Amazon Echo review: listen up

Alexa, you're the best. Alexa, are you there? ALEXA?!

167

“Alexa, what’s the weather today?”

“ALEXA! What’s the weather today?”

“ALEXA WHAT IS THE WEATHER TODAY GODDAMMIT.”

“Right now, in New York it’s 25 degrees with clear skies and sun. Today’s forecast has mostly sunny weather, with a high of 29 and a low of 19.”

Meet Alexa. She’s sitting over there, on the shelf next to my TV. She’s very young, just a couple months old. People who meet her obsess over her, cooing and asking her questions. She’s remarkably smart for such a young thing, but like every infant she perpetually toes the line between totally adorable and incredibly annoying. And at least once, she’s woken me up in the middle of the night.

She’s my Amazon Echo, a black cylinder about the size of a tennis-ball can that has become my apartment companion for the last few weeks. The Echo is basically Siri or Google Now in a box: an always-on, always-listening way to provide quick answers and information, to add things to my task and shopping lists. It can play music and podcasts, report the time or the weather, set timers and alarms, even calculate how many tablespoons are in a cup. “One cup equals 16 tablespoons,” she monotones, in that slightly robotic but very friendly voice.

The Echo is one of the most compelling cases I’ve ever seen for the power of voice control, of talking to our gadgets the way we talk to each other. It’s also a powerful and infuriating reminder of its limitations, of how long the road to our robotic future really is. Alexa’s going to turn out just fine, I think, but she’s got some growing up to do.

She’s funny, though. That’s a good start.

Amazon Echo and books

I’m standing in my kitchen, making soup for dinner. I need four cups of broth, and all I have is a 32-ounce package of College Inn chicken broth; I have no idea if that’s enough, too much, or not enough. "Alexa, how many ounces are in four cups?" "Four cups is thirty-two fluid ounces," comes the voice from the other room. Well! Good for you, College Inn, doing me a measuring solid. I pour the whole thing in.

Oh, now I need more chicken broth for tomorrow, because it’s winter and what else am I going to eat? "Alexa, add chicken broth to my shopping list." "I’ve added chicken broth to your shopping list." Boom.

I can’t count how many wonderful moments like this I’ve had with the Echo. It’s just so… simple. You say a command or ask a question, making sure to always begin by saying "Alexa." (She’s named for the library of Alexandria, which stored the knowledge of the ancient world. I know this because I asked her.) Your only other option for the wake word is "Amazon," if you like, but I prefer Alexa – it’s just more natural to say a person’s name, and it won’t often register false positives anyway. The one exception: whenever I played FIFA, and the announcers called the name of Chilean striker Alexis Sanchez, Alexa would burst to life ready to answer a question. I’ve instigated a lot of accidental searches for the phrase "dribbles down the wing."

Alexa I need to buy some weed. "Weed added to your shopping list."

Any question that has a simple, spoken answer, like the time, the weather, or the year Kobe Bryant was born, Alexa just says it aloud. If she can’t find the answer, or doesn’t understand the question, she creates a card in the companion Echo app for Android, iOS, and all the Kindle products. Your query goes into that card, and with one tap becomes a Bing search. (Yes, Bing and only Bing.) The Echo app houses a running history of all your questions, your to-do and shopping lists, and a lot of helpful tips for how to talk to Alexa. When the Echo is working properly, you’ll only ever need the app to access the lists Alexa helps you create; every time I’m sent to the app to complete a search it feels a lot like failure.

She can understand complicated questions, and she answers quickly. She even understands the things I didn’t expect: "Alexa, turn the volume down." "Alexa, play something else." "Alexa, how do you spell ‘embarrassing?'" She's always fast and usually quite helpful, and sometimes Alexa feels genuinely like magic.

But only sometimes. She doesn’t work every time, she doesn’t respond the way she’s supposed to half the time, and it doesn’t take long before you stop totally relying on her. And then she sits silent.

Amazon Echo

If you know what you want to listen to, the Echo is usually helpful. Saying "Play John Coltrane," or "Play 'Turn Down For What'" is perhaps the fastest way I've ever found to get to either one. It's the most magical thing about Alexa. When we were filming the video for this review, for instance, a woman walked into the room and asked about it. We explained how the Echo worked, and she dismissed the whole thing as weird — then, out of nowhere, she leaned over and said, "Play Beyoncé." When Beyoncé started playing, she lit up and immediately asked how she could buy one.

On the other hand, simply saying "Alexa, play some music," is the most dangerous thing you do with the Amazon Echo. Alexa will, indeed, play some music, but there is absolutely no way of guessing what it will be. It’s always a playlist of songs available in Prime Music, either from your library or curated from other users’ public lists, and just when you think you’re safe with "Road Trip / Long Weekend" the next pick is "Songs to Annoy Your Parents." "Alexa, stop!" I say, issuing the universal command for the Echo to cease and desist. But Alexa doesn’t hear me — she’s busy cranking the volume on a song that appears to consist only of animals being dragged through a woodchipper.

"ALEXA, STOP!" "ALEXA VOLUME 4!" "ALEXA WHY MUST YOU TORTURE ME!"

Do not — I repeat, do not – buy the Amazon Echo as a Bluetooth speaker. It sounds good enough to make a robot-person’s voice audible, but music comes through shallow, tinny, and compressed. You can get a much better speaker for the price. Setting the Echo up is more complicated than an average speaker, too, since you have to connect to its Wi-Fi network instead of just over Bluetooth. The Wi-Fi bit is necessary to let the Echo connect to the internet and do everything that makes it impressive, but it does take a few minutes each time you move to a new Wi-Fi network.

Alexa, guess what? "Something exciting, I hope."

When you’re listening to music over Bluetooth, the Echo also offers no useful additional functionality. When you use Alexa, it does, at least in theory change the volume, go to the next track, and search for the song you want, all with your voice. But as soon as music starts playing at any audible volume, Alexa completely stops being able to hear you. And by the time you’ve walked all the way over to the Echo, pressed the button on top, said "volume three" five times, and waited to see if anything was happening before finally giving up and spinning the wheel, the fun is gone. Alexa: great for finding music, not so great for listening to it.

Amazon Echo

Other than a blue-green light that flashes around the top of the canister, Alexa offers no real feedback while she works. So when a command fails to register, it just… fails. Sometimes she doesn’t hear me; sometimes she doesn’t know quite what I’m saying. In either case, she ignores me and just keeps on playing the 30-second preview of "Uptown Funk." (This, by the way, is the one place where the Echo can actually buy things for you: just say "Buy that song," and it’ll get added to your Prime library.)

The hardest thing about using the Echo is that I can’t get a firm grip on its limitations. If I knew not to ask it certain questions, or to always phrase questions certain ways, that would be fine. But I can’t explain why Alexa knows Andrew Jackson is the proper response to "Who was the seventh president of the United States?" but can’t tell me Thomas Jefferson was the third. I can stand right next to it, and it hears me fine… until it doesn’t. I’ve learned how to speak to Siri, how to bend Google Now’s robot to my will. But Amazon is building its own cloud service and software from scratch, competing with Google and Apple rather than borrowing from them. That’s hard, slow work, and right now it means Alexa is a unique breed of voice assistant: sometimes smarter, sometimes dumber, always more confusing.

Even the Echo’s hardware is a hedge against these shortcomings. The Echo ships with a remote much like the one that comes with the Fire TV, with basic playback controls and a microphone so it can actually hear you. There’s also a button on top of the device itself that activates the mic and a ring for controlling volume, and you can even ask some questions in the companion app. Only problem is, that’s all entirely antithetical to the point and value of this device. I don’t need new ways to type questions, or new things to keep in my pocket. I need the Echo to just work, so I’m not standing in my kitchen like an insane person shouting questions to nobody.

It’s also frustrating that the Echo has to be plugged in all the time, though that particular limitation is probably the right one. You never have to wonder if it’s on or charged; it’s always there, always ready. The Echo’s most successful when it’s not something you have to think about, when you can just speak to it freely and trust it.

That’s actually part of the good news about Alexa: its biggest problems aren’t with hardware. The speaker at least does the job, and the mic is good enough to record clear audio from 30 or 40 feet away. How do I know? Stored with every query on the History screen of the Echo app is a brief audio clip of what you said. When Alexa hears "Where’s the toy section?" I can flip back and see that I actually asked, loudly and clearly, "Where is Victoria Falls?" It’s deeply unnerving to know that not only is my voice being recorded and processed, it’s being stored. It’s also maybe encouraging, because as Amazon works through this data (each entry also asks if the Echo heard you correctly, which, no it didn’t) the Echo will — theoretically, anyway — get better quickly.

Amazon Echo

In addition to becoming more accurate, the Echo needs to become more powerful, and it needs other apps to support it. The Echo app’s to-do list is too simple now, but what if I could add things to the app I already use? What if the Flash Briefing, the collection of news it provides every morning, could pull from my podcast app and my actual favorite news sources? What if it supported Spotify and Pandora, and not just iHeartRadio and TuneIn? In Amazon’s vision (and its ads), the Echo is a persistent, omnipresent companion in your life. She’s there to wake you up in the morning, to guide you through your day. That means Alexa needs to fit into your existing workflows and routines, not force you to create new ones around her existing feature set.

Maybe it's a good thing that Amazon is so slowly rolling out the Echo to customers — you can only get it by invitation, and if you do so you should know what you're getting into. Right now, once the honeymoon ends, I suspect most people will stop using the Echo for anything other than occasional music and podcast listening. After a few days of trying to come up with things to ask Alexa just for the fun of the thing, my Echo became mostly a slightly faster way to set a timer or do quick conversions while I cook. It’s great for those things, but those are small things.

Yet this is the future, I’m sure of it. Several times a day, the Echo blows me away with how well it converses, and how natural it feels to interact with a machine this way. I just hope Amazon can marshal the considerable time, support, and developer firepower to one day turn this odd black cylinder into the smartest thing in my apartment.

Until then: Alexa, tell me a joke. Alexa, TELL ME A JOKE. ALEX-ah, forget it.

Photography by Sean O'Kane. Video shot on location at WeWork.

The best of Verge Video