One of my goals in writing this blog is to bring to my readers a better sense of how we have gotten to this point in time, with computers and the internet and Google being ubiquitous. So many people are surprised or even disbelieving to learn I've been using the internet since 1972 (called the ARPAnet back then). I think the human journey to this point is fascinating, from the development of the Roman system of roadways to a future we cannot imagine.
I have a host of topics in store, but today I'm going to jump ahead to a piece of breaking news, which fits into that picture somehow. I can't say "this is the future" -- but I do believe, this points to the future in a way that is both clear, and yet familiar, for it refers to capabilities we have long dreamed of for our computer assistants.
Stephen Wolfram, the guy who most successfully brought Symbolic Mathematic Software to the commercial market with Mathematica, and who sponsored MathWorld, is now about to unleash a new effort, which he expects to be even grander.
Of course, like any grand idea, it could be a bust -- at least for now. But if a simple idea like Google can change the world, as it surely has, this is surely worth a deeper look.
I've worked on Macsyma (a predecessor and at times perhaps a rival to Mathematica, a version of which is freely available as Maxima and commercially from Symbolics). And I've worked on Cyc -- Doug Lenat's massive effort to encode all of human common-sense reasoning into a knowledge base. This gives me an appreciation for the scope of what he's trying to do.
The opportunities are there, and growing all the time, as more and more knowledge becomes online, and in better and better forms for automated reasoning. Computers have long been able to do things that surprise and surpass humans. But the scope and variety of what humans do, and want help with, always is a challenge. One of the biggest issues with both Cyc and Macsyma was just the huge variety of types of reasoning humans bring to bear, and expect their computers to bring to bear on their behalf.
And the huge amount of knowledge required to be effective for Cyc made knowledge entry a Herculean task. Yet it would appear that Wolfram|Alpha manages to find a methodolgy and compromise that allows for curated knowledge -- separating the wheat from the chaff, without impassible economic barriers to encoding and capturing that knowledge. I am intensely curious about both the choices made in the approach, and the economics of those decisions, and the domains over which it will be effective.
So far, I have yet to see this, and I have a million questions. Rather than speculate further, I refer you to someone who has apparently seen it, Nova Spivack: Wolfram Alpha is Coming -- and It Could be as Important as Google.
For the moment, there's not much there at the WolframAlpha site -- just a couple tantilizing links, including one to request early access. It's expected to launch in May, 2009.
In the credit-where-credit-is-due department, I came upon this via Matt Marshall's blog entry Wolfram Alpha — it’s like plugging into an electronic brain, which was called to my attention by Don Hopkins on Facebook.
[Addendum: One reason I believe the future lies in this direction (in terms of capabilities) is that my 8-year-old daughter already expects Google to do this, because she's seen simple examples. I myself try it from time to time, knowing it probably won't do anything but a query, but accepting that as a fallback.]
Doug Lenat got a demo and posted his reaction: http://www.semanticuniverse.com/blogs-i-was-positively-impressed-wolfram-alpha.html
Posted by: David Andrews | March 11, 2009 at 08:29 PM
Thank you for the link, David! Very interesting, and really helps to put it into context.
For some reason your link didn't turn into a link. Here is Doug Lenat's commentary on Wolfram Alpha.
Posted by: Bob Kerns | March 11, 2009 at 10:15 PM