What Do You Think of Wolfram|Alpha?

Now that Wolfram Alpha is live, we can start to see if it fills a niche in our daily lives. When I used it, about half my queries were input that the engine did not understand. There were gaps I didn't expect. If it knows how tall a regular object is, like the Great Pyramid of Giza, why didn't someone input the length of the sides?

It did parse a question about the death of Bloody Mary, and gave me the right queen in English history.

Curiously, it know information that it doesn't seem to have connected. I asked how many ounces in a kilo of flour, and it didn't understand the input, but it gave me kilo of flour to click on. And at the very end, beyond the nutritional value of a kilo of flour, it said it had a mass of 35 ounces. It also had a serving volume of 58 fluid ounces. Now why didn't it connect those facts associated with kilo of flour with the question how many ounces in a kilo of flour?

So even within the information available on Wolfram|Alpha, there is a tendency to be too literal or straight line with its answers, and not make a web of connections. That is a problem in the way the engine works, not the information inside it.

What was your experience?

I'm finding similar results. While it's obvious that the engine is still in the new stages, I do like the concept and have high hopes for future progression.

I've also noticed that Wolfram Alpha doesn't seem to like "extra words" in queries, particularly question words. Maybe that will be something worked on, but it also may just be a thing that you get used to. It didn't take many sample queries before I got the hang of what type of words seemed to be throwing the engine off, and I just adjusted the phrasing of my input.

Either way, I definitely see where it has its use and am excited about the continued development. However, as for all the buzz about the new engine being a "Google Killer" - that doesn't seem to be the case at all. Wolfram Alpha and Google (as well as other search engines) are two different entities even if they do have some overlap. No matter how mature Wolfram Alpha becomes, the information that it returns is going to be based on facts and verifiable information. If you want to "dig" behind that information, look for more opinionated items (like product reviews), or be on top of late-breaking news and announcements, Google and other sources are still probably going to be the best bet.

I had the same experiences - especially with extra words. When I got the hang of searching like it was 1999, it was great information, nicely formatted and usable. Not a Google killer in my opinion, but a great search engine to use when you want stats or similar data all nicely bundled together - like the number of archivists in the US and the job trending details that go along with it. Will it be my 'go to' search engine like Google is? Probably not. But will I use it again? Probably so.