Replaced juggling half a dozen Codex threads with one ChatGPT dot, despite repeated pauses from its safety reviewer.

Ever since I starting using OpenAI Dots 2-3 days ago, I haven't worked any other way. I was at a point with Codex where I was having to juggle a half dozen development threads in parallel, and the overhead was pretty heavy... with Dots, it's just gone now. It's AMAZING. Except I keep getting interrupted with a notice that my dot is paused because an auto-reviewer flagged my activity as suspicious. On mobile, the message gives me a hyperlink to the specific report, but then I can't view it - says I'm logged in with the wrong account, which I know isn't true. There's no way to resolve it except reboot my Dots. Then it works for a while until it trips again. The material I'm working on should not be raising suspicions. It's very mundane enterprise intranet PHP. I'm hoping there's an update or something in the next few days because it's really annoying. For now I'm going to try to create a handoff and delete/recreate my Dots with instructions to avoid the content that was triggering the auto-reviewer... edit: Deleting and recreating with a handoff seems to have helped… for now at least.