|
After watching a lecture of Dreyfus on Heidegger and skimming his books I think I begin to see what's going on. Hubert Dreyfus teaches a specific set of philosophers beginning with Heidegger. His brother Stuart is an industrial engineer working on programming very early computers (50s) for Operations Research. They both worked at RAND corporation and involved in early "AI" research projects for military. Those AI projects of course were logic based problem solvers of 50s and 60s. But Hubert sees a problem with this work. What they do is incompatible with the philosophical tradition he belongs: "I began reading NSS's landmark papers with a mixture of excitement and fear. Perhaps Hobbes, Kant, and Husserl were right after all, and the human mind was an analytic engine. But then what about the seemingly plausible arguments of Merleau-Ponty, Heidegger, and Wittgenstein, which I had come to accept? As I read the RAND papers my excitement and fear turned to disappointment and relief." I think this was more like a cognitive dissonance than an actual contradiction. It was about choosing sides between Heidegger and maybe Descartes for him. That's why his objection sounds personal and dogmatic. So what was the big idea of Heidegger? Elephant in the room is the concept of "Dasein" (being-there), which Dreyfus think Heidegger is a genius for being the first philosopher noticing that. Dasein is an entity, with special mode of being. Only *human-beings* can be Dasein. So it's not an "object" like a table, that has properties (like in object oriented programming). It's not an equipment like a hammer (objects having methods). Human-beings, and only human-beings (definitely not LLMs, not dogs) can have this special way of being, or Existence. This idea of course has some roots in Christian Theology, as Heidegger himself. I think this is the reason of the strong opinions. A bit dogmatic. A belief that humans are fundamentally different beings. So an Equipment trying to mimic "Dasein" is categorically wrong (even impossible) in this belief system. The problem doesn't have to be Dasein itself, but you get the idea. It's either-or. If modern AI research is not flawed, their philosophy must be flawed. Since it can't be flawed, AI research must be flawed. Since research is trial and error, failures are the part of the process. But for them, each failure is an "increasing evidence". |