Not HAL: Computational Intelligence at Google
October 26, 2008
“Thinking Ahead with Google” by Elise Ackerman and Scott Harris is a very good article about a subject near and dear to some Googlers’ hearts–computational intelligence. You must read the full text here. The old term “artificial intelligence” or AI is not too popular. AI is science fiction. Computational intelligence is pragmatic. The story opens with a reference to a 2002 comment by Sergey Brin about the future of search. The analogy was to the HAL computer in “2001: A Space Odyssey”. HAL, as you may recall, went off his rocker. Accordingly, Google is the “borg,” shorthand for cyborg. The subject of “smart software” is not one that turns up in daily newspapers. I commend the San Jose Mercury News for tackling the subject.
Ms Ackerman and Mr. Harris report that Google will support the “Singularity University” announced at a conference called Singularity Summit. The idea is that “smart computing” is important and needs a focal point. The most important comment in the article for me was this:
The meeting was reported by technology writer Nicholas Carr in his blog Rough Type, after one of the participants blogged about it. But don’t look for the item on Google; organizers requested the information be taken down.
My recollection is that Mr. Brin delivered a talk at a Google developer conference in 2007. That talk did not become available on Google’s YouTube.com. Apparently, the support for a better HAL does not extend to making in depth information available. In my opinion, Google is the computational intelligence singularity. Google’s patent documents are chock full of references to smart software; for example, US20070198481 has little smart fellows named janitors running around autonomously. The janitors clean up data and resolve ambiguities in certain procedures. Check out the San Jose Mercury News story and take a peek at how janitors get smart. Like I said, the computational singularity is Google. I’m fuzzy with regards to “Singularity University”. It might be another Google recruiting method. If you are somewhat paranoid, don’t read Kevin Kelly’s “Evidence of a Global SuperOrganism” here. The creature is wearing one of those flashing Google lapel pins.
Stephen Arnold, October 26, 2008
Comments
3 Responses to “Not HAL: Computational Intelligence at Google”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
wow, the problem of defining AI and CI is not an easy one. CI has been around since the early 80’s and AI since the early 50’s. What gives?
They are 2 different things, but It’s not so clean-cut. One is top-bottom and one is bottom-up. CI is generally about problems that can’t be solved as computational algorithms. One way of differentiating between them is that AI uses symbolic approaches and CI subsymbolic approaches.
Then again…A.I is sometimes considered a part of CI, which deals with higher level cognitive functions.
AI has its “grand challenges” and so does CI (The rat).
I don’t think that AI is sci fi, any more than CI is really. When it comes down to it, the work in both helps both, so it’s all good.
This is always a good resource for that debate: http://cogprints.org/5358/1/06-CIdef.pdf
CJ
I agree that the conventional wisdom makes a distinction between top down and bottom up. My research suggests that Google combines up and down and tosses in side inputs as well.
Stephen Arnold, October 26, 2008
I think the whole research community is doing that 🙂
Makes sense to really. And more and more, research areas like psychology and even sociology are turning up in there too.
I think I just wanted to say that AI wasn’t Sci fi – There is GOFAI and there is quite cutting egde AI, it just focuses on different things to CI, although both overlap quite a bit.