I worry more it will kill the public’s interest in speech rec
I have been reading with great interest all of the back and forth chatter on the web regarding Apple’s Siri and its merit as a Google killer and as I read these articles I wonder if many of the people writing these stories have followed the speech technology space these past decades. In 2000, I ran a show called Communications Solutions and on stage I did a live demo of General Magic’s Portico calling it a killer app – the next big thing. It was incredible to me and it worked pretty similarly to Siri. Sure, Apple’s speech rec “wizard” Siri has enhanced functionality allowing you to do information searches but in a decade you would expect such advancements, right?
But one has to wonder why is Siri the Google killer and why now? After all, those of us in the tech space in the mid-nineties remember voice recognition phenomena Wildfire, one of the hottest pre-Internet startups that flamed out as spectacularly as it launched.
These general purpose speech technology companies seem to be the fastest rising and falling companies I’ve seen in just about any industry.
The point is voice rec and associated technology has been around for decades in call centers where the technology can save a company millions and in many cases hundreds of millions of dollars per year. And yet, the technology in the contact center isn’t perfect – it is far from it. Ditto Portico and definitely Siri.
In fact, the last Windows Mobile device I owned about five years ago allowed me to tell it via speech recognition to “play Def Leppard” – something Siri can do as well. But when Siri is asked to “play Paul Oakenfold,” it took three tries to get it right. And that is at my desk in a quiet office with little background noise.
The point is Siri still has a long way to go to be useful and in my tests on some days it works much better than others. Sure this could be related to background noise but the technology is far too immature to rely on and users will find after a number of failures that it makes sense to go back to using a keyboard.
Probably some of the best articles on speech rec and how today’s Siri news ties into the past comes from Eric Jackson over at Forbes. He used to work at speech rec company VoiceGenie when they were a customer of TMC (where I am CEO) and on a related article he details where some of the major speech players have gone to over the years – hint, Google. He has two other pieces worth reading. One says Siri is intended to be a Google killer and the other agrees with my premise that Eric Schmidt doesn’t believe for a moment that Siri is a Google killer.
This last point is interesting because Schmidt seems to be positioning Siri as a major competitive threat so US and possibly EU regulators don’t label them as a monopoly. But within Google, many disagree that Siri is such an important development. Perhaps the best comment about Siri came from my wife who said why would Apple release Siri when it is obviously not ready? This was an observation she came to after a few minutes worth of listening to me repeat myself to Siri with ridiculous responses.
In fact, in my testing, Google does a better job of recognizing my voice requests and responding to them than Apple – which of course when coupled with the fact that Android is outselling iPhones means that Apple shouldn’t be considered much more of a threat than Bing.
But Apple has brought much more attention to the speech rec space where we have known for years that small vocabularies are infinitely easier to recognize than large ones. I am just left wondering if we are too early – will we end up disappointing legions of smartphone users with speech rec technology forcing us to wait for the next generation (of humans not devices) before they try the tech again.
So no, Siri is far from a Google killer – if anything it may kill the popularity of speech recognition for years to come. Let’s hope the tech evolves rapidly so we can continue the speech rec momentum.