I’m seeing a growing number of errors in Facebook posts, tweets and comments that look an awful lot like speech recognition errors. That tells me both that we still have a ways to go before the technology is as good as we’d like it to be, and that a lot more people are using it.
Maybe the tech will catch up, with some combination of heuristics and better audio discrimination. But I’m not sure I want it to. Maybe if speech recognition had been around 30 years ago, Michael Stipe and REM might have produced songs with discernible lyrics. (By the way, my Mac’s voice recognition interpreted “Michael Stipe and REM” as “my goal star you and are in the am”. Which could actually have been from Automatic For the People.)
For now, we’re stuck with the digital equivalent of an aging uncle who’s in denial about the fact his hearing is going. And I’m going to have to run; my kids are yelling something from the living room, but I’m damned if I can make it out.