6-day, 31-PR chat compacted 27 times still works: dev argues context compacting is basically solved
holdenmatt · x · 2026-10-08
Pushing back on 'I'll never use a compacting harness again', developer holdenmatt argues context compacting is largely solved on newer OpenAI models. He shares a running Codex/Astra chat that's lasted 6 days and produced 31 PRs, compacted 27 times along the way—and 'it's fine?'—a stark contrast to how older models degraded after compaction.
More from coding & agent
- Musk promotes Grok Bot: X Corp.'s AI coworker app that logs in and works your tools 24/7 — elonmusk · 2026-10-08
- Demo is not delivery: what enterprise AI adoption actually takes — sujingshen · 2026-10-08
- DeLM's decentralized multi-agent system runs 2.49x faster, but MAS evals pick wildly different metrics — jyangballin · 2026-10-08
- Musk endorses Grok bot + Cursor cloud agents combo as a major productivity boost — elonmusk · 2026-10-08
- Dev: in the AI coding era, tabs-vs-spaces and style nitpicking never mattered — facontidavide · 2026-10-08
- RunningTab: environment-side task ledger consistently beats in-model tracking across 3 benchmarks and 3 LLMs — RexDouglass · 2026-10-08