I've always been a bit disappointed that evolution basically fixed all today's high level body plans about 500 million years ago. Lots of evolution since then h

I've always been a bit disappointed that evolution basically fixed all today's high level body plans about 500 million years ago. Lots of evolution since then has involved turning the crank on incremental and higher-level improvements (most famously, intelligence), but I don't see a lot of innovation in basic biomechanics. It reminds me of how everyone is willing to experiment with new neural network training algorithms in the software world, but you need tremendous resistance trying to change parts of POSIX to suck less. It's as if in both nature and computing, there's some conservation in the amount of experimentation we're willing to tolerate.