commentary What do Apple's Siri and Microsoft's Kinect have in common besides processing voice commands? Right from the get-go they were both a top target for tinkerers.
Look no further than this past weekend's very neat hack of Siri, which managed to get the voice software--which is perhaps best known for being the first major voice recognition system that has a personality--to control a home thermostat.
Not content with searching for weather reports and dictating voice notes, programmer Peter Lamonica put together a software workaround to funnel Siri's voice commands through a separate server, then used those commands to interface with the digital thermostat. The end result let Siri both check and change the temperature settings, using Apple's servers to do the heavy lifting when it came to transcribing. Keep in mind this is just a little more than a month out from the release of the iPhone 4S.
Is this a watershed moment for the kinds of things mobile phones can do? No, but it was a big one for Siri. Users took some of its basics, and rethought the kinds of things they could do with it.
Something very similar happened with the Kinect last year. Unlike Apple, which packaged Siri as an exclusive software feature within the iPhone 4S, Microsoft sold the Kinect as a $150 add-on for its Xbox 360 platform. The move gave owners of a five-year old piece of hardware new ways to control their system, and games with voice and motion controls.
Seeing a cool new gadget to hack, tinkerers--and not just Xbox owners--took to the platform immediately, wanting to have their way with the hardware, and use it in places Microsoft was not yet offering, like on desktop PCs. A week after the Kinect's release, that's just what happened.
Third-party Kinect software was published as a response to a bounty program put on by Adafruit Industries, which offered anyone who came up with an open-source driver for the hardware a small pile of cash. Microsoft responded by saying such a move was off limits, and that it would "work closely with law enforcement and product safety groups to keep Kinect tamper-resistant."
Just a week and a half later, two company representatives effectively did an about-face on the subject during an interview with NPR, saying that those who were writing software for the Kinect would not be pursued, and that Microsoft was actually interested in, and paying attention to what people were creating. In February of this year, the company followed up on that interest, announcing plans to release a software development kit for Windows users, allowing third-party developers to create software titles that make use of the sensor.
That Kinect SDK was only for non-commercial users though, with plans to bring it to commercial developers sometime in early 2012. In a post yesterday, Microsoft said that it was making some adjustments to the hardware to make it work better with desktop users, including things like shortening the cord, and enabling a "near mode" to help it detect things close up.
In the case of Siri, things are a little more complex. There may be hardware attached to it, but Siri is software that's reliant on Apple's servers to work. This means there's nothing to take apart, and fewer tech loop tholes to squeeze through. In Kinect's case it was intercepting and deciphering data sent through a USB plug after writing the software that would talk to the accessory. That's not the case here.
So how did Lamonica get his solution working? He used a crack published last week by development group Applidium that makes use of the server name and address that Siri uses to communicate with Apple. That signal is bundled with a security certificate that gets those queries through the door by pretending to be an iPhone 4S. For now, that amounts to something like the rag tag group of rebels trying to get into the Death Star in the first Star Wars film. The code works, but Apple could be wise to the fact that it's not so genuine, and eventually shut it down.
That begs the question of what Apple plans to do next. Will the company lock down these tinkering efforts, or open Siri up, the same way it did with the iPhone itself?
You might remember that the company faced a similar crossroads shortly after the introduction of the iPhone, when there was no App Store. Instead, Apple urged users to create Web applications tailored for mobile phones. The results, while admirable by some, were limited, underwhelming, and hard to make money on.
Apple eventually relented, opening up its devices to developers to build on top of, a decision that turned out to be very profitable, with a recent Piper Jaffray projection pegging Apple's cumulative app sales just shy of $5 billion. It stands to reason there's the same type of business potential for app developers to create tie-ins to Siri, adding more utility to their applications, and by proxy Apple's software and hardware business.
But by the same token, it would also mean Apple relinquishing control. Take for instance something like taking a note through Siri. Right now that goes through Apple's first-party notes software that's a part of iOS. But what if you're an Evernote user? Would Apple be willing to let you default particular voice commands to certain apps, or complicate its system by making users have to pick from a list? If you've ever used Siri, you might know how much an extra step, like having to pick from multiple phone numbers or e-mail addresses when trying to call or send a message to a particular contact, can aggravate the experience.
In Microsoft's case, there was little question of what to do. People were buying up the hardware in world record setting numbers, with many wanting to build things on it. Not to mention, the company already had a software platform (Windows), in place to let that happen on a much bigger scale than the Xbox alone. There was also the assumption that the Kinect sensor would continue to shrink, and eventually trickle down to become embedded in displays on computers, adding extra incentive to get people in the door and working on software.
With Siri that future may not be so easy to envision. It's not clear if it will make the jump to other iOS devices, or Mac OS X to go into Apple's computers. Frankly, it makes more sense in a phone right now, in part because doing certain things on a small device is hard, and Siri arguably makes them easier.
In the meantime, just ask Siri if she can be hacked. Right now her reply is "I don't know what you mean." Pretty soon, Apple may need to figure out a better answer.