@NymphUrban The official answer from the LCARS designer is "Whichever one you press will be the correct one" - that's literally what he said when actors were worrying about the same thing.
@OpenAI@sama This new "Potential risk detected" gate in Agent Mode, can't say I'm a fan, it interrupts previously-smooth flows. Would it be possible to disable this from the settings?
Also, Agent Mode being able to use credentials from custom instructions would be a big plus.
@GergelyOrosz Since a lot of people are expressing misgivings about Google over documentation, my instinct when googling is that the docs will be top of the list.
So effectively I'm still checking docs, just indirectly.