As a family, we’ve been playing around with Google Home for the past few weeks and one thing I can say right off the bat: it fits right in. Within a short period of time Home has become our central music player, our question answerer and fact-verifier, and even our own personal game show host. I wasn’t sure we needed an always on, digital personal assistant sitting pretty (and it is pretty!) in our living room, but it turns out that we do.
Here are a few of the features I like:
- Music. At the launch event in October Google touted its speaker quality and emphasized that music will be one of the most used features of Home. It’s powered by Google search so that you can search either by name and artist, an album or a playlist (for Google Music.) It also works with other streaming providers such as Spotify but I’m not sure how well the voice search is integrated with other streaming services yet. Google Music, however, was great at understanding our requests most of the time. For example, when we asked to play “Independent Ladies by Beyonce,” Home replied: “Playing Destiny’s Child, Independent Women Part One.” The more specific we were with our requests, the better Home understood them. Home is also great at creating a playlist on-the-go based on the first song requested One minor quibble is that Home doesn’t own up to when a song isn’t in its library, instead playing some lesser known variation of either that song or artist or even nothing remotely related. For example, when I asked to play Beyonce’s Sorry (which isn’t available on Google Music) I got the lullaby version: cute, but not really what I wanted to hear.
- General knowledge via Google Assistant. This has been perhaps the most fun aspect of having Home located strategically between the kitchen, dining, and living rooms, where it can pick up questions from the farthest reaches of those rooms. Questions at our house have ranged from what US presidents have been shot, the capacity of World Series ballparks, the length of the National Mall, and when a TV special was on (Hairspray Live, if you must know.) Assistant had an answer for each. That said, there were questions that it couldn’t understand and said that “it’s still learning.” but overall it provided the correct answer almost all the time. I also liked that the Assistant named the source before providing each answer, though most answers began with “according to Wikipedia…” What Assistant doesn’t yet know is how to connect questions and to understand the context between them. When asked “how tall is Curry” it gave the correct answer, but when the follow up was “how old is he?” Assistant had no idea who we were asking about. That said, Google claims that its Knowledge Graph, “the easily accessible information that pop up under the search bar for certain queries, now encompasses 70 billion facts.” So hopefully it is going to find the right answer most, if not all, the time, depending on whether it understood the question, which brings me to…
- Understanding speech. Over the Thanksgiving weekend we had guests with varying accents and while Home occasionally misunderstood the sentence as a whole, overall its success rate was high. That said, we have learned that questions must be asked in a very particular format to ensure understanding and we (the humans) are training ourselves to ask questions correctly. I call these “magical phrases” – present sentences that users know Assistant will recognize. This reminds me of the time in ancient tech history when we all learned a different way to write the alphabet because it was what the device understood. I assume that just like improvements in touch-screen navigation over the years, thus we will see improvements in speech recognition, both words and phrases, in the coming years.
- Localized answers for topics such as weather and news so that all we need to ask is just “what is the weather” without a location to get the local forecast. Home can also answer traffic questions based on location.
- Personal answers. Here’s where Home is limited. Asking “what is my day like?” is a magic phrase (see item 3) that reads back items from a calendar. But whose calendar? Home only supports one account at the moment which is limiting for personalized responses. For us, the decision was made to add the account that also pays for Google Music but this limits the rest of the members of the household. Google Home is truly a family device, serving everyone in the home (and guests!) so it’s a shame to limit features based on a single-user use case.
- We haven’t yet connected smart devices to Google Home. It would be nice to turn on the lights with a simple voice command.
- Productivity. We enjoy using the shopping list feature on Home. Saying “add flour to my shopping list” adds it to a keep checklist. To overcome the fact that the list is owned by the single Home user, it was shared with other household shoppers. The nice thing about the keep list is that it can be checked at the store on a mobile device.
- External services: it’s possible to connect an Uber account to Home, but not yet possible to order directly from an ecommerce site.
Why did we adopt Google Home so fast? It turns out that efficiency wins: it’s much easier to use voice commands than to pick up a device, unlock it, bring up a browser or the google app, and type in the query. As Home improves word and phrase recognition, requiring less magic phrases and more contextual association, and add third-party services beyond Uber (such as food delivery, shopping and travel reservations) it will become the preferred way of interaction with these services. Brian Roemmele says web sites will have to learn how to interact with voice devices such as Home and Amazon’s Echo. He adds that “by the end of 2017 5% of consumer-facing websites have a voice first interface” and, even better, “by 2020, 30% of web browsing sessions [will be] done without a screen.” Even at today’s technical level for voice interaction demonstrated by Google Assistant, it’s clear that in private spaces such as the Home, voice becomes the interface of choice.