Baruch, the flaw in your argument is that a real intelligence, artificial or not, has to work in the real world. The reason we have intelligence is not to be intelligent, but rather we're intelligent because our intelligence helps us deal with problems in the real world, and help us survive (to certain values of "help" and "survive"). Said problems tend to be macroscopic in nature — it doesn't really matter much what the individual molecules or atoms or subatomic particles are doing in the system presenting us the problem, as such individual deviations average out over the untold bajilions of particles that constitutes the system.
Remember, people find QM baffling precisely because we just don't see it on the level we operate on day-to-day. Our intelligence simply isn't built around trying to make sense of it, because under normal circumstances QM simply doesn't enter into it — the quantum mechanical nature of the universe is at the large scale blurred by its thermodynamical nature.
In the real world, not the world of a theoretical undisturbed quantum system, there's thermal noise and outside interference, and as such any system that works in the real world has to be robust to these disturbances. An intelligent system that is too frequently prone to making inappropriate responses to consistent situations due to thermal noise or interference will find themselves quickly weeded out in favor of more robust systems.
In short, a system you describe that is so equisitely sensitive to the exact state of the system as it could not be acceptably emulated by a computer would have behavior that is on any practical level indistinguishable from true randomness, and wanton randomness in life tends to be disfavored and selected out.
We operate at a temperature of around 300K, and that's an appreciable amount of noise introduced into the quantum systems making up our thoughts. It translates into about ~25.8 meV, and any state that differs in energy by less across some scale of interest would be utterly lost in the noise, and each brain state that matters (macrostate) is represented by an ensemble of untold fucktons of actual quantum states that may be taken (microstates). Indeed, the degree of coarseness actually implemented in the brain is not single electrons, single microtubles, or even single neurons, but most likely a few tens of neurons, each composed of around 1014 atoms.
The statement that life/intelligence lies in the undescritized continuity is therefore very much premature, and given what else we know, highly suspect. Evolution would absolutely disfavor biochemistry that is so sensitive to thermal noise, where the border between life and death is so razor thin. Indeed, evolution would much favor biochemistry that works to stay as far the hell away from those borders as practically possible (homeostasis). Same with intelligence; evolution would disfavor intelligence so sensitive to practical randomness as to keep their behavior squarely within predictability, and thus respond reliably to similiar stimuli. This is also borne out by psychology experiments, which have revealed that we are actually remarkably predictable.
You can't rely on quantum mechanics to save us from the implications of AI. That is just Deepak Chopra crapola. There's no evidence or research that supports the notion that quantum mechanics or "analogue computing" is vital for our kind of intelligence.
