5 ms·
> If two different species both want the same thing and only one can have it—in this case, to be the dominant species on the planet and beyond—they are going to
by golergka 9y ago
> If two different species both want the same thing and only one can have it—in this case, to be the dominant species on the planet and beyond—they are going to have conflict.
The term "dominant species" really jumped at me as I was reading this piece. It's so fantastically vague that it begs the question - would us and AI have the same definition for what "dominant" is?
We're not building AI individuals with self-preservation instinct, hunger for resources and sexual drive. We're building super-individual systems like Google, Facebook and trading systems, that can be much more intelligent than we are but also have vastly different built-in "purposes" (just as evolution have "built-in" purpose of eating, having sex and caring for our family into us).
I think that in the end, the thing we're building will end up much closer to Solaris ocean than Terminator.
- pixl97 9y agoWell, the issue is we are not building general AI systems at all yet. If and when we can, the game changes, and we pretty much cannot figure how it will until it happens. There is no other general intelligence at this point that can rival human intelligence, we have a systems sample size of 1. Making predictions based on such a wide range of possible technologies is hopeless at this point. Why would a fixed placement general learning computer system that takes input from millions of possible sources evolve in the same way that a mobile robot with general learning intelligence? Will both be possible? What about swarm intelligences? Too many questions with no possible way to answer them.