Yesterday there was a lot of talk here on machine intelligence and I know you had expressed some concerns about how machines are making leaps in processing ability. What do you think we should be doing?
Gates:
There are two different threat models for AI. One is simply the labor substitution problem. That, in a certain way, seems like it should be solvable because what you are really talking about is an embarrassment of riches. But it is happening so quickly. It does raise some very interesting questions given the speed with which it happens.
Then you have the issue of greater-than-human intelligence. That one, I’ll be very interested to spend time with people who think they know how we avoid that. I know Elon [Musk] just gave some money. A guy at Microsoft, Eric Horvitz, gave some money to Stanford. I think there are some serious efforts to look into could you avoid that problem.
Gates goes into a little bit more detail on his views on AI.
Interviewer:
Gates:
Horvitz’s thing.
Musk’s thing.
If Gates were to really get on board, that would be huge, I think. Fingers crossed.