Stephen Wolfram is known both for his monumental The New Kind of Science and Mathematica developed by Wolfram Research. Now he has entered search engines’ domain with his new product. It’s called Wolfram|Alpha.

Google is really good with providing you with a list of websites which contain the answer to your query. Wolfram’s approach is different though. Instead of the website providing the answer his aim is to provide you with the answer itself. When you type in Poland’s GDP or rule 110 to Wolfram|Alpha it produces accurate result and there is no need to go anywhere else on the web. Google directs you to Wikipedia’s page as a first result in case of CA rule and gives you the number in case of GDP, but not accompanied by any additional data. Yet when you try terms which are slightly more vague, say “mass customization”, Google produces accurate findings (mass-customization.de on 2nd place), while Wolfram|Alpha get’s confused.

This is understandable since Wolfram|Alpha is not a tool to browse the web, but a tool to browse knowledge. To say it in the words of it’s developers, it’s a tool for new kind of knowledge based computing. It’s creators say that it is not a search engine (sorry for the misleading title), but a computational knowledge engine which aims to provide you with accurate answer to a query about any kind of a fact. For every query Wolfram|Alpha computes the answer composed out of information from numerous sources. List of those sources is available next to the answer in most of the cases.

Whether this can be called artificial intelligence able to “understand the meaning” of it’s users’ queries is a question somewhat outside of my scope of interest. What’s really interesting for me is the amalgam of algorithms behind it (the number of algorithms and models harnessed exceeds 50 000 already).

I have learned about Wolfram Alfa thanks to Edwin Bendyk’s article in Polish weekly magazine Polityka.