Comments on: I'll Take "Corporate Stiffs on Cheesy Sets" for $200

By: Kevembuangga

Kevembuangga — Mon, 21 Feb 2011 12:50:44 +0000

@Malcolm MacIver – Sorry I screwed the link, a PDF can be found at Carson Chow’s blog “A description of the strategy and algorithms…”.

@Angelisa Josalisa – It is indeed very unfortunate that the general public is seeded with uncalled-for fear and paranoia by the Singularitarians and their ilk (Wired, Bill Joy, etc…).
To find more reasonable views Google for “nick szabo singularity” (cannot put two links in the same comment on pain of “spam” deletion )

By: Angelisa Josalisa

Angelisa Josalisa — Sun, 20 Feb 2011 23:32:46 +0000

My Mom’s response at the end of the program was this. Imagine a 70 year old black woman as you read the lines. “That dirty Watson just pretended not to get final Jeopardy, so we don’t feel that bad when his ass takes over everything.” My Mom saw the entire program as a introduction to the Terminator reality she has been preparing for over the last 20 years. I felt at the end it was a much to do about nothing. But I know for people who may not be use to seeing AI, in movies, comics, this blog, etc. Watson was their introduction. The problem is, I don’t know how they received him. Other than my Mother’s Sarah Connor like cry of war over the machines, I really don’t know how people who are not really into this sort of thing felt about the introduction. Now my parents are afraid Watson will become their future Doctors. I wonder how many other people feel this sense of paranoia about the event?

By: Malcolm MacIver

Malcolm MacIver — Sat, 19 Feb 2011 15:18:10 +0000

@Kevembuangga – thanks (was your second link the same as the first?) Langford mentions that the technical papers will come out in the summer. But, as far as I can tell, other than a general description of the component technologies we won’t know the details until then. Langford also makes this comment: “For a random person, this might seem evidence of serious machine intelligence, while for people working on the system itself, it probably seems like a reasonably good assemblage of existing technologies with several twists to make the entire system work.” It’s almost certainly the latter. If they had a new algorithm we’d probably have heard it mentioned. But, and I say this from years of of struggling with baroquely complex animal nervous systems and biomechanics, I expect a machine that eventually reaches the designation of “intelligent” to be an enormous inelegant collection of hacks.

@Jillinthebox – absolutely – IBM (or IMB!) deserves to promote this and some advertising for their brand during the show would have been appropriate. How much and the manner it was done in counts for a lot. It’s like a bunch of ad execs who usually sell mainframes to big institutions were trying hard to reach a different demographic. So it seemed stiff in manner. And there was just much too much of it. In beating humans at one of their favorite games (and by clear implication, besting them at one, albeit limited, type of intelligence), a little more restraint on tooting the horn of the company that did it would have been better.

By: Jillinthebox

Jillinthebox — Fri, 18 Feb 2011 18:11:45 +0000

I thought the whole thing was amazing.
I agree with Malcolm, the IMB commercialization of it all seemed odd, but perhaps a necessary evil. This type of research is not cheap and a little self indulgent promotion never hurt any scientific endeavor. Besides the Jeopardy challenge was just a gimmick to begin with, what did you expect?

By: Kevembuangga

Kevembuangga — Fri, 18 Feb 2011 06:59:54 +0000

Aaawwwww… silly “anti-spam”!

It would be great to know more of the technical details

There are some here

By: Kevembuangga

Kevembuangga — Fri, 18 Feb 2011 06:57:04 +0000

By: Kevembuangga

Kevembuangga — Fri, 18 Feb 2011 06:55:15 +0000

It would be great to know more of the technical details

There are some here, via http://sciencehouse.wordpress.com

By: Malcolm MacIver

Malcolm MacIver — Fri, 18 Feb 2011 00:14:57 +0000

@FoxtrotCharlie: Yes, absolutely, the breakthrough here is in the segmentation of the question. Once that’s clear — and it can be arbitrarily difficult if word play and ambiguity are present — then you’re right, it’s a reasonably doable search process.

I’d characterize what IBM did as more than a baby step – as I wrote, I believe it will go down in history as the date of a very important milestone in AI. Jeopardy is hard, and I wouldn’t have imagined – based on extant natural language processing approaches that I’m aware of – of something that would handily beat a world champion. While with many of the easy questions it looked like Watson only beat the humans because they couldn’t hit the buzzer fast enough, there were quite tough ones that it did very well with as well.

My issue was with regard to the PR aspects only. The scientists and engineers behind the work are first rate, and what they’ve accomplished is really impressive. It would be great to know more of the technical details, but I imagine they won’t be revealing too much until things have been patented, if then.

@TerryS – thanks for your feedback. Offending readers was not what I had in mind, but instead some reaction on whether the unrelenting IBM PR and commercialism through the episodes was as dreadful as I felt it was. In your case, it seems I failed.

By: felix

felix — Fri, 18 Feb 2011 00:01:09 +0000

Watson seems to be very similar to Google search on a static database. So what’s really amazing IMO is that anyone with an internet connection has instant access to this AI.

By: TerryS.

TerryS. — Thu, 17 Feb 2011 21:24:59 +0000

I quote: “…came in the form of a small cheesecloth bag of genetically modified goat genitalia inserted into the head node of the aforementioned Power Server cluster.”

Having a bad day Malcolm? Was that smarmy statement really necessary? Actually, the tone of the whole blog entry was quite offensive.

Just my two-cents worth.