[HamSession]

It is good to note that Lauder's principal is a statistical argument for degrees of freedom in a system. The Landauer principal on a statistical averaging of this effect, meaning you will sometimes be able to store a little more information.

I think you have your calculations off for the amount of energy required to store a human brain. Assume 8.6 x 10^10 neurons (more accurate number as its simple density measurement) each holding a 64-bit weight which gives us 5.5 x 10^12 bits. This multiplied by the lower bound e.g. k(1)ln2 or 9.56 x 10^-24 J, gives us 5.3 x 10^-12 J which is lower than the energy required to type a letter on this keyboard. Now even if we increase this by body temperature say 37 C we get 310.15 K or about 1.7 x 10^-8 J. This is of course a lower bound on the problem but even if you pump up the number of neurons or weight value you still get less than a joule of energy required.

[scythe]

>each holding 64 bits

There is no way in hell you're going to store the entire physical state of a neuron in anything near 64 bits. The question is not "information stored in the brain". The question is "information required to recreate the brain". In fact a single neuron has many more than 64 synapses, each of which itself has a highly intricate structure. All of the information describing this structure must be retained intact in order for the image brain to be equivalent to the substrate.