Wolfram Alpha is gonna be pretty fucking neat. From the creators of Mathematica, Stephen Wolfram and Wolfram Research, the system proposes to be “an authoritative source for data” from the sciences, social science, finance, pop culture, and damn-near anything else that can be quantified and calculated. Ask it for the volume of your favorite lunar crater, expressed in wheels of parmesano reggiano, and it’ll run the numbers. Math, FTW!
But as far as I can tell from Wolfram’s recent presentation at Harvard, Alpha will tell you the source of the data, and will sometimes show you an example formula for a type of calculation (how to figure the volume of a cylinder, etc.), but it will not show you the actual calculations necessary to arrive at the answer.
Instead, we’re supposed to trust the system. Alpha’s code is so complex, says Wolfram, that it would be vastly inefficient for a mere human to read the steps. The more sensible approach is to “try to do the best QA that we can.” (QA = quality assurance, a.k.a. testing) And I’m certain that he’s right — that thing has got to be a bear to test.
If your mother says she loves you, check it out
The problem is, if I want to use Alpha’s answers in a news article, scientific paper, or anything else requiring an authoritative source, I need to know that it’s right. Unless we arrive at some kind of universal agreement among scientists, academics, mathematicians, and everyone else that Wolfram’s creation is always right, I can’t believe Alpha’s answers if I can’t test them.
I’m baffled by this omission. Alpha can do extraordinarily complex work, but it wouldn’t pass high school physics. You’ve got to show your work! Scientists validate hypotheses by repeating experiments and comparing the results. “Trust me” is not an authoritative answer.
“An authoritative source for data” is misleadingly simplistic. If it does what it’s supposed to, Alpha is more like “an authoritative source for quantitative thought” — the ultimate almanac, complete with a staff of uber-geeks from every field worth researching, backed by machines capable of turning around complex calculations in microseconds.
It’s cool as hell.
It’s a pocket calculator for *everything*.
It’s the realization of Leibniz’s characteristica universalis.
But is it correct? Only one way to tell. Mr. Wolfram, will you please show your work?