A few years ago the NYT editorial board argued that inflation was going to be transitory because "it's not the 1970s." It was a smug, dismissive, and ultimately

A few years ago the NYT editorial board argued that inflation was going to be transitory because "it's not the 1970s." It was a smug, dismissive, and ultimately incorrect argument. Appealing to the past doesn't mean much because we don't live there. This argument about automation is no different. We don't know what is going to happen because of AI. We don't even know where AI is going to go as a product/feature/whatever. It could totally make obsolete an entire class of lower-earning white collar workers whose main contribution is writing things. On the other hand, it could fizzle out and we could be comparing Bing Chat to Clippy a few years from now and laughing at all the investors who wasted their money on the next big thing after crypto. Pointing to times in the past when people feared automation, and showing how those fears did not pan out, is not a good argument. Everything in history is contingent, everything could have gone a different way if a few people made different decisions at a few key times, and fear of automation may have motivated some of those people to make the decisions that prevented automation from destroying all the jobs. To put it another way: If you play Russian Roulette with a six-shot revolver and you survive five rounds, should you be more or less likely to believe that you will survive the sixth round? People who look to the past and say "bad thing that was predicted did not happen then so it won't happen now" are guilty of thinking that if they survived five rounds then they will definitely survive the next one.