Alexa and Siri sitting in a tree

I’ve had an interest in voice user interfaces for some time but living in Taiwan kept my experiences limited to Siri, due to I suppose limited support for local languages and I’m sure a host of legal/licensing issues. I’ve in the past been underwhelmed with Siri’s natural language processing, perhaps do to my dreams of interactions more akin to science fiction, like in the film Her.

Siri does often surprise, as she did by suggesting on my iPhone lock screen that I call my wife on her birthday.

When Amazon had the Echo Dot’s for sale at half price, I bought four, which is excessive, but I justified the expense since they can function as a decent bluetooth speaker for each of the kids room. The sound that comes out of them is about what you would expect from a speaker of this size, and better balanced than most of the bluetooth speakers we have purchased in the past. Pairing two together makes a for a pleasant enough bedroom listening experience. I haven’t spent a great deal of time with the devices but here are some insights thus far:

  • Siri is interesting in that after a short training period it will only respond to your voice, which is great when you consider the hilarity that would ensue with the potential number of iPhones in one room. Alexa out of the box has no such limitations. So unless you speak quietly, one person activates all the echo dots at similar times. You can set-up a voice profile but it’s buried within the iPhone app. and not readily that you can do so.
  • It’s much more enjoyable interacting with a device with decent sound output but I don’t find Alexa’s responses to be as smooth as Siri’s.
  • Apple Music skill is US only. A big disappointment to my kids.
  • My kids love music, but Alexa won’t play anything on their accounts. Alexa won’t play podcasts either. I have prime so I get a subset of music, but the catalog and experience pales in comparison to Apple Music.
  • I paired 2 Echo Dots together for stereo separation but this only works for music and not for anything else. When playing a sleep sounds skill it plays through 1 speaker only. This seems like an incredibly glaring oversight on their part which I hope they fix in the near future.
  • When you pair 2 speakers, the volume increases when you play anything other than music. It’s like the Echo is compensating for the loss of the other.
  • Alexa skills have discoverability problems, and
  • Skills are only interesting in theory as it requires the user to remember a long set of commands. UI should not add to the users cognitive load – I feel like I am back to having to write down common commands for Unix because I don’t use the system enough to commit them to memory. While I understand you are supposed to be able to ask Alexa what skills are available, I’ve never been able to get it to work. There is no equivalent to typing “help” in Alexa.
  • Skills seem very similar to traditional interactive voice-response systems, very unnatural.
  • I’ve yet been able to get it to play a podcast. This is likely due to not linking one of the music sources that contains podcasts, but Alexa describes it as an unknown error.
  • All the common requests (weather, etc.) seem to be handled with aplomb.
  • You need to think very clearly before you ask Alexa (or Siri) to perform a request. I suspect most don’t. There is no mid sentence error correction.
  • There is no way to get Alexa to repeat a response – no please repeat that again. Often times a response to a request is too long and it’s difficult to fit all the information into working memory.
  • Most surprising to me was when I asked a more advanced query, “Alexa, I would like to buy toothpaste”, it responds with an error message stating that I would have to change the primary language of my account. I thought Amazon would have this part nailed down tight. At least I can be assured that my kids won’t be buying their favourite treats via the Echo in my bedroom.
  • The most fun usage (for me) thus far is the Rooster skill. I walked into my sons bedroom early this AM and asked Alexa to play Rooster and it proceeded to loudly play a variety of Rooster crowing sounds. He hates me now.

Other than hardware that many people can afford, I as of yet see no major advances with Alexa over my experiences with Siri. It’s a pity that Apple hasn’t developed a similar smart speaker “for the rest of us”, $450 for each HomePod is not money well spent I think. Despite their limitations, I do look forward to digging deeper into skills and routines – in our effort to keep kids eyes away from screens, I hope to develop my own.