August 10, 2007

A bet against speech recognition

From blogger Russell Shaw at ZDNet: Man endures thumb surgery to better enable iPhone use. The subject of the surgery, one Thomas Martel, tells a reporter:

"Sure, the procedure was expensive, but when I think of all the time I save by being able to use modern handhelds so much faster, I really think the surgery will pay for itself in ten to fifteen years."

When I read this story I couldn't help but think back to ten years ago, when Bill Gates was predicting that speech would be a standard part of the PC's human-computer interface within a decade. Blogger Matthew Paul Thomas has a collection of Bill Gates quotes on speech recognition showing how this projection has evolved over time (with links to sources), e.g.

Bill Gates, 1 October 1997: “In this 10-year time frame, I believe that we’ll not only be using the keyboard and the mouse to interact, but during that time we will have perfected speech recognition and speech output well enough that those will become a standard part of the interface.”
Bill Gates, 24 March 1999: “Speech recognition … I don’t think you’ll see dictation as something that most people will use in the next couple of years. The extra processing power, getting the extra memory I think has us on a track to provide that, but for most people, I think it will be more like a five-year time frame before that’s a standard way of interacting.”
Bill Gates, 26 March 2001: “Because of where we’re going with real-time communications, including the instant messaging that will be included in Windows itself, voice annotation, voice communication and speech recognition are becoming mainstream capabilities. And so we believe that virtually all PCs should have that right out of the box.”
Bill Gates, 25 February 2004: “Now, with speech it’s not as easy. Speech is another one that will be solved, and will be solved for a broad range of applications within this decade.”

Speech technology has definitely been improving -- beyond the formal studies in the research community, witness all the interaction we can do these days to get telephone listings, driving directions, movie tickets, etc., not to mention the whole variety of successful speech transcription products out there in the market. But will the technology advance quickly enough to make Mr. Martel regret his choice? Time will tell.

[Hat tip to Greg Schnitzer]

[Update: it turns out that the surgery story was a hoax. Oddly, the newspaper in which it appears says that it "features stories from the paper as well as other news, satire and commentary", but nothing at the site gives any indication of which is which. Nonetheless, it was an interesting opportunity for a ten-year perspective on Bill Gates's predications...] Posted by Philip Resnik at August 10, 2007 01:48 PM