Now that we know what most of us suspected--that Alexa et al listen more than is altogether proper, perhaps the next generation will double-down on that feature and be called Boswell.
Right now the computing power needed to process natural speech into actionable instructions is only available in relatively large systems and so necessarily links individual speakers to the Amazon/Googles system via open TCP/IP internet.
In my early adulthood, the computing power needed to process keystrokes into editable documents was only available in relatively large systems and so necessarily linked typists to 8 inch floppy disks and "Winchester" drives owned by the university or corporation via proprietary LANs.
Right now the computing power to process keystrokes or taps on a display into editable documents can be purchased as a stand-alone device at WalMart for a day's pay, carried in a pocket, password protected, and destroyed -- if necessary -- with a dunk in the kitchen sink. The device ALSO connects to the phone and internet systems, but you don't have to do that.
It seems safe to predict that "soon" a cheap, portable, personal-private computing device will be sufficiently powerful to process speech-to-text with local, on-board, resources. Technically, that's a safe prediction. Whether or not the alliance of Government, Google, Apple and Amazon will allow such devices to be brought to market is political rather than technical and so not such a safe question upon which to prognosticate.
2 comments:
Why would anyone purpously hire one's own live-in Stasi agent?
Right now the computing power needed to process natural speech into actionable instructions is only available in relatively large systems and so necessarily links individual speakers to the Amazon/Googles system via open TCP/IP internet.
In my early adulthood, the computing power needed to process keystrokes into editable documents was only available in relatively large systems and so necessarily linked typists to 8 inch floppy disks and "Winchester" drives owned by the university or corporation via proprietary LANs.
Right now the computing power to process keystrokes or taps on a display into editable documents can be purchased as a stand-alone device at WalMart for a day's pay, carried in a pocket, password protected, and destroyed -- if necessary -- with a dunk in the kitchen sink. The device ALSO connects to the phone and internet systems, but you don't have to do that.
It seems safe to predict that "soon" a cheap, portable, personal-private computing device will be sufficiently powerful to process speech-to-text with local, on-board, resources. Technically, that's a safe prediction. Whether or not the alliance of Government, Google, Apple and Amazon will allow such devices to be brought to market is political rather than technical and so not such a safe question upon which to prognosticate.
"Soon" is, IMHO, about 4 years.
Post a Comment