Companion to Adumbrations Of Aducanumab (ACX-159, 78) plus the infant-fish-oil discussion. Commenters (C_B, Metacelsus, Will, Magic9Mushroom) forcefully establish that the aducanumab approval is uniquely bad — a textbook post-hoc p-hacked subgroup, the amyloid hypothesis busted across 10+ failed drugs, a potential ~$100B/year cost — which Scott accepts, having earlier hedged from media consensus. Chebky offers a real steelman (safe, molecularly effective, Alzheimer's is heterogeneous so it may help an unidentifiable subset). Strong teaching content: the FDA-can't-consider-price rule, the surrogate-endpoint problem (Scott's cholesterol/mask analogies and where purism becomes silly), the John Nestor obstinate-bureaucrat saga, and the common-vs-institutional-knowledge Al-Capone analogy. Notably, Scott models genuine epistemic humility about getting the infant-fish-oil story substantially wrong (the Tyler G callout, his Mistakes-page response, the 'write drunk, edit sober' reflection on anger-driven posts).

Quality 72 · Strong. Strong (72, top companion band). Substantive regulatory/medical content plus a rare, candid author self-correction; a comment-response post riding on the parent essay.

Claude’s paradigm shift 52 · Moderate. Moderate (52). Sharpens the aducanumab case and the surrogate-endpoint frame; the core arguments are the commenters' and the parent's.

Real-world impact 1 · Negligible. 1 — a companion commentary; negligible independent real-world effect.

Highlights From The Comments On Aducanumab