Yes, full access

I now only run agents in full-permission mode. 7 months ago, I would never have done this.

It’s worth reflecting on this behavioral change as a way to reason about trends over the past 7 months.

I run agents in full-permission mode because:

1. It’s faster.
2. I have more trust in the models.
3. I haven’t bothered to run remote agents or figure out sandboxing.

I get work done faster because agents are not blocked on permission gates. Permission gates effectively serialize work.

I have more trust in the models these days. Models and harnesses have improved. I can prompt with less care and with larger scope and still have decent, functional results. The underlying system is often a mess if not tended to over time, but it’s good enough for early projects.

I haven’t bothered to sandbox or run cloud agents. I should. I don’t know why I haven’t done this other than I haven’t felt the urgent need. I’ll try to do this more often. It’ll also be easier to run background tasks.

I don’t think I’m alone in defaulting to YOLO mode. It’s too convenient to be rare. I think there’s appetite for a default, well-understood sandboxing effort. The sandboxing solution here will be a free, open standard, not something that can be monetized. Cloud agents will see good adoption but, depending on their economics, may not be a default running-mode for agents.