>If there's anything it is like to be a worm, THAT'S the hard problem.

Well, the problem is to give some definition of "what it is like to be X" we can cleanly separate from the "mere" sensorimotor statistics of X.