(These are excerpts from my book "Intelligence is not Artificial")
Recap
Here is what i told you so far. There are assumptions underlying the belief that super-intelligent machines are coming soon. The first one is that A.I. is making staggering progress and the second one is that progress is accelerating like never before. I showed you that both statements are wild exaggerations. There are still colossal gaps in the program of A.I. and very few creative ideas on how to fill those gaps. Brute-force A.I. is unlikely to succeed outside problems of pattern recognition and brute-force A.I. has relied too much on faster and faster processors. Now that Moore's Law is coming to an end, we will need more (ehm ehm) intelligent ways to do A.I. than brute force. I am not denying that there is progress in the way machines can work for us, but I will demystify why they can work better than in the past (it is more about the environment that we structure for the machines than about their intelligence). This explains why most of the machines around us are pretty stupid and why i don't see any robots walking around the streets of Silicon Valley. Precisely because of the limitations of today's A.I., you don't fear that machines will steal your job unless your job is so stupid that even a stupid machine can do it.
As for the "accelerating progress" of our age, in most cases it is neither "unique" nor "progress". One century ago the world was completely changed by a series of inventions that happened one after the other within a few years: the telephone, the radio, the car, the airplane, the record, Quantum Mechanics, Relativity, etc. Are you sure that today's progress is more dramatic than that one? When answering this question, let's keep in mind that change is not always progress. Change can go in both directions: forward or backwards. Change is not necessarily in the direction of progress. There have been a lot of changes in Syria since 2011, but only ISIS would call it "progress".
More criticism of today's A.I. is coming in the next pages, some of it more philosophical than practical, and in particular we will discuss the concept of super-human intelligence. But first let's go back to the fundamental question: why A.I. at all? And this will also link to what i wrote at the very beginning: I am not afraid of A.I.; i am afraid that it will not come soon enough.
Back to the Table of Contents
Purchase "Intelligence is not Artificial"
|