APLawrence.com -  Resources for Unix and Linux Systems, Bloggers and the self-employed

Wolfram Alpha - Incredible!

I mean "Incredible" as in I don't believe it for a minute. Stephen Wolfram says he will shortly release search software that actually "computes" answers rather than just looking up matches in a database.

Someday search will understand what you want, what answers it actually has, and the relationship between the two. We're not talking about "find me stuff about fixing a forgotten password". That's easy for Google to find. It's not so easy for Google to answer factual questions, but Stephen Wolfram says his software will be able to do just that. Here's Nova Spivack describing his impressions:

Instead, Wolfram Alpha actually computes the answers to a wide range of questions -- like questions that have factual answers such as "What is the location of Timbuktu?" or "How many protons are in a hydrogen atom?," "What was the average rainfall in Boston last year?," "What is the 307th digit of Pi?," "where is the ISS?" or "When was GOOG worth more than $300?"

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.

My reaction is simple: I doubt it.

Oh, not that we won't someday have software like this. We definitely will. But to live up to the implied promises, this would have to be very advanced AI. Extremely advanced. So advanced that if Stephen Wolfram really has done what he claims, our world will be radically changing as a result.

I'm suspicious on two counts. One is that quantum leaps in AI are suspicious enough by themselves. Perhaps even more suspicious is the bald faced marketing tactic of announcing that they WILL do this two months from now. That smacks of vaporware and perhaps an upcoming effort to get funding. If you really have something this incredible, you don't need to build buzz. Just release it: I and the rest of the world will fall over in amazement.

I suspect instead that Wolfram has nothing more than the kind of basically useless efforts that are out there now. Try MIT's START, for example. Just now I asked it "When did SCO first sue IBM?". It doesn't know who or what SCO is and certainly doesn't know that they sued IBM. I asked "When did Reagan first speak of supply side economics?" and got referred to Wikipedia. I got better and quicker answers from Google and I bet that's just what is going to happen once Wolfram Alpha goes live. There WILL be questions it can answer, but like START, they'll be very limited.

And if not? Well, then we are well on our way to the kind of robotics envisioned by Asimov et al. so many years ago. The world will change, and rapidly. Future historians will mark May 2009 as the beginning of the Robotics Era.

Don't hold your breath.

Got something to add? Send me email.

(OLDER)    <- More Stuff -> (NEWER)    (NEWEST)   

Printer Friendly Version

-> -> Wolfram Alpha - Incredible!


Increase ad revenue 50-250% with Ezoic

More Articles by

Find me on Google+

© Anthony Lawrence

Sat May 9 14:40:36 2009: 6350   TonyLawrence

Seems that Google doesn't believe it either: (link)

Sun May 24 12:06:57 2009: 6399   TonyLawrence

Wolfram is dangerous.

Example: I search for Sharon, Massachusetts. Wolfram tells me that it is at 42.11N, 71.19W, which is correct. It says the population is 5941 people - that was probably true sometime in the 50's.

If I search for "zip 02067" it finds that town again, but now the population is more accurately said to be 18049 people.

You can't trust Wolfram. I'd take Google for that search any day.

Mon May 25 19:36:55 2009: 6404   AndrewSmallshaw

I agree entirely. Wolfram Alpha promises much but delivers little. I want a good next-generation search engine to recognise the likely context of my search phrase and present details accordingly. However, it should also break down its search results by context rather than by rank so I can quickly find the kind of information I am looking for.

Search for something obvious: hydrogen. It actually works very well. It recognises that in isolation I am probably referring to the chemical element. If that is wrong it presents me with a list of alternatives. However, it got that right so it presents me with a list of various properties of hydrogen. There's no narrative description, just tables of raw data. The data is probably what I was looking for but some additional context would be welcome. In particular some of the tables are missing interpretive details. The table on abundance is by mass rather but number of atoms or by volume for instance which really needs to be stated explicitly. Overall it does a reasonable but not exceptional job. However you would expect that - for anything presenting itself as a general reference the chemical elements should be a high priority as they are pretty fundamental. However, it is not really a search engine - all it has done is present me with its page on hydrogen. It has not provided references where I can find out more.

Now try something slightly less obvious but still pretty mainstream. I went for HMS Pinafore. Not because I am a huge fan of opera, but because it is the first thing that came to mind that is slightly more obscure yet still easy to see what the ultimate search engine should provide. An average person will immediately recognise it as a musical. Those with a more literary bent will immediately think Gilbert and Sullivan. Nevertheless it should be obvious without additional context that I am referring to the play.

What should the ultimate search engine provide? It should break down answers into distinct categories so I can decide what I want. I might be after primary source material - a copy of the script and music for instance. I might be looking for a critique of the play. I may be interested in current theatre productions of the play. I may be interested in film, radio or television adaptations, or reviews of any of these. I probably would welcome being told if any of those adaptations were going to be broadcast over the next few weeks. There's a lot of things that I could be looking for and if Wolfram Alpha could categorise results and present them to me in an ordered manner it would be an immense tool.

What does it do in reality? It doesn't recognise the term at all. Anything presenting itself as a comprehensive tool should do. This isn't one of the millions of articles on Wikipedia that have article yet don't deserve them - it is something that many genuinely are going to want to look up for reasons other than to make sure that there is coverage of it.

Ultimately for all the AI hype it is evident that there is a lot of human work going on behind the scenes. While that is still the case coverage always is going to be patchy. Sites such as Wikipedia only work at all because of the huge number of people working on them, and even then it is littered with glaring errors (which if you attempt to fix usually turns out to be a huge waste of time). For all its faults it isn't feasible for a private operation to attempt something similar without real automated intelligence behind it. Wolfram Alpha lacks that so for now it is merely an advertisement of what Mathematica can (not) do.

Mon May 25 21:04:43 2009: 6405   TonyLawrence

For breaking down search, Google's new "Wonder Wheel" is interesting (hit "Show Options" on any search and then look for "Wonder Wheel" in the sidebar). It does a nice job with "hydrogen" and "HMS Pinafore" :-)

Kerio Samepage

Have you tried Searching this site?

Support Rates

This is a Unix/Linux resource website. It contains technical articles about Unix, Linux and general computing related subjects, opinion, news, help files, how-to's, tutorials and more.

Contact us

We're terrible animals. I think that the Earth's immune system is trying to get rid of us, as well it should. (Kurt Vonnegut)

This post tagged: