Well, it is true that Siri lives in a cloud, but she really isn't a she. Siri is an intelligent software assistant and knowledge navigator. The voice is totally computer generated.
For those who are not familiar with Siri, it is an Apple iPhone4S application that is a voice recognition program that is being actively advertised and marketed as the next breakthrough in voice recognition software. And, watching the Apple advertising videos it seems like magic. And that is what Apple wants us to think. But if you watch the mistakes that Siri makes on Youtube and elsewhere, you may rethink the magic. But, let me go back a few years to put Siri in context.
The idea of speech-to-text has been around for many years and was once the basis of science fiction. It was a great idea to be able to talk to a computer in a meaningful way, and have it not only understand what we said, but to carry out our verbal instructions. Forty years ago, DARPA, (Defense Advanced Research Projects Agency), an arm of the Department of Defense, not only funded the development of the internet, it also funded several universities to work on artificial intelligence. Voice recognition was one part of the search for computer applications.
Over time, mainstream companies worked with speech-to-text and have developed programs like Talk it Type it, Dragon's speaking naturally, and many others. Most have relied on training to "teach" the programs to recognize individual voices. It is necessary to think that within each language there are dialects, inflections, and voice differences. So, getting voice recognition perfect is a monumental task. Accepting the fact that Siri makes mistakes is understandable. But now comes the bad, the ugly, and the trend.
Apple did not design Siri, they bought it in 2010. At that time it was being planned for the BlackBerry and Android platforms. Apple ended that plan and introduced it for the 4S iPhone and the iOS system. Remember, Apple is not an open source company. They are not going to share this technology with anyone. And, why would they? It is a great marketing tool. If you ask Siri for directions to a resturant, it will list 5 others who are likely advertisers. It is another threat to Google. And, Apple was smart to have Siri work in the cloud and not on the phone or other device. It allows Apple programmers to watch and learn from the user interactions and improve the program.
So, while the Siri program may be an order of magnitude better than current speech-to-text applications, we may never see the result in other applications that are non-Apple. Of course, Apple is not the only smart player. We will see true voice control of computers, but not anytime soon. The trend is that Siri will get better, but we won't know how it was done. Better to curse the darkness than light a candle...
For those who are not familiar with Siri, it is an Apple iPhone4S application that is a voice recognition program that is being actively advertised and marketed as the next breakthrough in voice recognition software. And, watching the Apple advertising videos it seems like magic. And that is what Apple wants us to think. But if you watch the mistakes that Siri makes on Youtube and elsewhere, you may rethink the magic. But, let me go back a few years to put Siri in context.
The idea of speech-to-text has been around for many years and was once the basis of science fiction. It was a great idea to be able to talk to a computer in a meaningful way, and have it not only understand what we said, but to carry out our verbal instructions. Forty years ago, DARPA, (Defense Advanced Research Projects Agency), an arm of the Department of Defense, not only funded the development of the internet, it also funded several universities to work on artificial intelligence. Voice recognition was one part of the search for computer applications.
Over time, mainstream companies worked with speech-to-text and have developed programs like Talk it Type it, Dragon's speaking naturally, and many others. Most have relied on training to "teach" the programs to recognize individual voices. It is necessary to think that within each language there are dialects, inflections, and voice differences. So, getting voice recognition perfect is a monumental task. Accepting the fact that Siri makes mistakes is understandable. But now comes the bad, the ugly, and the trend.
Apple did not design Siri, they bought it in 2010. At that time it was being planned for the BlackBerry and Android platforms. Apple ended that plan and introduced it for the 4S iPhone and the iOS system. Remember, Apple is not an open source company. They are not going to share this technology with anyone. And, why would they? It is a great marketing tool. If you ask Siri for directions to a resturant, it will list 5 others who are likely advertisers. It is another threat to Google. And, Apple was smart to have Siri work in the cloud and not on the phone or other device. It allows Apple programmers to watch and learn from the user interactions and improve the program.
So, while the Siri program may be an order of magnitude better than current speech-to-text applications, we may never see the result in other applications that are non-Apple. Of course, Apple is not the only smart player. We will see true voice control of computers, but not anytime soon. The trend is that Siri will get better, but we won't know how it was done. Better to curse the darkness than light a candle...
RSS Feed
