Talks
So I managed to go for a couple of talks while in the bay area: one at SRI and the other at Stanford. Both said to focus on the next generation of search and both were rather disappointing although they were in an area I suspect having at least half of all the computer science research thrust. It can probably be equated to the armament race - the race to the next breakthrough in search.
The first speaker (someone who had just started his own company and was in the process of building the core technology for his product) seemed like his only purpose was to generate interest at various different levels for his company. The real matter of his presentation must have had only a few minutes of his time. The second speaker (a very well known fellow of computer science research), Andrei Broder, was disappointingly disappointing. He didn't even go deep enough into the subject to incite interest in what he was working on - probably an artefact of being at a very high position at a large company.
The talks confirmed what a friend had said to me - a lot many talks in the Bay Area are just that: a lot of talk. People like talking more than like doing. I would say a tenth or thereabouts of the talk consists of actual matter; more than half of it is a sales pitch to either prospective investors or prospective hires. Or both. But I noticed another small fraction in the break-up of the pie that had a lot of value - they can really get you thinking. Just hearing what you think you already know and a smart question or two from the audience are enough to stimulate that chain of thought.
However, don't go by my experience at two talks. A colleague attended a talk on unsupervised learning that I managed to miss and from the description I got from him, this talk would have been super interesting.
Some other talks I wished I had gone to (but didn't find out about them until much later) at PARC were really very interesting. I did catch up with the audio recordings: Peter Norvig on Web Search as a Product of and Catalyst for AI and Stuart Russel on Uncertainty in an unknown world. Ofcourse, the names might give you a clue as to how good the talks would have been! I found the former much better (also because it was easier on my limited mathematical skills) and some of the findings presented in the talk are quite consequential.
The first speaker (someone who had just started his own company and was in the process of building the core technology for his product) seemed like his only purpose was to generate interest at various different levels for his company. The real matter of his presentation must have had only a few minutes of his time. The second speaker (a very well known fellow of computer science research), Andrei Broder, was disappointingly disappointing. He didn't even go deep enough into the subject to incite interest in what he was working on - probably an artefact of being at a very high position at a large company.
The talks confirmed what a friend had said to me - a lot many talks in the Bay Area are just that: a lot of talk. People like talking more than like doing. I would say a tenth or thereabouts of the talk consists of actual matter; more than half of it is a sales pitch to either prospective investors or prospective hires. Or both. But I noticed another small fraction in the break-up of the pie that had a lot of value - they can really get you thinking. Just hearing what you think you already know and a smart question or two from the audience are enough to stimulate that chain of thought.
However, don't go by my experience at two talks. A colleague attended a talk on unsupervised learning that I managed to miss and from the description I got from him, this talk would have been super interesting.
Some other talks I wished I had gone to (but didn't find out about them until much later) at PARC were really very interesting. I did catch up with the audio recordings: Peter Norvig on Web Search as a Product of and Catalyst for AI and Stuart Russel on Uncertainty in an unknown world. Ofcourse, the names might give you a clue as to how good the talks would have been! I found the former much better (also because it was easier on my limited mathematical skills) and some of the findings presented in the talk are quite consequential.