I'm seeing that same sequence of stages with my org. In our work, we're currently at stage 3 Bargaining but there are teams here and there leapfrogging depression straight to acceptance. Exciting times if you ask me. The key is really focusing on the test and verification approach imo. Automate what you can, measure/log or proceduralize what can't be automated.

Some people seem to absolutely refuse to understand that a lot of open source is done for the authors' personal enjoyment. There is absolutely no reason they have to accept AI contributions, no matter how much the author of this tweet wishes, and plenty of excellent reasons why they might not (e.g. they actually enjoy programming and the practice of programming). This would be true even if AI generated code were completely indistinguishable from human authored code, which it is not.