Ambition is Path-Dependent
This is my first post. Actually, it’s copied from something I recently wrote on Facebook, but there is more to say than really fits in a Facebook status message. I hope to expand on this soon with longer (medium-size?) posts.
Marvin Minsky, one of the founders of AI, reflects on the history of the field in this video http://www.technologyreview.com/video/543031/marvin-minsky-reflects-on-a-life-in-ai/
He expresses disappointment with the current state and rate of progress, saying “big companies and bad ideas don’t mix very well.”
I don’t share Minsky’s assessment. Minsky was one of the most important and influential people in AI. Arguably, he was too influential at times. But he was undeniably brilliant, so I think it’s worth trying to understand his perspective.
Like any technological discipline, AI has followed a path-dependent trajectory over the last several decades. Improvement has been discontinuous and iterative, with innovations built upon earlier successful ideas. But our ambitions also follow a path-dependent course: what we want to do changes based on what we discover we can do. Perhaps Minsky perceives these changing ambitions as a loss of ambition.