To Build Truly Intelligent Machines, Teach Them Cause and Effect (2018)

In the 1980s, when everybody was trying to do AI with some flavor of predicate calculus, he extended that to probabilistic predicate calculus. That helped. But it didn't lead to common sense reasoning. The field is so stuck that few people are even trying. Working on common sense, defined as predicting what happens next from observations of the current state, is a classic AI problem on which little progress has been made. I used to remark that most of life…

I think to build "truly intelligent" machines, what you need above all else - and which I never hear discussed - is want. Everything about natural intelligence derives from the organism's wants - we want food, we want safety, we want sleep, we want to reproduce, etc etc - even fidgeting comes down to wanting physical comfort. Every single motion and action derives from want. As long as machines are simply executing instructions and have no want of their own…

>> Mathematics has not developed the asymmetric language required to capture our understanding that if X causes Y that does not mean that Y causes X. There we go again. These assertions by Pearl discredit him. Yes, mathematics can capture this perfectly well, it's the implication relation in a nutshell: X → Y means that X causes Y but we don't know whether Y causes X. I guess, strictly speaking, implication doesn't say anything specific about causality, so you could…