The money spent on CPU/GPU manufacturing improvement is a sunk cost and more than paid for itself along the way in economics benefits; the cost of the cluster t

The money spent on CPU/GPU manufacturing improvement is a sunk cost and more than paid for itself along the way in economics benefits; the cost of the cluster that beat him is already around what he was paid for this one set of matches. In contrast, making a Lee Sedol will never be cheaper, and he lacks all of the other myriad advantages of software, like being instantly clonable and continuously improvable. You will never be able to mash up Shusaku and Sedol to see what they do, train them against each other for subjective millennia, examine the internals of their position evaluation, increase or decrease the sizes of their brains, compress them into much smaller neural nets, combine with other algorithms like MCTS, etc. Energy consumption is among the least important aspects of AI at the moment. We are not in a situation where we know how to make a feasible human-level AI but because electricity costs 10 cents/kwh no one can run, but they could if electricity cost 5 cents/kwh...