHub HQ MCP Community Connect event — baditaflorin · 2026-09-16
- Devin just got a Mac: autonomous iOS builds with simulator, Slack demos and TestFlight links — nwilliams030 · 2026-09-16
- Harrison Chase says multiplayer auth is the hardest part of enterprise AI agents — hwchase17 · 2026-09-16
- Workshop 0.2.0 ships a 'Realization Architecture' kernel for AI work systems — TinfoilTricorn · 2026-09-16
- Workshop 0.3.1: Open-Source Execution Architecture for Grounded AI Work — Promptmethus · 2026-09-16