Wolfram Alpha and the expectations game

Wolfram|AlphaStephen Wolfram made a post on his blog last week that sent hearts in the geekier corners of the Internet a-flutter. If Steve Jobs is the prototypical rock star geek, known the world over for his genius and one-dimensional wardrobe, then Mr. Wolfram is the influential indie rock act, influential in all the right places if not quite a household name. Wolfram's post unveiled his latest project: a search engine called Wolfram Alpha.

Within a couple of days, news of Wolfram's announcement started to spill over into the mainstream media. "British scientist developing 'Google killer' search engine" exclaimed the Telegraph. Fox News reported that Wolfram Alpha was positioned to "challenge Google". A ZDNet blogger asked if Wolfram Alpha was just "lots of hype".

The "hype" in question was, of course, entirely the creation of the media and blogosphere. All Wolfram did was announce his project (on his blog no less, with nary a press release to be seen) and he notably did not mention Google once in the entire post. News of a new search engine caused some tech writers to begin rubbing their hands together and cackle, hoping that they would have another delicious disaster story in the vein of last year's Cuil implosion.

There's a reason that Wolfram never mentioned Google in his post - Wolfram Alpha is not competing in the same universe: it's a step towards semantic search and semantic computing. A semantic search engine would actually work the way that beginning web users expect search engines to work - by answering questions. As Wolfram explains in his post:

The way humans normally communicate is through natural language. And when one’s dealing with the whole spectrum of knowledge, I think that’s the only realistic option for communicating with computers too.

Experienced web users know to tailor their queries around Google's capabilities: drop prepositions and focus on unique keywords, learn how to use operators. The end result is that an effective Google search string doesn't resemble a proper English sentence in any way. By contrast, Wolfram hopes to create a semantic search engine that understands your question.

The semantic computing part is why Wolfram Alpha isn't going to be putting Google out of business though. Wolfram Alpha will answer your questions not with pages of web links, but with a single definitive answer. This fundamental difference leaves plenty of room for Google. If you want to know the boiling point of the transition metal mercury, then Wolfram Alpha will be the place to turn. If you want to know which Italian restaurants are within 2 miles of your office, or what news commentators are saying about Bernie Madoff, Google will still be your best bet.

So until Wolfram Alpha is actually released, take what you read about it with a grain of salt. You can sign up for the Wolfram Alpha Beta (ha!) at http://www.wolframalpha.com/.

Comments

The problem with Wolfram's

The problem with Wolfram's idea is that all it takes is someone to ask the following question to destroy the entire internet:

"What is 1 divided by zero?"

We can only pray that there

We can only pray that there is some sort of failsafe to prevent that from every happening.

Or maybe that's what the Large Hadron Collider was for.