What about this bug: setup refresh: processed N write roots line and no setup binary completed line follows.
2. No command-runner launch is logged for that command, so the command itself does not appear to start.
3. A few seconds later Codex starts the next command. It spawns another helper, which logs the same grant on the same folder.
4. The helpers pile up. Task Manager showed about 45 at once on the morning of Nov 13 and 20 that afternoon. Each used about 1 MB of memory, about 1% CPU and 0.2 to 2.8 MB/s of disk writes.
5. They keep running until the app is quit and the helpers are killed.?
Are you f*cking kidding me. Got fired from OpenAI today.
In my defense, being a $300k human API between Claude, ChatGPT, and Gemini worked perfectly until today...
I wanted to test Fable 5.2's quality with one-shot prompts without human feedback.
Everything in this game is from scratch: models, sounds, textures. ~$550 at API pricing, ~2.7M output and 445M input tokens, heavily cached. Took 8 hours!
Play it here: https://t.co/fazOCcyNIH