A 3 a.m. handoff to Codex
OpenAI says one of its engineers used a ChatGPT “dot” to help triage an on-call incident in the middle of the night. In an Oct. 16 post, OpenAI Developers said the engineer, identified as Ali, put his dot to work at 3 a.m.
According to the company, the dot investigated the issue and started a Codex thread. Ali then steered that thread from his phone in the ChatGPT app. OpenAI did not disclose the cause or severity of the incident, what the investigation found, or whether the dot helped resolve it.
The account is a company-reported example rather than an independent test. It shows the workflow OpenAI wants dots to support: an assistant can gather context and delegate coding work, while a person remains involved and directs the next steps.
What OpenAI says a dot can do
OpenAI described the underlying Codex handoff in an Oct. 9 Developer Community announcement. The company said a dot can start work in Codex, continue an existing thread and draw on a user’s ChatGPT conversations, Codex threads and automations. Dots can also be created in the ChatGPT mobile app on iOS and Android.