IANAE but, transmitting power from cables has several drawbacks 1) Anything that adds weight to the cable is a very bad thing. 2) the power loss due to extreme length. 3) Redudancy directly affects 1 and 2. 4) 1 and 2 are mutually exclusive.
And as far as wireless power transmission goes, if its focused i think the powerloss is quite acceptable over those distance relatively to other means. The redudancy is exceptionnal too, you can have multiple beaming stations on the ground and/or in the air, ready to send their own beam in case the primary goes down.
The power of chaos (call it randomness or QM) and really, really long spans of time. Plus a lot of evolutionary stresses. Reusing simpler/earlier organisms as basis for interaction and possibly sustenance for more complex ones is only natural. Throw in a few well chosen mass population genetic bottlenecks and there ya go
Training is crucial. At the moment we have the tools to create a virtual environments in which we can breed populations of incremental complexity (genetic/coding-wise) through evolutionary pressure. As we learn to create these training and bottleneck methods for simpler "life" forms, the computing capacity needed for complex organisms will probably be there by the time we perfect such methods.
My guess is yes, about 10 years for an intelligent ANN. Also, when we do reproduce the brain it will have seemingly limitless potential for medicinal uses. It will also bring a terrifying potential for abuse. Such mastery of human intelligence could lead to disaster...
Karl's version of Parkinson's Law: Work expands to exceed the time alloted it.