>>12690585It can be achieved only when the computers are as strong as the brain's data and information processing.
The size of the computation required to simulate consciousness is staggering. If you believe it's on the cellular level, we aren't so far today, we have computers with order 10^10 bytes, so we can simulate millions of neurons at the neuron level, assuming the data processing is at the neuron level.
But since the data processing is pretty obviously deeply intracellular, and is fundamentally intracellular and mediated by RNA, the computational information in the brain is most well approximated by the total number of nucleotides of RNA in the brain.
This amounts to 10^20 bytes, a billion gigabytes, and is comparable to all the artificial computers put together. The scale for bit storage is then comparable to the atomic scale, it's only about 10 cubic Angstroms per bit. At the moment, our bits are at the lithographic scale, or about 1000 square Angstroms per bit, so the computers are much more primitive.
This is simply comparing raw processing power, but I think this is the most important thing to compare. A computer the size of 10^20 bytes, even if programmed inefficiently, should be able to do some amazing things, and tinkering might produce intelligence without much additional design required.
But of course, one should try to understand the algorithms of the brain as well as possible, perhaps one has to mimic it quite closely, not just get the processing and order of magnitude of memory right, and have any old learning algorithm.
