They estimate that if Moore’s law continues, we will have the computational capacity to emulate a human brain at the level of its spiking neural network by 2019, or at the level of metabolites and neurotransmitters by 2029.
...for one million dollars. Note that that would not be cost-competitive with humans.
That said, there are individuals that corporations pay more than a million dollars to rent the time of. If we assume that decision is cost-effective (which is a big “if”), getting to own those individuals outright for a million dollars might be a bargain.
Here, we risk crossing over from the realm of wondering “how much computer power it would take” into the bizarre fantasy realm—where emulations actually happen before engineered machine intelligence does.
Agreed. OTOH, to my mind we’d already made that crossover earlier in the discussion, as well… once we have engineered human-level machine intelligences in the mix, all assumptions about how much anything costs are just pretty-sounding numbers, so to talk about emulations costing a million dollars (or any other particular number) already presumes that we don’t have engineered human-level machine intelligences yet.
...for one million dollars. Note that that would not be cost-competitive with humans.
Typical humans, no.
That said, there are individuals that corporations pay more than a million dollars to rent the time of. If we assume that decision is cost-effective (which is a big “if”), getting to own those individuals outright for a million dollars might be a bargain.
Here, we risk crossing over from the realm of wondering “how much computer power it would take” into the bizarre fantasy realm—where emulations actually happen before engineered machine intelligence does.
Agreed. OTOH, to my mind we’d already made that crossover earlier in the discussion, as well… once we have engineered human-level machine intelligences in the mix, all assumptions about how much anything costs are just pretty-sounding numbers, so to talk about emulations costing a million dollars (or any other particular number) already presumes that we don’t have engineered human-level machine intelligences yet.