| |
|
Saw this Wired article though I haven't found any transcripts of the speech it's talking about.
"AI has been brain-dead since the 1970s," said AI guru Marvin Minsky in a recent speech at Boston University. Minsky co-founded the MIT Artificial Intelligence Laboratory in 1959 with John McCarthy.
Such notions as "water is wet" and "fire is hot" have proved elusive quarry for AI researchers. Minsky accused researchers of giving up on the immense challenge of building a fully autonomous, thinking machine.
Much as Minsky is seminal as a whale's scrotum, I don't agree with him here. Or, rather, I agree, but I don't think that's the point any more.
I've said numerous times here and elsewhere that "building a fully autonomous, thinking machine" is not what we should be thinking about in AI. It's vastly expensive (and not likely to be funded), by the most optimistic projections extremely difficult and time consuming, and doesn't even have a good philosophical grounding. What's more... what's the point? We already have billions of autonomous thinking machines, and more every day. Understanding how they work, sure, very useful. Using AI as a tool in this, great. As a goal?
"We're building systems that detect very subtle patterns in huge amounts of data," said Tom Mitchell, director of the Center for Automated Learning and Discovery at Carnegie Mellon University, and president of the American Association for Artificial Intelligence. "The question is, what is the best research strategy to get (us) from where we are today to an integrated, autonomous intelligent agent?"
How autonomous and intelligent would we ever want an integrated agent to be? More so than today, less so than a human, in my opinion. What we really want is an idiot-savant agent, one which is easily controlled in most respects but does some things - which we don't want to do - really, really well. This is the pattern for most devices. We don't want to walk and carry things, so we build cars, which do nothing except move things around, but do that a lot better than we do.
"The worst fad has been these stupid little robots," said Minsky. "Graduate students are wasting 3 years of their lives soldering and repairing robots, instead of making them smart. It's really shocking."
(Actually, I have some sympathy with him here, but I think that's my distrust of hardware coming out.)
There's a very good point relating to public perception of AI at the end:
AI researchers also may be the victims of their own success. The public takes for granted that the Internet is searchable and that people can make airline reservations over the phone -- these are examples of AI at work.
"It's a crazy position to be in," said Martha Pollack, a professor at the Artificial Intelligence Laboratory at the University of Michigan and executive editor of the Journal of Artificial Intelligence Research.
"As soon as we solve a problem," said Pollack, "instead of looking at the solution as AI, we come to view it as just another computer system."
Spellcheckers - not AI, just computing. AI is only AI if it doesn't actually work. Or comes back from the future to kill Edward Furlong. |
|
|
