The Neurological Scaling of Human Expertise

摘要

Although chip clock rates seem to have plateaued, the inexorable rise of computing power in accordance with Moore's law continues. We can easily measure the increase in performance using a portfolio of metrics or a Pareto surface across them, including clock rate, memory latency, bus speeds and so on. In this paper, we address two questions. The first of these is what it would mean to scale a human brain, in the way that the primate brain has been getting steadily bigger and more powerful in the lead up to homo sapiens. The second is whether, if we could scale the human brain at the same rate as computer power, human algorithms and computational processes would continue to dominate in the domains where humans still reign supreme. To consider these questions we will phrase much of our practical considerations in terms of board games, particularly the games of Go and Chess.