Well duh, curve fitting isn’t new, that’s why we try to make predictions before we know the result and try to keep the hypothesis simple. Of course falsifiability isn’t enough alone, but it certainly hasn’t lost its place.
Your comparisons are ridiculous because you’re comparing things which are testable (genetic variances, etc) with hypothetical differences between ancient brains we don’t know the structure of. We still don’t even know enough to make deep comparisons between brains of related animals. Until you can both synthesize and simulate the brain of ancient genomes you have absolutely no idea if you’re on the right track, you can’t know at all. There’s so many different ways a brain can implement the same behavior with so many different unpredictable side effects that you can’t say more than “they behaved in a way that kept them alive long enough” with any reasonable certainty. Do you know at what rate brains have changed biologically? No?