How much computing power is needed to clone your brain? The human brain has about 86 Billion neurons. However, most of the neurons are in the cerebellum, like the relatively small granule cells, and may not be needed for consciousness or a reasonable approximation thereof. After all, a sperm whale has a 7.8 kg brain while a human brain is only 1.4 kg, so much of the brain must be dedicated to controling other parts of the body. The cerebrum may only have 20 Billion neurons. The cerebral cortex of the cerebrum has even fewer, and many are dedicated to controlling movements, processing sensory information, and other physical functions. As few as 10 Billion neurons may be involved in consciousness. Of course, that is a guess.
Currently a True North neural net chip simulates about a million neurons with 4096 hardware GPU cores. So 10,000 True North chips may be enough to simulate consciousness. That would be a very large, very expensive machine. However, most neural networks are run on GPUs that were designed for accelerating the graphics of your video games. A number of semiconductors designed specifically for neural networks are being designed, and showing impressive performance gains. Iif Moore's Law continues to hold, it is likely one could simulate 10 Billion neurons on a single chip as early as 2038. Less convincing mimicry may be possible in just 8 years.
However, more powerful computing environments will be easily available in even less time. One can rent arrays of processors, GPU arrays, and soon neural network chip arrays. The needed computing power may be available very soon, and be available on your phone some years later.
Of course, that assumes that all of those neurons are really necessary. In extreme hydrocephalus cases, fluid often fills 95% of the cranium, leaving only 5% for neurons. Yet, these people with almost no brain often have IQs of over 100, in one case over 126! It may be possible to organize a neural net that can simulate consciousness with far fewer than 10 billion neurons, perhaps as few as 100 million along with access to traditional data storage that can swap in and out less used networks. 100 million neurons is not that far off. The SPAUN simulation is an example of what can be done with 2.5 million neurons.
We are not that far the technology to generate a convincing simulation of you, if you are ready to take advantage of the opportunity.
Dr Suzana Herculano-Houzel came up with the 86 billion neuron number based on a count of neuron nuclei.