AI is "comically good" at detecting small earthquakes–here's why that matters

1. These are smaller earthquakes, monitoring them is neat but not as important as the larger ones.

2. This is less "AI" and more "machine learning" which essentially is just applied statistics. It's about as "intelligent" as a digital guitar pedal.

3. FFT analysis is just as old as the field of realtime computing itself. The ML research and application is almost the easy part given the cost of GPGPU compute today.

It's cool tech, but it won't back up your "AI is a panacea" argument if you try to cite it.