Microsoft Doesn't Get It, Episode 65,536: TellMe versus Siri

As with all new technologies developed by or associated with Apple, competitors are stumbling over one another to dismiss Siri. Google has already played the "yawn" card, and now Microsoft's Craig Mundie has joined the fray during an interview with Forbes. Here are Mundie's remarks when asked about Siri, proving for the 65,536th time that Microsoft Doesn't Get It:

"TellMe facility's been in the Windows 7 Phone (sic) for more than a year! I just think that people are infatuated with Apple announcing it... it's good marketing. At least as a technological capability, you could argue that Microsoft has had a similar capability in Windows phones for more than a year. You could take these Windows phones and pick them up and say, 'Text Eric,' and it'll say, 'What do you want to say?' and it transcribes it... You can query anything through Bing by just saying the words. I mean, all that's already there. Completely functional, been there for a year."

Can you really just pick up one of these Windows phones and use it right out of the box in the way Mundie claims? Not if the results of a video comparing TellMe versus Siri are any indication. Using Siri and TellMe simultaneously, an Australian user tries to create a meeting, send a text, check the time in Perth, and play songs by a certain artist. Siri doesn't miss a beat, while TellMe fails every time, often with hilarious results.

Let's issue some apologetics for TellMe: maybe it's tripping over the Aussie accent. Maybe it needs to "learn" the speaker's voice over time to deliver better results. Maybe the handset's microphone wasn't picking up his voice well enough to decipher his commands accurately. But the side-by-side comparison doesn't lie; Siri kicked TellMe's butt up and down the block.

The comparison video even handicapped Siri by using the sort of stilted, standardized voice commands that Android and Windows Phone 7 have indeed "already had for over a year." But the thing that Siri's detractors must keep in mind is that Siri is not just speech recognition. In fact, if you're one of Apple's competitors, and you're going to try talking about Siri intelligently, I'd say the first step is writing "Siri is not just speech recognition" on a chalkboard 100 times.

Certainly Android and Windows Phone 7 have had basic speech recognition functions before Siri's debut, but can you have a conversation with those implementations, or do you have to stick to a strict set of commands that the OS can interpret? You're certainly going to have a hard time convincing me that I can have this sort of interaction with TellMe on Windows Phone 7:

"I'm in Chicago. I want pizza." (I'm actually over 13,000 kilometers from Chicago, but never mind that.)
"I found a number of pizza restaurants in Chicago, IL, US."
"How about New York City?"
"I found a number of pizza restaurants in New York City, NY, US."
"Maybe hamburgers in Miami instead."
"I found a number of burger restaurants in Miami, FL, US."
"How windy is it?"
"The wind speed is currently 26 kph (ENE) in Miami, FL, US."
"No, how windy is it here?"
"Wind speed is currently 21 kph."
"Is it shorts weather?"
"Here's the forecast for today through this Thursday."

What makes Siri a revolutionary voice-controlled interface is the fact that you don't have to stick to the standard "Search for pizza restaurants in Chicago" or "What is the current wind speed in Palmerston North, New Zealand" language that other interfaces require. Even more than that, Siri understands conversational context in a way that competing voice interfaces don't.

Fine, Android and Windows Phone 7 beat Siri to market by being able to ask your phone questions, but that's like the difference between issuing commands to your dog versus talking to a three-year-old human being. My greyhound is pretty much a canine genius, but once you stray from the hundred or so commands and concepts I've programmed into her doggy brain, she's every bit as lost as TellMe would be if I asked it "Is it shorts weather?"

Siri is far closer to being able to understand natural speech the way a human would; it's not quite there yet, but it's often amazing how close it gets. On my first day with the iPhone 4S, I set a half-hour timer and later asked Siri "How's that timer doing?" fully expecting it to choke. Instead, it showed me the timer immediately. I was startled that it worked at all, and worked well, and moments like this prove that Siri isn't the also-ran that people like Microsoft's Mundie are trying to claim it is.

Here's the funniest part: just like Multi-Touch, I suspect Apple's competitors are only being dismissive about Siri for as long as it takes for them to copy it... badly. Meanwhile, Siri and I will be over here waiting for you to catch up.

"Play some Led Zeppelin."
"Here's your Led Zeppelin."

And of all things, what plays? "Communication Breakdown." Siri, you so crazy.