I really do think more of us should be focussing on the "reasonable person" approach to this tech. Yes, it feels like magic. Yes, it's an amazing technological leap forward. But going all-in on the hype and seeing every world problem as a nail in need of the "AI" hammer doesn't feel like the best way to treat it.

Methods like those in this post that leverage the real-world useful-right-now aspects of GPTs are genuinely useful to a lot of people. I'm pretty sure that treating it "reasonably" and with a healthy does of scepticism is what's going to yield the greatest productivity boosts.