Marvin Minsky, an artificial intelligence pioneer and co-founder of the M.I.T. Artificial Intelligence Lab, passed away on January 24th.
Artificial intelligence went into a period of reduced interest called the "AI Winter" after long being over-hyped. But more recently, AI has resurrected itself in different forms, finding its ways into everyday life through as self-driving cars or virtual assistants like the Amazon Echo, Google voice search, and Microsoft's Cortana. We must remember that Minsky's lab's efforts from the 1950's laid the groundwork for what we have today.
Minsky believed that there was no fundamental difference between machines and humans, and that machines can mimic human behavior. He was not a fan of Intelligence Augmentation (machines helping people do better work). Some of this was discussed in a prior blog of mine.
He described research dollars spent on work by Doug Engelbart's on NLS (what can be described as an early hyperlinked information system and graphical UI) as a glorified word processor. In practice, both approaches continued to succeed. Minsky's lab also created an early neural network system, and its descendants help drive anything from machine translation to image recognition.
Some lesser known contributions to popular culture include;
- Advising Stanley Kubrick on computer graphics for the science fiction movie, 2001: A Space Odyssey (the actual graphics systems available in the lab was never used in the movie since the devices available in the late 1960's were too crude for what was expected to be available in 2001)
- The quote, "once the computers got control, we might never get it back. We would survive at their sufferance. If we're lucky, they might decide to keep us as pets" Life Magazine (20 November 1970).
I suppose the fate of being a pet is better than being terminated by Skynet in the Terminator movies, but one can argue whether or not it follows the Three Laws of Robotics. You have to hand it to Minsky for being optimistic about AI capabilities!
As for me, an analyst covering networking technologies, I continue to hope that AI can assist in network diagnosis, automatic traffic management, or configuration. True AI may still be years out, but we're making good progress. On a different note, a computer has beaten some top Go players just this week, thanks to work by Google researchers. Good computer-playing Go programs has been a tough nut to crack, and this is a major development.