http://www.gazettenet.com/2011/10/28/robert-ackermann.

The exception that proves the rule !

]]>I was wondering though about whether people would say that we “know” a prime if instead of having a simple closed formula like the one in the essay, we had a more complicated one, that we couldn’t evaluate so efficiently. I suspect many people would be satisfied with it. In fact, I suspect many people would be more satisfied with a complicated closed formula for the answer of a problem than with a simpler algorithm to compute it.

It seems to me that the possible complexity-theoretic arguments that might be true against AI, might also be made in areas other than AI where something should be possible to build, but we cannot build it. (Of course, a more likely candidate is “engineering problems”, but if there is something deeper behind those problems, then that something might be related to computational complexity, and efficiency in general).

Also, I think there is a typo in the last paragraph of page 14, when it says “then that would certainly strong excellent empirical evidence”.

]]>In normal circumstances, the distinction is indeed that you can go any direction in space but only one direction in time. Within the framework of our two best theories (relativity and QM), this is not a conjecture but a straightforward fact (in relativity, the difference is that time has the opposite metric signature).

–Scott (on a ship off the coast of Norway)

]]>