>>As soon as the simulation behaves like a human being, it changes behavior when observed.
>
>
>I think we're already miscommunicating.
>
>The idea is to create a model, a simulation, of a bunch of fundamental particles and quarks.
>
>Through that, build up atoms, molecules, cells, and eventually, a human being.
>
>So there is a simulation of physics, and within that simulation we've put together a human being.
>
>With me so far?

So far yes, but I went further. Eventually, the wet dream of all Harry Sheldons of this world is to build a simulation of a society and thus predict the effect of whatever they intended to do, which is usually to get rich, famous or in power.

Which will never work, because as soon as you change something in the environment, the people will react in unpredictable ways. Or, to predict that, you would have to know them so well that you wouldn't need a simulation.

If you meant this to stay at the biological level, to, say, try new drugs on the simulation - you'd be wrong only inasmuch as the mind has ways to alter the works of the body, so you could draw a much better line between organic and somatic. Perhaps the line would be much wider than we thought, i.e. that line would be the margin of error, for cases where it predicts certain behavior which then doesn't happen because in real life people have minds and all kinds of influence - placebos, psyhosomatic illness/self-healing, not healing when it should, healing when it shouldn't etc etc. Getting to know that kind of stuff would be a worthy result in itself.

>> It may even develop that sense of being watched that we all get at times. And then it doesn't behave the same anymore. It starts acting, and interacting with the observer.
>>
>>Which is why I tend to discard about 90% of psychological "research". Simulation would add another set of quotation marks somewhere.