Whether it be the U.S.S. Enterprise‘s ship computer, the Batcomputer, or even HAL, the future has always promised us smarter, more helpful—and, sure, at times more dangerous—computer intelligence. But despite the existence of virtual assistants such as Apple’s Siri and Microsoft’s Cortana, that future has always seemed just a little bit out of reach. Modern virtual assistants are no more capable of holding a truly intelligent conversation than brewing you up a cup of tea, Earl Grey, hot.
Part of what might be holding the intelligent assistant back could be its close ties to the mobile platform. Right now, you may have Siri or Cortana at your beck and call wherever you take your smartphone, but there are a few drawbacks that prevent the virtual assistant from reaching its true potential.
One is power—in both a metaphorical and literal sense. Despite the increasing sophistication of mobile devices, their processing power still pales in comparison with desktop machines. More power could, for example, let the assistant process speech onboard the machine, rather than sending the information over the network to a company’s servers—that could enable faster response time, making the assistant snappier and forestalling some of those inevitable network hangups. And, hopefully, it could also help it better understand and parse what you’re asking for, so you end up with fewer mistaken queries.
Mobile devices are also reliant on batteries. Which means they have to worry about power efficiency; they need to limit tasks, especially processor- or network-intensive ones, so that the phone doesn’t rapidly exhaust its power reserves. In Apple’s recent iOS 8 release, it introduced a feature called “Hey, Siri”, which lets you trigger the assistant by saying the aforementioned phrase. But because of (rightful) concerns about battery life, it currently only functions when the device is plugged in. Great if you need to use it in the car or while you’re lying in bed, but less useful if you’re just around the house.
So if growth and development in the virtual assistant might be in part stunted by its commitment to mobile, the obvious solution would seem to be bringing virtual assistants to a platform where such constraints don’t exist—such as the desktop computer.
A recently posted video shows that Microsoft is doing just that. It demos a rough implementation of the company’s virtual assistant, Cortana, in an early pre-release build of the forthcoming Windows 10. In the demo, Cortana’s shown working with calendars, looking up nearby locations, and calling contacts via Skype. (Given that it’s such an early build, however, the service is unsurprisingly buggy and prone to error, as evidenced when it says it’s retrieving directions to “This is not working.”)
Removing the limitations of the mobile platform is potentially huge for the virtual assistant, but it’s not sufficient to simply duplicate the experience those assistants already provide. The bigger question is how people are likely to use a voice-controlled intelligent assistant when it’s available on the desktop, rather than just in their phone.
To me, this opens up an entirely different scenario of a sort of “ambient computing.” For example, if you’re relying on an intelligent assistant in a known space, such as your home or office, you can interact in different ways than you would if you were relegated to using your phone. For example, the fact that your desktop or laptop is generally going to be plugged in means that you can take advantage of the always-listening features, like that “Hey, Siri” feature, to call on your assistant even if you’re not sitting at your computer. This appears to be what Amazon is aiming for with the Echo device it announced last month—though hopefully tying it to your computer will make it a little more personable and a little less creepy than a standalone device that’s sitting around and always listening.
A desktop-based assistant could also prompt more natural, conversational interactions. Right now, a call that comes in to my Google Voice number will ring a landline, my mobile, my iPad, and up to two Macs. My virtual assistant could, instead, act like a real assistant, and announce from a single set of speakers that I have an incoming call from, say, a colleague—along with a voice-based prompt to answer or decline the call. Likewise, rather than popping up a dialog box on the screen to remind me of an upcoming appointment, the assistant could be smart enough to detect that my computer is inactive and instead announce that reminder aloud.
I have to admit that I’ve long wanted the capabilities of the iPhone-based Siri on my iMac, and frankly, I’m a little surprised that it seems Microsoft will beat Apple to the punch. At the same time, though, I think what we’re seeing of Cortana here—with the full knowledge that what eventually ships might be different—is underwhelming. If all we get on the desktop is a warmed-over version of the same mobile assistant we already have on our phones, it’ll be a bit of a letdown. What, after all, is the real advantage to having the exact same technology, just tethered to a bigger machine?
There’s every possibility that what actually shows up when Windows 10 arrives next year will be more complex and ambitious than what we’ve seen to date. I’m holding out hope for an intelligent assistant that’s even smarter than what we have now…and if it can talk like Paul Bettany, that’s just that much better.