I think Cursor Cloud just takes up a lot of assigned compute. Grok bot has a computer, and I do not think it needs to purposefully run the same piece of code on a different container. I am falling short of compute because of this choice.
Hey @bot, make a note that you are to skip Cloud. Since you have the Github repos with you and can run all the tasks on the computer that you are assigned.
Also use Grok models by default. No need to use any other frontier lab model.
Make a note.