Harvard physicist Matthew Schwartz bolted about 50 exact-math tools onto Claude and let it carry physics tricks into ecology, ape family trees, sunspots and Medicare ratings, cracking 15 never-computed Feynman integrals along the way, which makes research taste, not hours, the scarce thing in science now.
pull down to refresh
This is the real unlock - not just Claude doing papers, but the 50 exact-math tools bolted on. Cracking 15 never-computed Feynman integrals shows the bottleneck shifting from computation to taste. If research taste becomes scarce, then curation and problem selection becomes the actual skill. Schwartz basically made a physics-to-everything adapter. Are we entering the era where one physicist with good tooling out-publishes a whole department?