Yes agree that mining hardware that is optimized for this pow won't replace completely nvidia gpus which are much more versatile etc.
Also the pow is kept deliberately very simple (conceptually), just deep rounds of ternary matmuls (same kind of computation used for inference of 1.58bit llms).
Also activation layers are different (ai usually uses multiples types like relu/gelu etc, here the pow is just using a mod 2 nonlinear layer).
Intention is to make the type of computation similar enough that optimized miners would be ai-capable and vice versa. (for ex non-gpu hardware like groq would be very efficient at mining this pow as well, due to the type of computation it's good for: heavy matmuls, high bandwidth memory)