Andrew Huberman· PhD
what if we could utilize artificial intelligence to scan across all drugs and all diseases to find the best opportunities?
Loading…
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.
what if we could utilize artificial intelligence to scan across all drugs and all diseases to find the best opportunities?
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.
So- - Mm-hmm. - we know, in the example earlier, that one lab found increased PD-L1 expression in this one form of cancer, and this drug inhibits PD-1. So therefore, let's make a connection that no one had made yet.
So what I really am bullish on using AI for is not to simulate something that hasn't been done yet, but it's actually to find connections between what has been done.