To build on this, as people still commonly ask "but surely we can predict earthquakes": even if we had perfect precise knowledge of the stress field in the crust (we don't, it's a rough approximation even with measurements), the literal process of an earthquake is one of fracture and failure.

The difference between fault slip causing a tiny earthquake and a major earthquake is whether the crack/slip propagates or not. And that is entirely dependent on the exact micro and macrostructure of the rock around the initial slippage. Even atomic level imperfections in crystal structure could be the difference in propagation Vs not.

(In fact even the initial slip is determined by the relatively weakest point to accommodate current stress field)

Therefore, to even begin attempting prediction, we need atomic-level understanding of the crust and stress field dozens of km deep. Completely impossible.