The headline is broadly defensible, but the qualifications matter. Effect sizes vary by population, the strongest claims rest on shorter trials, and credible voices push back on how it's typically framed.
The amount of usage is huge. Um so the more you use AI, the more likely this is to happen to you. But I I kid you not, I'm really I'm not trying to be alarmist. As a psychiatrist, when someone comes into my office, I ask them, "Do you use meth?" because I'm trying to assess their risk of becoming psychotic based on something that is not like schizophrenia or type one bipolar disease. Now I'm starting to ask people, do you use AI? How much?
Every Sunday: the week’s new conflicts and verdict changes — and nothing else.
Native comments, Twitter mentions, and Reddit threads about this claim — surfaced together so the conversation isn't fragmented across platforms.
Bookmarking — the dossier-vs-overview split is the right call. Most of the time I want overview; sometimes I want receipts.
Would love a "what would change this verdict" RSS feed. Sign me up if it exists.