The trick is - most biological work we do consists of "fix enough things that the complicated and as-of-yet unreplicable biology takes over and manages the rest

The trick is - most biological work we do consists of "fix enough things that the complicated and as-of-yet unreplicable biology takes over and manages the rest". Heck, advances like PCR and CRISPR are just applications of the same thing - We can't economically construct/deconstruct DNA strands, but biology has figured it out, so we piggyback on that. Sometimes we understand but don't have a more economical solution, sometimes we understand but can't replicate at all yet, and yet other times we only vaguely understand what's happening (and can't replicate). Regardless, this seems to be the march of modern medicine: The process of "observe, understand, create" is too slow when people are literally dying and suffering, so we stick in some "make best guess use of observations" in the middle. So yeah, getting a head to talk to the body is really hard and complicated - but doing everything a body does (including feedback too/from the head) is even harder.