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.
You could say- - Let's run 10,000 cell cultures in parallel. The graduate student cost is nothing. They don't need to sleep. It's AI, after all. - Mm-hmm. - And with all the properties of this immune cell type, different concentrations of drug, and while it's not a real-world experiment, you can get an indication of what the outcome might be and what might be worth taking a better look at.
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.