Blog Post

Wolfram Alpha: Impressive, But Not the Future of Search, Yet

alphalogoWolfram Alpha, the  “computational knowledge engine” that hopes to complement traditional search engines, debuted over the weekend. Watching Stephen Wolfram demo the service in a screencast last week excited the geek in me: A search engine that can do calculus?

Try the app yourself, and you’ll find that it has an easy-to-use interface and can provide some very interesting results. Want to know the distance between San Francisco and Tokyo? Alpha can do that. It will add some other interesting facts, too, like the populations of the two cities. Enter a date (like my birth date) and Alpha will return a bunch of facts about that date. It can also do some even more impressive things. Enter any equation (including calculus) and Alpha will plot out an answer for you. It can even display geometeric shapes and show you which notes to play to make a chord. On the results pages you can  drill down to get more interesting information, change the units, or fiddle with the inputs.


Unfortunately, once you play with the app for a little while, you’ll realize that there are some large gaps in its abilities, which is frustrating, especially given how the demo made it look like you could type in nearly any question and get back a useful result. However, as Wolfram says, Alpha “is the first step in an ambitious, long-term project to make all systematic knowledge immediately computable by anyone.” It’s important to realize that Alpha is still at an early stage: Whether it is ultimately successful depends on if the team behind it can fill in those gaps.

Knowledge engines like Wolfram Alpha and True Knowledge (previously covered on WWD), are not intended to be “Google killers,” (s goog) rather they could act as a useful complement to traditional search engines. I imagine that Wolfram Alpha, in particular, could be a useful tool  for school and college students, journalists and bloggers.

Wolfram Alpha suffered from traffic overload errors over the weekend as people rushed to try the service (the computing power requirements of the service must be huge), so if you’d like to play with it, you might need to be a little patient.

Share your thoughts on Wolfram Alpha in the comments.

11 Responses to “Wolfram Alpha: Impressive, But Not the Future of Search, Yet”

  1. tdhurst

    This is an exciting step forward for cloud computing. Very stoked to see what this means for future technology and collaboration.

  2. @Dima I’m not sure that Google will be able to rush out something does exactly what WA does, unless the have had something in the works for a long time already. It’s not exactly a trivial system and look how long it took them to make GrandCentral into Google Voice…

  3. First thing I looked at is the API and they have it.

    This really an ultimate example of Web 3.0 and cloud services and it’s something really new since Google. This is why some people say it’s Google killer – of course it’s not and of course I would expect same thing done by Google pretty quickly and that will be WA killer :)

  4. @Yuri “at least not what we call search” – exactly right. With tools like Wolfram Alpha at our disposal, we’ll be able to do a lot more in the future. Will we still call it “search”? Probably. The definition will just have been widened to take in knowledge-based tools like Alpha, too.

  5. @Charles – yes, that’s a good point. It does present answers very authoritatively with no clue as to where the info came from. It would be good to have links back to the original source of the info (or at least a reference). I’m not sure how easy that will be for them to do, though. {EDIT} Actually, I take that back — it has a “Source information”” link at the bottom with references.

  6. Charles Young

    …however, there are other considerations. Once the initial excitement calms down, people will start to think a little more critically about the service. One obvious question is why should we trust the information WA provides when the provenance of the data is not clear? WA authoritatively states statistics, facts and figures. Some, obviously, are beyond question. Others are highly questionable. With Google or Wikipedia, at least you get some cues to suggest how trustworthy the data is – obviously a long way from being infallible but better than nothing. With WAA, all you get is a ‘Source’ link at the bottom of the page with a long shopping list of references. Early experimentation with WA shows that it does indeed provide some ‘facts’ which are simply not accurate. A recent Radio 4 (UK) programme discussed the often quoted statistic that the biggest cause of mortality for women between 19 and 44 is domestic violence – turned out that this stat (or variations of it), which has been quoted in official government policy docs, etc., originally substituted the word ‘mortality’ for ‘morbidity’ (ill-health), and is not really based on any actual statistical information at all, but is more of a guess. Will WA result in similar ‘true lies’?

  7. Wolfram Alpha can’t and won’t be the future of search for simple reason: it is not a search engine, at least not what we’ve been calling a search engine. It may look and feel like one, yet it’s something different. In essence it’s a growing knowledge base with a solid set of search/analysis capabilities. And it become quite successful as such, it users start coming in with right expectations (and “Google killer” isn’t one of them).

    Meanwhile it’s good at things you describe in the post as well in engaging in nearly philosophical dialogs, answering questions like “what are you?” and greeting you “Hello, human”: