There is nothing more important, either in the long or the nearterm, for information science and technology than funding the C-Prize. It solves the AI problem if the AI problem is soluble.

TIBET(tm) is the web's nearterm future.

Relation arithmetic is more important than TIBET(tm) but further out (and not as important as the C-Prize).

If I were in Ray Ozzie's shoes I would apply something like the C-Prize to the entirety of MS's source code base.