Ghosts in our machines
At itch we like to learn about new types of tech and new interfaces as they emerge, and to evaluate the context where they may be employed (if at all).
This past year has seen a lot of interest in speaking with computers.
This is both in line with the larger trend of appification, where services exist as compartmentalised, on-demand resources.
As users struggle to find behind which colourful door lies the answer to what they want to achieve. The number of these silos of activities and contexts that we can keep track of is quite small, so there is plenty of opportunity in opening up the possibility space as well as guiding the user from the very first interaction.
There is also an element of users wanting to escape the unnavigable jungle or modes, channels and conflicting interface conventions.
Does the answer lie in voice or text as a unifying interaction paradigm?
We’re eager to tinker with this, so where do we start?
Amazon’s Alexa in its various physical incarnations has now been available for two years, Google have announced their version dubbed Home and Apple’s Siri lives across the roundrects in your life; in your Mac, iPhone, iPad and Apple TV.
We wanted to see how spoken commands would work in our office environment. Evaluating the available options left us with Amazon Echo that arrived a couple months ago. It was fairly easy to set up, despite the dedicated app not being available outside the U.S.
So far, the applications have been limited to controlling office music playback on Spotify. However, more complex queries don’t work, and some artists’ names are difficult to pronounce in a way that is recognised.
There are a lot of third-party and community-developed “skills” available for Alexa, but when browsing the available options few are services that we already use. The number of options is increasing steadily, with more announcements made every month
We haven’t gotten into hacking the device just yet, but the first steps would probably be to enable local weather and we were planning to set up the skill to control a Sonos system. Sonos recently announced that official support will be coming in early 2017 so we’ll likely look for something else when we find some time to dig into it.
Little ghosts
It’s still early days for this type of interface, but Facebook’s and Apple’s progress means that there can be accelerated experimentation and learning, if still a bit limited, on the majority platforms of the West. We’re still quite a way from the vision of a virtual assistant that can handle our every request.
App makers on iOS so far seem to regard it an opportunity to drive more traffic away from the current context and into their app, which was perhaps not quite the point. Other early examples include leading you through a “phone tree” type menu which is sure to simplify certain interactions but that is way too simple for many.
The promise of an artificial intelligence that can join a conversation, lurk in the background and provide support as well as “do the boring/messy stuff”, is not quite available to most people.
One quibble we have with Alexa is that her understanding of “Skånglish” seems to be very limited. Skånglish is a regional variant of English spoken only in the south of Sweden that happens to be the native tongue of one of us itchers. By the way, this colleague has also been known to trigger Siri completely out of the blue in other people’s phones (even without “Hey Siri” being turned on). So we don’t rule out some kind of virtual agent whispering. Or something like an ongoing Telephone Game with technology.
