One might think with that amount of scratch, and the fact that the company Top Dog Bill Gates has talked up speech recognition since Hector was a pup, that they might finally have something that works.
Ummmm. Think again.
The poor designated demo doer at today's Financial Analyst Meeting in Redmond had not much luck at all wowing the crowd with Vista-related speech recognition.
After hushing the multitudes (ambient noise apparently is a no-no in speech reco) he dictated into the machine: "Dear Mom" which translated into " Dear aunt, let's set."
Then: "Delete all" morphed into "double the killer select all."
How might that system translate "Back to the drawing board." ?? Or: "Thank God Bill Gates is in Africa." ????
Directions on Microsoft's Paul DeGroot once explained the Microsoft Research dilemma as well as anyone. Asked how a company which claims to pour money into developing new "innovative" stuff mostly ends up with stuff (finally) that is ahem "derivative" to use a polite term. DeGroot's take: Ninety percent of Microsoft's R&D budget, is probably in "D" leaving just 10 percent in "R."
"D" would cover the such mundane blocking-and-tackling as programming and, well, development versus the perhaps-more-sexy pure research that might result in innovation-- as in the iPod.