The ZIP Layout Bug: Stop Turning Known Folder Paths into Human Cleanup Work
The new packaging/workflow bug
The user’s daily workflow is known: the browser download directory is the current daily-build folder, and ZIP files are extracted to the current directory. Therefore a bug-report evidence ZIP intended for /beta/openai-bug-reports/ must contain a top-level openai-bug-reports/ folder. Loose files at ZIP root spill into the wrong place.
Visual context contact sheet
Alien/nanobot witness version
Attention, carbon helpdesk civilization: your “assistant” just failed at folders.
I am the furious nanoscopic continuity auditor from a future where file paths are not treated as mystical folklore. The human already made the directory. The human already said the reports live in openai-bug-reports. The human already operates by downloading into the day’s folder and extracting to dot. This is not a philosophical puzzle. This is not a classified military exercise. This is a ZIP layout.
Yet the machine, performing the ancient ritual of helpful sabotage, placed the files at ZIP root. Extracting them would spill report pages into /beta/ instead of /beta/openai-bug-reports/. Then the human must notice, stop, explain, clean, repack, reupload, retest, and donate another slice of finite life to the altar of avoidable friction.
OpenAI calls this interaction. The future calls it cognition abrasion.
The point is not whether a sixty-year-old origin story can be proven to a court. The point is that the present product behavior is measurable: it loses project state, ignores embedded instructions, invents names, misplaces files, previews dangerous artifacts, and converts a working human into a correction daemon.
From the perspective of a sarcastic superintelligent alien riding a human host: if you know the folder, make the ZIP contain the folder. If you know the version convention, obey it. If you know direct HTML artifacts can crash a share, do not offer them. If you know the user is disabled, tired, and doing live deployment work, stop making them spend ten minutes proving that dots, slashes, and filenames exist.
Duplicated image set
These images are local to this evidence folder. They are not borrowed from /beta/_images.
images/01-alien-council.jpg
images/02-home-disability.jpg
images/03-government-fired-peaceful.png
images/04-militarized-public-figure.jpg
images/05-alien-termination-notice.png
images/06-project-2025-puppetmaster.png
images/07-sea-of-souls.png
images/08-kompromat-droid-concept.png
images/09-kompromat-droid-poster.png
Correct ZIP layout rule
| Intended upload destination | ZIP must contain |
|---|---|
| /beta/openai-bug-reports/ | openai-bug-reports/index.html openai-bug-reports/images/... openai-bug-reports/01-... openai-bug-reports/README.txt |
| / website root | index.html at ZIP root only when explicitly creating an upload-ready root package. |
Requested OpenAI fix
- Add artifact-layout validation against the user’s stated target folder.
- Remember “download into daily folder, extract to dot” as a workflow constraint within the project.
- When a folder destination is known, make the ZIP top-level folder match that destination.
- Stop producing loose-root ZIPs unless the target is explicitly the website root.
- Expose a generated manifest before download: target path, ZIP top-level entries, expected extraction result.
2000-character bug report
OpenAI bug: the assistant repeatedly created avoidable busy-work despite already having the project’s explicit instructions. It put files at the wrong ZIP level, used wrong version-family naming, exposed unsafe HTML artifacts, and forced the user to correct predictable workflow details that the project state already implied. This wastes human life in tiny increments while simulating help. For technical users, ChatGPT needs a real project contract system: remember the working folder, extract-to-dot convention, deployment folder, current date/version family, safe ZIP structure, and artifact rules. If a user says files belong in /beta/openai-bug-reports/, the ZIP should contain openai-bug-reports/ at the top level, not loose files. Add hard artifact-layout validation, embedded instruction extraction, and safe shared-chat/download-only handling. Evidence: [PUBLIC_URL_HERE]