This paper is entertaining but it depends on the assumption that computing power grows exponentially forever and will eventually be able to solve all P problems at negligible cost. That assumption is clearly not justified.
And the belief that any market perfectly incorporates data, not a single point off in any stock, is basically a strawman anyway.
