The Noisy Channel

 

A New Kind of Marketing (NKM)

March 9th, 2009 · 13 Comments · General

The blogosphere is a buzz with hype about Wolfram Alpha. Stephen Wolfram writes:

It’s going to be a website: www.wolframalpha.com. With one simple input field that gives access to a huge system, with trillions of pieces of curated data and millions of lines of algorithms.

We’re all working very hard right now to get Wolfram|Alpha ready to go live.

I think it’s going to be pretty exciting. A new paradigm for using computers and the web.

That almost gets us to what people thought computers would be able to do 50 years ago!

And Nova Spivack shares his own excitement:

Stephen was kind enough to spend two hours with me last week to demo his new online service — Wolfram Alpha (scheduled to open in May)….

In a nutshell, Wolfram and his team have built what he calls a “computational knowledge engine” for the Web. OK, so what does that really mean? Basically it means that you can ask it factual questions and it computes answers for you….

Think about that for a minute. It computes the answers. Wolfram Alpha doesn’t simply contain huge amounts of manually entered pairs of questions and answers, nor does it search for answers in a database of facts. Instead, it understands and then computes answers to certain kinds of questions.

I haven’t seen this much excitement about a search-related product since the pre-launches of Cuil and Powerset, and we know how those played out. In fairness to Wolfram, however, he did bring us Mathematica, which is more than a legitimate claim to fame.

However, I’m not so persuaded by his more recent accomplishment of publishing A New Kind of Science, a best-seller and 1200-page coffee table book.  Here’s what Wikipedia tells us about its critical reception:

NKS received extensive media publicity for a scientific book, generating scores of articles in such publications as The New York Times, Newsweek, Wired, and The Economist. It was a best-seller and won numerous awards. NKS was reviewed in a large range of scientific journals. Several themes emerged. On the positive, many reviewers enjoyed the quality of the book’s production, and the clear way Wolfram presented many ideas. Many reviewers, even those who engaged in other criticisms, found aspects of the book to be interesting and thought-provoking. On the negative, many reviewers criticized Wolfram for his lack of modesty, poor editing, lack of mathematical rigor, and the lack of immediate utility of his ideas. Concerning the ultimate importance of the book, a common attitude was that of either skepticism or “wait and see”.

If Wolfram has built a breakthrough tool to support  information seeking, then he should let it prove itself by unveiling it and letting other people test it. We aren’t talking about some kind of esoteric science where only a few intellectuals can hope to understand it. Rather, his product purports to be some kind of search / answer / knowledge engine. It’s 2009, and we’re all used to the general vision. What we’re holding our breath for is execution.

I’m open to the possibility that Wolfram has built something that will change the world. But I’m extemely skeptical, and this hype campaign hardly instills confidence. Apparently he told Nova that the product will be launched in May. Two months: not so long to wait to see how well reality matches the hype.

13 responses so far ↓

  • 1 Jon // Mar 9, 2009 at 9:30 am

    Even if he completely nails it with this QA system — approaching perfection in answering general knowledge factual questions — would that really be the next big thing in “search”? What is the real utility of a system like this?

    IMO, factoid QA is a good place to demo NLP or knowledge representation technologies, but there are very few real use cases for a system like this. Factoid answers need context to be really useful beyond just filling out crossword puzzles, taking tests or answering trivia questions.

    My prediction — when it launches, there’ll be an initial peak of interest when people see that its possible to perform general factoid QA, but that’ll quickly fade when everyone realizes that a factoid QA system doesn’t solve any real problem.

  • 2 Daniel Tunkelang // Mar 9, 2009 at 9:58 am

    There’s a peak of interest already. What I want is a peek at the system. And I agree, he’ll have to show that his system can solve real problems that existing ones can’t–at least if he wants my vote for greatness.

  • 3 Neal Richter // Mar 9, 2009 at 12:00 pm

    Even if this is just QA, we’ve seen that fail before .. even when in my opinion it’s super useful for CRM applications.

    I hope that the True Knowledge (http://www.trueknowledge.com/) people will do it right, they looks like they are working hard, raising a bit of capital, and shipping code/data frequently. These attributes and not hype are better predictors of success.

    For all their hard work Harabagiu, Moldovan et al. were not able to execute a large commercial launch of the QA platform that has been winning the TREC QA track for approx 10 years. See Language Computer Corp and Lymba Corp. Either they tried to get big VC money and failed or they are smart enough to make a modestly successful business as an OEMer of QA software.

    Do we really think that Wolfram is going to be able to beat people who have been working on QA for a decade?

  • 4 Christopher // Mar 9, 2009 at 1:04 pm

    I hope he manages to pull off an unbounded QA system but I am highly skeptical.

    As Neal says “Do we really think that Wolfram is going to be able to beat people who have been working on QA for a decade?”

  • 5 Christopher // Mar 9, 2009 at 1:45 pm

    Interesting side note to this…

    Wolfram’s Mathematica team consults on a TV show called Numbers, essentially a drama where a match genius professor helps his FBI brother solve crimes using math.

    Last week’s episode featured a computer which could pass the Turing Test, essentially coming up with the best answer to any unbounded question…

    Now that’s a new kind of marketing. Coincidence? I think not.

  • 6 Daniel Tunkelang // Mar 9, 2009 at 1:49 pm

    That’s awesome! Someday I hope I have access to that kind of subliminal product placement. Readers in the media industry, please take note.

  • 7 erica // Mar 9, 2009 at 10:49 pm

    with regards to comments that this sort of thing wouldn’t really be useful: really? I can think of plenty of times when I’ve wanted to know the answer to some fairly specific factoid-type question, and either had to figure out how exactly to phrase it, or wade through Wikipedia articles or other web pages to find the answer. For example:
    ‘what is a word that means “too wordy”?’
    ‘what is the least expensive way to get from Boston to New York?’
    ‘which grad schools in North America have programs in underwater basket weaving?’
    ‘how do I use Emacs to make a sandwich?’
    ‘when is CKY parsing better than Earley parsing?’
    ‘is it possible to perform a Fourier transform using some duct tape and my left pinky toe?’
    and plenty of other sorts of random but specific (and less absurd) questions that are probably better examples that I just can’t think of right now — does Wolfram claim that his system will be able to answer questions like these? I think it would be super cool and quite useful if they got this to work. I am pretty skeptical about what can be done in two months’ time, but nevertheless I look forward to seeing what they come up with.

  • 8 Daniel Tunkelang // Mar 9, 2009 at 11:01 pm

    Erica, I’m with you on this one; I’d be impressed with a robust question answering system. The Google / Wikipedia / rest-of-web combo gets me pretty far, but certainly there are a lot of questions it can’t answer because there aren’t sentences on the web that answer them. But I’m not holding my breath that Wolfram Alpha (which sounds like an Ayn Rand protagonist, no?) will do any better, other than perhaps on a carefully tuned subset.

  • 9 erica // Mar 10, 2009 at 12:46 pm

    (or a spaceship)

  • 10 Ben // Mar 10, 2009 at 4:50 pm

    On the question of whether ‘factoid’ Q&A can make a business, its appears that the SMS services such as texperts or aqa are turning a profit even with the overhead of having human beings answer the questions. Clearly some people have the need for this.

  • 11 Booster // Mar 18, 2009 at 11:30 am

    Looks like Language Computer is finally going public with QnA: check out Swingly.com, whose CEO is the same as the CEO of Language Computer.

  • 12 Wolfram Talks About Wolfram Alpha | The Noisy Channel // Apr 6, 2009 at 10:31 pm

    [...] from a software service. Maybe I underestimate him. Needless to say, I’m a lot closer to my initial skepticism [...]

  • 13 The Wolfram Cometh | The Noisy Channel // May 15, 2009 at 9:33 am

    [...] The initial marketing and buzz offered a level of hyperbole comparable to the hype that surrounded the publication of Wolfram’s New Kind of Science (appreciated NKS by Wolfram and fans) seven years ago. Regulars may recall that I responded by calling it “A New Kind of Marketing (NKM)“. [...]

Clicky Web Analytics