Robots queuing up to fail

I feel safe in predicting a continued barrage of stories about extraordinary advances in artificial intelligence (AI) and machine learning, with plenty of claims about their revolutionary implications. I should note immediately that I’m not advocating for a blanket dismissal of all of them, because you never know: some of these things might indeed prove to be great progress. But what I will make the case for is pausing to think about any such claims in their proper context.