A continuing investigation into OpenAI product behavior, technical failures, support practices, and accountability.
My GPT-5.5 archive documents premature declarations of progress, repeated failed repair loops, capability confusion, and a system that kept transferring verification work back to me.
I spent six days diagnosing a catastrophic Codex Desktop failure, found the 71 GB conversation that caused it, and fixed it myself. OpenAI Support documented the evidence, refused to move the case forward, and…
New from NOMOTO MEDIA
Occasional email with newly published stories, investigations, films, photography, music, and podcasts. No daily newsletter.
Support independent work
NOMOTO MEDIA publishes essays, investigations, fiction, audio, and films without a paywall. If the work is valuable to you, help support the next piece.