"The nightmare we should be focusing on is institutional: generative AI breaks long-standing correlations that society uses to infer things like effort, sincerity, authenticity, and credibility."

Spot on, I believe. It might be a good idea also business-wise; solving the authenticity issue alone might open new opportunities now that the "big platform economics" are dying (by hypothesis).

I really like Tufekci’s framing. The debate still feels weirdly powerless. You are either "pro" and in love with text parrots, or "anti" and dreaming of blanket bans, like you could outlaw 2+4 or 43x24. New techniques do not get uninvented. They just show up and stick around.

The real question is not "should AI exist?" It is "how do we keep our democracies sane" once cheap, scalable tools start melting effort, trust, and credibility. That part feels fixable.

Let’s get practical. Phones out of the bedroom. Kill infinite scroll. Notifications on a diet. One no-screen evening a week. Easy wins. And bigger picture: how do we relearn distrust without going full tinfoil? How do we stop asking screens for comfort and start asking them for facts, even when the facts are boring, ugly, or inconvenient? Screens are amazing at dopamine. They are bad at truth. The good news is we can design around that.