Yeah I don't think we're getting there in our lifetimes, especially not with how neurobio/psych research has been progressing lately, but I agree it is theoretically possible. And along the way there are many imperfect models that could still be highly useful - I don't think replicable is a binary thing at this level of abstraction.