Only if the desires of AI researchers are what "fuel" such developments. More likely, and in fact what historically I saw happen, was that companies who had acquired lots of data but had very narrow business models tried to find other ways to monetize the data by processing it further, ending up with what we see in the form of presently-existing surveillance capitalism. This highly processed human behavior pattern matching is to general artificial intelligence what Cheezwhizz and Velveeta are scalable artisanal cheese-making.

The history is very useful in explaining what's happened, as is usual with history. Some remarks on the structure of other "AI" "successes" such as programs that win chess and go games without knowing that chess and go are games, or that they are playing, would be helpful. This helps to show what Dreyfus was talking about.

But it would also strengthen the draft to discuss what the CS people who didn't believe in AI did believe in, and what they did with it. What Seymour Pappert was doing with logo, what Sherry Turkle's thinking and writing achieved, how that affected what Richard Stallman and I were thinking, and what we did with it—the other story that runs alongside the AI idea is about the humanities and the technology. It's finest day has not yet come, either.