Monday, January 25, 2016

Marvin Minsky, 1927-2016

Prof. Marvin Lee Minsky, or just Marvin to those who knew him, died last night.  We have lost one of the great humans of our time, perhaps of all time.  A humble, brilliant, passionate man, with a blazing intellect and an amazing zen, a childlike curiosity, Marvin pioneered much of the early work in AI, together with John McCarthy, Herb Simon, and Allen Newell and others.  He opened our eyes to much that was new, and he and his academic progeny have shaped a lot of what we know about AI today.

With a grad school recommendation letter of all of a single line, Marvin changed my life.  And his work, his teachings, his ways of exploring the unknown, his ability to span and to combine several widely varied disciplines, have been a great lesson to me, a source of great inspiration over more than 25 years, from my graduate studies, to our recent work on AI, both at Infosys and with OpenAI.  Indeed when I started my AI lecture recently for Infoscions, not happy with any of the recent work I saw, I went back to a paper Marvin wrote before I was born ("Steps Towards Artificial Intelligence").

As sad as I am, and countless others are at his passing, perhaps even sadder is his recent statement, on the current state of the work in AI, at a time when we hear about AI and its impact on our world and our lives all around us.  Last week I was in Davos at the World Economic Forum's mtg, and AI, and its feared impact on people's lives, and jobs, was the talk of the town.  I hosted a panel with some key experts as well, to try and add something hopefully thoughtful, to all these voices, but perhaps just added more to the noise, and all along I kept thinking of how Marvin would have reacted to all that sound, all those alarms.  Perhaps he'd have chuckled, before unleashing a typical Marvin zinger that would put things in perspective, and yet enlighten.  Despite the widespread interest in, and hype around, AI, we are nowhere close to implementing many of Marvin's ideas, including his work in the society of mind, which he published ~25 years ago.

So in looking back on his life, and reflecting on his passing, perhaps the best we can all resolve to do is to live his dream, his aspiration, of building systems that get ever closer to Artificial Intelligence, but to do so in a way that he would have been proud of; his purposefulness and integrity, his gang of experimenters, his childlike curiosity, his "model railroad club", his instinct to look at things from many different perspectives.  That a purposeful, unencumbered, pursuit of artificial intelligence may in the process get us that much closer to our natural spirituality.

Upon hearing the news of Marvin's passing, Alan Kay, his friend for the last 50 odd years, said to me "... there was no one ever like him."  So true.

R.I.P Marvin.  There was no one ever like you.  We will miss you...