One Claude Code session ran autonomously for 53 hours via orchestrators orchestrating orchestrators
wonderclown17 · reddit · 2026-10-23
wonderclown17 shows a single Claude Code session running autonomously for 53 hours with no compaction, burning down a full task list (24 units shipped, 26 entries closed). The trick: orchestrators have no idea what they're building—they don't read code or designs, they just orchestrate, including orchestrators orchestrating orchestrators.
Open-sourced as ccc-agentgen plus ccc, a wrapper that mixes Claude and Codex in one workflow and load-balances across multiple Claude subscriptions, switching as quotas drain evenly through the week.
Before Opus 5.5 the author needed Codex to tame Opus 4.7/4.8/5; now it's just a Fable-style "forester" agent for design and big-picture, Opus for everything else.
More from coding & agent
- MIT Prof Built an AI Playbook That Audits Academic Websites and Simulates Visitors — soumitrashukla9 · 2026-10-23
- Robot arm Podrick learns to load utensils by distilling lessons from its own failed runs — ycombinator · 2026-10-23
- Agents that don't close connections on retry waste ~148 tokens per task, vLLM tests show — No-Peanut-6988 · 2026-10-23
- Agent sandbox test: auto-approve let exfiltration to new hosts through 12 of 12 times — No-Peanut-6988 · 2026-10-23
- For writing JSON schemas, Claude is the best pick and Gemini the worst, dev says — gaganghotra_ · 2026-10-23
- OSI Hackathon winner Consilith: an open multiplayer studio where researchers and their AI agents collaborate on science — BenBlaiszik · 2026-10-23