It's all downhill from here. Well, okay, not so much downhill as it is a matter of figuring out all the edge cases, which, as we data miners know, is always the time-consuming part. I'll even toss out a prediction: we'll spend at least the next ten years working on exactly that problem, until someone comes up with a generalization that works and stuns everyone with its simplicity and common sense.

Philosophical question for you, me, and Dr. Shulgin: how would the DEA respond to a researcher using a non-scheduled halogenated phenethylamine with a radioactive tracer to study 5HT-2 receptor activity in a neuron-brained robot? What if the neurons are cloned human neurons?