Thursday, August 24, 2023

Alchemy lives

Moral: Machine learning's recent splash brings Herbert L. Dreyfus back to the conversation. Let's look at that. 

---

I first heard of Dreyfus while listening to a discussion about artificial intelligence and databases at a conference that was sponsored by the "very large-database" group. The meeting was in the 1980s and was held at Kyoto, Japan. The reactions were varied, but one could see the positions being taken. He didn't seem to have many friends there defending him. 

Okay, leap forward. Looking further, I agreed with the guy. However, at the time, my focus was on implementation of algorithms for knowledge based systems (and, later, knowledge based engineering modes) that were highly effective in providing solutons that mattered. Needless to say, subsequent work involved a broader scope for computing that suggested its potential. 

Ubiquity? The concept was not unknown. However, the release of IP changed the whole tone. That was in the mid-1990s. Since then, we have had several cycles of boom and bust. The first one? Go look up the tech bust of 1999/2000 to read about one. 

There were others before and after 2000. A couple of the ones before related to artificial intelligence. This post provides a brief summary of Dreyfus's involvement in the discussions. 

Now, the theme of this post. Here is a little blurb from Bard (Google's xNN/LLM)  with respect to Dreyfus's importance.  
  • Business is often seen as a way to make money, and in some cases, this can lead to people trying to get something for nothing. For example, some businesses may engage in deceptive or fraudulent practices in order to make a profit.
Oh, was this prompted? Sure. The idea was to tell it that the ca-pital-sino (coined in this blog in a post on Tuesday, 26 Jan 2010 - Shell games) deals specifically with this issue. 

Remarks: Modified: 08/23/2023

08/23/2023 --