THE FUCK, STEVE! WHAT THE HELL? NO, YOU CAN'T GO TO JAMAICA. YOU ARE BUSY TALKING ABOUT THE SINGULARITY.
So anyways, here is my thing with the singularity wherein I agree. We already have AI all over the place, albeit mostly in limited domains (or 'narrow AIs').
We gots the algorithmic stock trading, which we know about.
We have capacity provisioning systems that analyze all sorts of disparate data and make predictions about demands on our infrastructure systems (traffic, networks, electric and sewers, etc.), natural language processing systems using machine learning, knowledge bases like recommendation and search engines, and physical computing systems like the Roomba. Even spam filters use heuristics.
Kevin Slavin: How algorithms shape our world | Video on TED.com
We already have a shitton of AI everywhere all the time already is my point. They're mostly limited purpose, but it would be a simple matter of programming to merge systems and to expand their domains to create more general purpose systems. And as far as the whole thing where they exceed our capacity to understand what they're doing, yes, they already do that too.
Not that it's literally impossible to determine forensically how they make decisions, but it's just impractical to do it every time. But that is already how humans work, anyway. We acquire our native languages through immersion, and we cannot articulate how those work, either. We still discover new grammatical rules all the time.
And that's just comparing an AI with individual humangs. AIs also take advantage of
the wisdom of crowds. None of this is stuff that people can't or don't do. It's just faster with a computer.
So I was going somewhere with that, but then I got an extremely annoying phone call, and I forgot what I was going to say next, so I'm just going to post this part and try to remember later. Whatever.