Watch Out Siri: Amazon Buys Voice Recognition Technology

Page 2 - Seeking answers? Join the Tom's Guide community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
@ericburnby

try not to sound a little disappointed, but Siri has very minimal AI functions, although 'she' may have descended from an AI project, SIRI in itself is closer to a user preference algorithm than an AI and probably outsource alot of the heavy lifting to wolfram alpha, SIRI converts the voice information determines the context and parses it on

now saying that, i do know a few folks who can make SIRI feel like real intelligences....
 
^ And I know a few people who can make Voice Actions appear to be the same as Siri (mostly fandroid idiots who know nothing of the underlying technology, like yourself).
 
The thing that missing is AI's inability to handle follow-up questions using pronouns. Every question has to be very explicit.
 
Speech recognition is nothhing more than a keyboard alternative. The reason few people use it is because it's a very poor keyboard alternative. If keyboards were as big a resource hog and was as fussy and unreliable as speech recognition software few people would be using computers. Keyboards function admirably while causing no trouble and using trivial system resources. Until speech reconition is the same it's going no where. I use Dragon Naturally Speaking but only for occiasional transcribing jobs because it's a perpetual trouble maker on my computer unless it's freshly booted.

I should be able to buy a peice of hardware, a card, with all the physical and software resources Dragon needs to function so that it's demand on resources used for other applications is as trivial as my keyboard. But that's not enough. There need to be bots built into the package so users can verbally train the speech recognition package to learn and execute fuzzy language commands, the kind of casual speech typical between a boss and an a well trained personal assistant. The bot needs to have enough "personality" and social finesse so it can earn the support and acceptance of it's users. All the techniques and technology needed for speech recognition to be a smashing success are at hand. It's not happening because of unwillingness to address ALL the issues that matter to users.
 
Status
Not open for further replies.