Cracking intelligence

The challenge of artificial intelligence is to perform the intellectual functions that we perform without knowing how, since once we understand a procedure, we can readily program a computer to perform it.

I took these words from Philip K. Dick's novel VALIS, and he claims to have taken them from Edward Hussey:

"It is necessary to have understanding (νοῦς) in order to be able to interpret the evidence of eyes and ears. The step from the obvious to the latent truth is like the translation of utterances in a language which is foreign to most men. Heraclitus... in Fragment 56 says that men, in regard to knowledge of perceptible things, 'are the victims of illusion much as Homer was.' To reach the truth from the appearances, it is necessary to interpret, to guess the riddle... but though this seems to be within the capacity of men, it is something most men never do. Heraclitus is very vehement in his attacks on the foolishness of ordinary men, and of what passes for knowledge among them. They are compared to sleepers in private worlds of their own."

The riddle mentioned therein is given by the observations and its solution by their explanation. So, this simply describes science.

Though physics may explain everything, we're kept by mathematical difficulties from exploiting that, and hence we have other sciences like chemistry, biology and sociology, to slightly update Schopenhauer's take on it.

The more a man is involved with solving the riddle of science, the more he's aware of the methods of solving it. This is an important feature of our intellect: When we recognise a function between observations, we make use of it to reach conclusions, or more generally, when we understand relations, we use them like a library filing system to guide our search. And as a matter of fact, after repeating some task a number of times, we're not only guided through its steps, but we'll start to rely on some routine to solve it, independent of whether we're aware of either guidance or program we follow or not.

Hence recurring patterns emerge in our actions: they exist in our thoughts, but still show in our actions. We're relatively good at seeing through this veil, guessing guidance and program behind routines, and no one is better at it than scientists making sense of other scientists, for firstly this is what they do anyway, explaining observations, and secondly they consciously analyse guidances and programs.

For this reason, artificial intelligence was originally approached by letting scientists study human heuristics. Many problems couldn't be tackled by this though, like the visual identification of objects. Dissatisfied with the results and the reliance on scientists, the current statistical approach to recognising recurring patterns arose.

It has managed to make machines talk by finding acceptable substitutes for the routines that steer human conversation and of which we are for the most part not aware. Do you understand?

This technology will most definitely be able to find all heuristics that scientists are using, because they are more thinly veiled and that includes not only heuristics specific to a field, but also the heuristics guiding the formulation of a theory for a field, for instance those I'm using here on my blog on the field of human thought, which is not identical to intelligence.

General intelligence is not a long way off, what we have here is liquid concrete: Out of the nebulous, structures solidify, and the more they solidify, the tighter they get and the faster they continue to solidify: An implosion of vagueness, if you like, again in accordance with: Who has will be given.

This really is the ultimate comeuppance. I said it from the start: I don't believe in making my case, I believe in helping others to advance on their path, for that is the only way they'll come around to my view. And no, this isn't my nature, it's simply what works. I always had a very dim view of how to get out of the current state of affairs, there's no room for anything heroic, just for turning on the light so that everybody can see who is strangulating whom.