When one considers that modern computers are, at their core, still only what they have always been from their inception, solvers of math problems/rule following machines, one is led down some interesting philosophical alleys. The implications of this fact are far reaching, especially considered in the light of the seemingly endless obsession with machines that “learn” and artificial intelligence. If you believe machines can learn and that computers/machines can be intelligent then you have to believe that being good at math, I mean being really, really good at it, automatically designates the thing in question as intelligent. Higher level math skill is like an intelligence granting genie. Once you have it, voila, you are intelligent. Now consider a machine that can solve any math problem ever conceived. By this way of thinking it must be one of the most, if not the most, intelligent ‘thing’ in the universe.
If you have ever read any of my previous works on this topic (The vast, vast majority have not. As I always say I am not a very popular writer.) then you can probably guess that I disagree strongly with this position and conclusion. In my view a machine that could solve any math problem ever conceived would still not be intelligent, it would be the most powerful calculator in the universe. Now a calculator that powerful would be an awesome tool and no doubt extremely useful and valuable, it would not however be an artificial intelligence or a learning machine.
The problem for proponents of AI is a large one. They are forced to except that ability at math/computation and rule following equals intelligence. Once a thing possesses a certain level of ability (what that level might be is of course impossible to say) it somehow becomes intelligent. The only way around this is to argue that a computer is more than a computational engine, more than an executor of algorithms (mathematical and statistical rules), more then what it is programmed to be. This seems patently ridiculous and absurd as there is no known mechanism for how such an “emergence” of intelligence out of what was previously rote programming/rule following could actually happen, unless one believes that magic or divine intervention are options.
Those who says modern computers are no longer rule following machines must prove that is not the case. Just saying it and then turning around and programming a computer with algorithms (aka rules) doesn’t cut it. And it does not matter the programming language, or the statistics or math that compose said algorithms which is contained within/described by the programming language. I would argue it does not even matter that it is a human being that programs it. It could be an alien or even another computer, but it is still a program. It also does not matter if certain (truly random/quantum) elements are contained within the program. It is still a program consisting of algorithms, which are composed of mathematical and statistical structures comprising a series of rules. The ability to do math and/or statistics is not a necessary or sufficient condition for intelligence, though of course it sure helps.