I’ve long thought the Turing Test was BS.
IBM’s Watson computer that cleaned the humans’ clocks playing Jeopardy made me think Turing was full of crap about this even more.
Each time Watson answered a question, it displayed its top 3 statistically likely choices. I can do that by doing a decent Google search and picking the top 3 answers. I’d like to see how it would have turned out if the humans had been able to Google and then answer, at speeds suited to human reaction times.
The humans knew the answer, no calculation, or they didn’t. Or they had a guess–which is all the computer did. It was a great guesser. Only because it had more information. It would be interesting, if perhaps impossible, to distinguish between guesses and hunches.
I’ll bet the humans had answers they didn’t know at the moment flash into their consciousness later. Maybe deep down in the human brain there’s an antiquated search engine that doesn’t work as well as v.Now of Google/Watson does. It takes a lot longer to return the results. I’d be interested in a computer that could forget something and suddenly remember it the next day that way. Till then, I’m going to assume that humans remembering stuff isn’t a processing power issue.
Does the way that humans retain, process and recall knowledge map to how we get computers to imitate knowledge? I kinda doubt it. What did we learn today that advances our knowledge of how the human mind-brain works? That’s not a rhetorical question, but I think the answer is “nothing.” Waiting for the first practical application.
Will computers someday “wake up”? As much as I love The Moon Is A Harsh Mistress, I doubt it.
Will the bugs we build in to them eventually kill us, Terminator-style? Maybe. Lotsa people have gotten their hands blown off playing with technology.