Wolfram Alpha, the new computational knowledge engine

By David Barnes, Director, Business Intelligence

I was outrageously excited when I heard about Wolfram Alpha, the new computational knowledge engine designed by Stephen Wolfram. Finally Google has competition I thought. But this was a short-lived notion and confirmed to be false once I ran a hand full of queries on the Mathematica powered engine. As stated on Wolfram’s homepage, the goal of the project is to make the world’s knowledge computable (no small task and certainly something humans have been driving towards ever since the advent of the transistor).

Wolfram boasts some 10 trillion+ pieces of data, 50,000+ types of algorithms and models, and linguistics capabilities for 1000+ domains. Yet when I “ask” Wolfram for the results of last night’s Cubs game, Wolfram responds; “WolframAlpha isn't sure what to do with your input.” Partially accepting blame for the results, I adjusted my syntax. In fact I adjusted it 4 times until I was left with the most generic term: “Chicago Cubs.” But at last, results! I scanned the results page - not quite what I was looking for. Where are the scores? Where are related articles, drill downs, etc? Unlikely Google will return zero results for this query.

So what exactly is Wolfram good at? Enter a name and Wolfram will give you census data including ethnicity breakdown by surname. Enter an airport code and you could quickly educate yourself on the number of runways an airport has (don’t expect to see a list of departures and arrivals). Enter most words and Wolfram will oblige with a definition (don’t stump it with complex acronyms like L.A.S.E.R.). In other words, nothing you won’t find on Google or Wikipedia.

I know Wolfram is a new release. But releasing a product that claims to be a computational knowledge engine that falls over on 7 out of 10 questions is just frustrating to the end user. In the meantime, I’ll stick with Google.