If you're on a phone you might get tricked into thinking it is paywalled by the giant advertisement for the print replica of the article but if you scroll down past that you can view the article in its entirety.
Turn off Javascript. Click the uBlock symbol and the symbol that looks like an empty closing tag. Burning extra cpu cycles to read the news always seemed strange to me anyway.
In a way, this sorta seems like it was an inevitability; humans are a very social species. It only makes sense that our advancements in tech would, first and foremost, promote that.
> Martin E. P. Seligman, professor of psychology at the University of Pennsylvania [...] has been monitoring the temperature of emotional incidents on the computer network to see how they differ from "real time" emotion. The constraints of the form tend to flatten anger and push toward civility, he finds.
I wonder if this was just population effects. The group of people communicating over the internet in 1993 is not the same as the group of people doing so today.
>"Oh," he says. "You mean why no voice dialing? Saying 'Get Fred' or 'Get Sam' or something? I don't have a voice dialer." He will soon, surely. Any day now, he will be able to pick up a phone anywhere and say "Call home" to a network that will recognize his voice and look up his home number (verifying his credit by means of his voiceprint).
Only way I see something like that working is AI/ML and you'd need to vastly expand voice bandwidth past what PSTN is in order to have enough information for it to work. And even then who's to say people's voices, verbal idiosyncrasies, accents aren't still too similar to differentiate?
For whatever it's worth, I've seen AT&T experiment with voice recognition for authentication. One of the Achiles' Heels of the system is you can collect a lot of the samples you need to bypass it from peoples' voicemail greetings. Especially if like for some jobs, they're required to re-record the greeting every day, i.e. "Today is Monday, June 17th, and I'm in the office..."
That being said, linear predictive codecs used in mobile/lower bandwidth IP networks have damaged the sound quality of the PSTN quite significantly (some are even vocoderized). If you're just dealing with, say, g.711 in wireline networks (when this was written, aside from g.722 in some ISDN networks, g.711 was used almost exclusively; GSM was very much a niche in its infancy), you might be able to squeak by with enough data to reliably recognize people.
18 comments
[ 2.8 ms ] story [ 50.1 ms ] threadI wonder if NYT has the paywall differently depending on your region.
> Martin E. P. Seligman, professor of psychology at the University of Pennsylvania [...] has been monitoring the temperature of emotional incidents on the computer network to see how they differ from "real time" emotion. The constraints of the form tend to flatten anger and push toward civility, he finds.
edit: at least the terms, more so than the behaviors they describe
In hindsight that's a... yesn't.
That being said, linear predictive codecs used in mobile/lower bandwidth IP networks have damaged the sound quality of the PSTN quite significantly (some are even vocoderized). If you're just dealing with, say, g.711 in wireline networks (when this was written, aside from g.722 in some ISDN networks, g.711 was used almost exclusively; GSM was very much a niche in its infancy), you might be able to squeak by with enough data to reliably recognize people.