The Turing Test

If I devise a test to distinguish between the weak and the strong, and that test is the lifting of a 10 pound weight, then the fact that someone lifts the 10 pound weight doesn’t convincing show that they properly belong to the class of the strong. If by contrast the test involves lifting a 500 pound weight…

The Turing Test is a similarly inadequate way to  distinguish between computer and human.

Because computers have a finite number of rules, it follows that when a problem is more complex than the computational means of solution, a computer will fail to solve it in finite time. An example is the Travelling Salesman Problem, when cities represent instructions, and a set of instructions is executable if and only if it is possible to visit every city/instruction exactly once before returned to a home city/halt state for a certain cost, and when the number of cities is greater than the number of instructions in the computer’s program.

There is a class of problems whose solutions depend upon the notion of all possible rules, and these problems are beyond the capacity of any finite set of rules, and hence beyond the capacity of a classical computer. They are not however beyond the capacity of the human mind (nor that of a natural quantum computer). A test that is able decide if a putative intelligence is human-like is therefore that of solving a problem unsolvable by a man-made computer. If and only if a computer can per impossible solve one of these problems (e.g. P versus NP or one of the Millennium Problems), then I will agree that a computer deserves to be called intelligent, for that is the only way that a computer could “deceive” me into believing it was human. I’m not going to believe for a moment that a computer is intelligent in an infinite-human-sense unless it can pass a far sterner test than that wishfully devised by Turing, who apparently found it comforting for some strange reason to believe that humans are intelligent in a finite-machine-sense. He surely cannot by the limits on his own Turing Machines have believed that machines possess an infinite intelligence…