Looking forward to the postmortem over this outage.
OpenAI has really gotten terrible overtime with maintaining uptime with the likes of Google or Meta. Perhaps it is this pervasive lack of experience with site-reliabililty engineering that is missing at OpenAI.
But once again, local modals such as DeepSeek-R1 don't need this sort of waiting or multiple outages.
