> If there is anything to learn from the history of science, it is that breakthroughs happen via better or new theory and not by brute-force compute

that might just be an illusion. humanity as a whole applies brute force by just trying all the reasonable new theories until one of us discovers something that works and declares it a breakthrough