Microsoft’s Bing—Introducing One of Wolfram|Alpha’s First Commercial API Customers

November 11, 2009

Having announced the Wolfram|Alpha API just over three weeks ago, I am pleased to share in announcing Microsoft’s Bing decision engine as one of our first API customers.

Starting today, Wolfram|Alpha’s knowledge, computed from expertly curated data, will enrich Bing’s results in select areas across nutrition, health, and advanced mathematics. Wolfram|Alpha provides immediate, unbiased, and individualized information, making it distinctly different from what has traditionally been found through web search. By using Wolfram|Alpha, Bing recognizes the complementary benefits of bringing computational knowledge to the forefront of the search experience.

By using our API, Bing will be able to seamlessly access the tens of thousands of algorithms and trillions of pieces of data from Wolfram|Alpha, and directly incorporate the computations in its search results.

Microsoft’s initiative and interest in Wolfram|Alpha began earlier this year. In fact, there is an interesting story that circulates within our walls around some of our early discussions with Microsoft.

Highlighting examples of Wolfram|Alpha to the most senior executives at Microsoft, Stephen Wolfram entered the query “2^2^2^2^2”. Upon seeing the result, Bill Gates interrupted to say, “What, is that right?”

A profound silence fell over the entire room.

Stephen replied, “We do mathematics!”

Amused, Stephen, Bill, and the other executives dissected the calculation and determined that the result was, indeed, correct. Microsoft continues to pepper us with questions to this day, reflecting its continued enthusiasm in Wolfram|Alpha.

We applaud Microsoft’s vision and foresight in augmenting their search with Wolfram|Alpha, and we look forward to a fulfilling and productive partnership.

27 comments to “Microsoft’s Bing—Introducing One of Wolfram|Alpha’s First Commercial API Customers”

  1.  

    I saw ads for Bing doing functions in their website and stuff. Quite interesting that I found out why. :D

  2.  

    Congratulations on the partnership! :)

  3.  

    I wonder how much strain this will add to the WA servers. I mean, Bing is huge…

    Tomi
  4.  

    I went to bing’s website and typed 2^10 and it gave me web results… It’s not functioning yet. It would have been wiser to implement with a system already in place, such as Google and how it already does calculations, except do it with the wolfram API for greater expansion.

    GamefreQ
  5.  

    There’s a bit of a difference between Bing’s answer and WA’s answer.

    Bing = 65,536
    WA = Well, a lot more than that…

  6.  

    … “What, is that right?” …

    I don’t get it…?

    Kristjan
  7.  

    Well, it depends on operator associativity.

    65536 = (((2^2)^2)^2)^2
    WA’s answer = 2^(2^(2^(2^2)))

    I worked the first one out in my head. The second one was a bit harder.

    Matt Powell
  8.  

    Bing says 2^2^2^2^2 == (((2^2)^2)^2)^2

    Wolfram says 2^2^2^2^2 == (2^(2^(2^(2^2))))

    Vastly different answers. Which is correct?!

    Michael
  9.  

    To be fair, there is a difference between the following three lines
    2^2^2^2^2
    2^(2^(2^(2^2)))
    (((2^2)^2)^2)^2

    I guess a human mathematician interprets the first line as equivalent to the second, but computers , and computer geeks like Bill Gates, consider the first as equivalent to the third. Too bad this blog glossed over the “dissection” of the calculation and just left it as Wolfram Alpha right; Bill Gates wrong, but hey I guess I can “do mathematics” too.

    Ash
  10.  

    It’s been a long time since any mathematical education, but I don’t understand why the formula calculates out as 2^(2^(2^(2^2))) rather than (((2^2)^2)^2)^2. That’s a massive difference & pretty counter-intuitive. It was my understanding that equations play out from left to right unless coerced by an overriding operator.

    Timothy King
  11.  

    thats great.. bing becomes more powerful..

  12.  

    ^ always operates right to left, so your first example is always wrong, and the second is always correct.

    power
  13.  

    PLEASE TELL ME WOLFRAM|ALPHA WILL NOT CHANGE!!!!! I’d hate to see something so perfect get ruined by bing. I’m happy that WA is helping bing out, but don’t change this site for the worse….. please.

    Andrew
  14.  

    If x^y^z were (x^y)^z, then you would just write x^(yz).

    (Think of e^x^2, which occurs all the time, and never means e^2x.)

    Therefore x^y^z can only mean x^(y^z).

    samuel black
  15.  

    I tried the query in everybody’s favorite search engine (Google). It does not seem to handle 2^2^2^2^2. But it did 2^2^2^2 nicely, and also disected it:

    2^(2^(2^2)) = 65 536

    http://www.google.com/search?hl=en&safe=active&q=2+%5E+2+%5E+2+%5E+2+&aq=f&oq=&aqi=g2g-m1

    WA will definitely bring power of maths to Bing. The question is, will WA allow it to be licensed by others too (e.g. what if Google does a partnership too)?

    chirag
  16.  

    Power and Samuel Black are right, of course. I probably would have made the same mistake as Bill G., though.

    Telanis
  17.  

    SO — Bing is incorrect, Wolfram Alpha is correct.

    That is not so surprising.

    What is surprising is that once Bill Gates and the other executives saw that Bing was wrong, that they did not take immediate action to fix the error.

    Can anyone here explain why they would not care to have Bing be correct?

    Joel
  18.  

    As for visualizing the problem while reading it left to right, just substitute some of the ^2’s with other numbers… 2^3^4^5^3. Reading it left to right, you’re raising 2 to some power that you need to figure out, and you don’t do that by using the base 2 right off the bat. That first number, when the operators are all ^, must be the base. Once you realize that, then you can see why you don’t just start off from the first 2. That’s why ^ must operate from the right to the left, so you solve for the power that you’re raising the base to.

    Mike
  19.  

    Well, one week has passed since the last comment, and bing.com doesn’t seem to be generating WA answers either to the exponentiation query or the other cited useful area of nutritional facts. So it’s back to using Giggle for normal web trawls, WA for really fun research, and put Bonk back in its box for another year.

    Rob
  20.  

    What is the resullt of this info ?
    This a expriment or for fun only?

  21.  

    You’re taking the power of a power of a power of a power of a power of a base.

    Bill Gates
  22.  

    I tried 2^2^2^2^2 in Bing and gives an incorrect answer. It does provide a link to the (correct) calculation by Wolfram|Alpha:

    http://screencast.com/t/MTVjNTJmZGIt

    Oscar C.
  23.  

    If bing will show web results then the webmaster will be happy about it, but if bing will give result like google do.. then webmasters will be a little sad about it. Users will really love wolframalpha because it has all the different types of results.

  24.  

    Congratulations

  25.  

    To this day, Bing returns the wrong result.

  26.  

    Is there a good laymen’s source guide to all these parameters?

Leave a Comment