I've been reading The Media Lab: Inventing the Future at M.I.T., by Stewart Brand. It includes conversations with Marvin Minsky, one of the founders of Artificial Intelligence. And Minsky has staggering ideas. "Imagine," he commands, "what it would be like if TV actually were good. It would be the end of everything we know." I want to show you his ideas about consciousness and the evolution of AI, in this discussion with Brand.
"What happened to artificial intelligence?" I asked. "How come thirty years later it's still trying to succeed?" AI was set up as a receding goal, Minsky replied, "a way of asking, 'What are the twenty most important ways the mind works?' 'Intelligence' is a collection of a lot of mysteries. 'Artificial' just means we can make other things than people do them, so we can explore the horizon of unsolved problems in psychology and computer science. Another advantage back then was that psychologists were so repelled by the idea of artificial intelligence that they didn't consider it competition. They would have if we'd called it Cognitive Something.
"Has any of this cleared up what consciousness is or is for?" I asked. "People have such a small number of memory registers," Minsky said, "that we can't think of much. Everything has to be on automatic. Consciousness is not a window. It's more like a debugging trace you use for reprogramming around problems. Humans are really amazing, considering. Just think of what we must be the next step toward. Imagine having a four-megabit consciousness chip in your head."
Me: "Do you have a standard timeline for when machine intelligence catches up with human intelligence and goes rolling on past?"
Minsky: "Yeah. Between 100 and 500 years. Intelligent evolution is unprecedented. Nobody's ever seen one. So in few hundred years it could do trillions of years of ordinary slow evolution." Me: "And make enormous mistakes." Minsky: "That's the trouble. There's no time to iron out the bugs. It might fill up the universe with styrofoam or something because it has some wrong ideas about how the cosmos needs a shock absorber."