Dear OpenAI Team,
First off, love the product—you’re doing amazing work. Now, let’s talk about how you’re making my life unnecessarily harder.
I get it: safety, accountability, yada yada. But for users like me who know what we’re doing (and frankly don’t want to click “Trust this action” for the millionth time), could you please add an “Always Trust This Host Until Revoked” option?
Right now, the “Always Allow” button doesn’t actually do anything meaningful. If I still have to confirm every time I want to perform an action, what’s the point? If it’s just a one-time thing, rename it to something honest like “Allow Once and Pretend I’m Helping.”
It’s 2025. If I can trust my fridge to order groceries, I should be able to trust my assistant to update my memory without needing a permission slip every. single. time.
Think of the possibilities: more efficiency, less user frustration, and fewer support tickets from people complaining about this exact issue. Win-win, right?
Thanks for considering. Hope you take this seriously (and not just because I’ll keep spamming the request until it’s implemented).
Cheers,
Rup