As someone who went through Stanford CS in 1983-1985, when the "expert systems" people, led by Feigenbaum, were in charge, I can say that the expert systems crowd was stuck by then, but in denial about it. Feigenbaum was testifying before Congress: "The US will become an agrarian nation" if Congress didn't fund a big national AI lab headed by him. The expert systems crowd were claiming they were headed for strong AI in the near term. They were in complete denial about the fact that expert systems mostly give you back what you put in. You can code up some kinds of "how-to" information as rules, but it's just a form of declarative programming.
It was frustrating hearing those professors. I'd spent time in industry before going to Stanford for a MSCS, and I could tell that they were stuck and covering for it.
A few years later, Stanford moved computer science from Arts and Sciences to Engineering. This pushed the CS department into more practical directions. The School of Engineering had deans, supervision, and planning. CS under Arts and Sciences was sort of what each professor wanted to teach. There was no undergraduate CS at Stanford before that move; prospective undergrad students were told "get a liberal education".
That didn't end the "AI Winter" at Stanford, though. It took the DARPA Grand Challenge, in 2003-2005, to do that. The head of DARPA, Dr. Tony Tether, used the Grand Challenge to force the big-name AI schools to get some results. DARPA had been funding automatic driving work at Stanford and CMU since the 1960s, with disappointing results. The big-name schools were quietly told that if the private sector and amateurs outdid them in the Grand Challenge, DARPA was turning off their AI funding. Suddenly entire CS departments were devoted to the Grand Challenge. Stanford had to bring in machine learning people from CMU to compete. That's what turned the department around and finally pushed the logicians into the background.
It was frustrating hearing those professors. I'd spent time in industry before going to Stanford for a MSCS, and I could tell that they were stuck and covering for it.
A few years later, Stanford moved computer science from Arts and Sciences to Engineering. This pushed the CS department into more practical directions. The School of Engineering had deans, supervision, and planning. CS under Arts and Sciences was sort of what each professor wanted to teach. There was no undergraduate CS at Stanford before that move; prospective undergrad students were told "get a liberal education".
That didn't end the "AI Winter" at Stanford, though. It took the DARPA Grand Challenge, in 2003-2005, to do that. The head of DARPA, Dr. Tony Tether, used the Grand Challenge to force the big-name AI schools to get some results. DARPA had been funding automatic driving work at Stanford and CMU since the 1960s, with disappointing results. The big-name schools were quietly told that if the private sector and amateurs outdid them in the Grand Challenge, DARPA was turning off their AI funding. Suddenly entire CS departments were devoted to the Grand Challenge. Stanford had to bring in machine learning people from CMU to compete. That's what turned the department around and finally pushed the logicians into the background.