Several years ago Amazon quietly acquired some speech technology by buying Yap, a small company founded by Igor Jablakov and several other speech veterans from IBM and elsewhere. Yap had made some real noise in the Voicemail to Text arena, but Amazon quickly exited that category, clearly with other things in mind.
The first product heavily influenced by the Yap people might have been the Dash. Dash is a combination scanner and voice note taker, and sort of looks like a little bottle opener. It’s a purpose-built, one-handed gizmo with one job: making it easy to increment (Amazon Fresh) shopping lists, either by scanning a barcode of something you already have, or speaking the name of the item. The voice transcription presumably happens in the cloud, using who knows whose speech engine. It’s a cool, simple idea. There was a very similar idea on Kickstarter a while ago called Hiku which did not get funded.
My life is not so optimized that I need Dash; my wife and I manage just fine every Sunday with a pencil and a piece of torn up junk mail. I love how Amazon had the mom holding a kid and only one arm free in the video. Course, she’s using the scanner on things she hasn’t run out of, which is a bit odd..(Dash is the tool for those families who are so optimized they need to order a backup jar of pickles). I would have changed the video to taking scans of empties).
Next up is the Amazon Echo, a much more ambitious gadget released yesterday. It is a dedicated box/speaker with network-tied Siri that stays put in whatever room you choose to put it in. Like Siri, Cortana and Google, it will trap a particular word (“Alexa”…great name, secure trademark, or both?) in order to invoke commands. Watch the video for a pretty thorough demonstration.
At Jott back in 2008 we had pretty much the same idea (w/o the BT speaker), but we had neither the cloud-based, accurate speech recognition tech, the simple home networking tech, or the microphone array. You could do Amazon product searches with Jott, but otherwise we had bupkis.
A few observations:
- The snarky might say that Echo is a way of getting Amazon’s version of Siri out there in a way that doesn’t rely on Fire phone sales…
- When the Fire Phone was rumored to be coming out, a lot of people assumed that it would barge its way into the market by being heavily discounted to Prime customers. I don’t think that happened immediately at launch, but did once it was clear sales were struggling. It appears that Echo is going after Prime customers in a heavily discounted way immediately (makes sense, they’re also the heaviest shoppers.)
- Though I understand why these are likely tied to Amazon lists/services, I think Echo in particular could be a very cool tool that will get some real adoption, and I’d love to have it a bit more accessible to developers. In fact, if it were open, Echo might have a better chance of getting real traction first, and pulling Amazon services along second. For instance, “Hey Alexa, put an appointment with Lauren in Google Calendar for 4pm on Sunday.” Or, “Hey Alexa, tell everybody it’s dinner time.” Or “Hey Alexa, play my Spotify playlist ‘Abba and friends'”. “Hey Alexa, call the police.” Embrace and extend.
- I wonder what the folks at Jibo think about this? Echo is a poor man’s Jibo in some ways.
- Amazon have done some work on the microphone array and beam forming, and I’m glad they’re trumpeting it a bit. It’s difficult to isolate a voice in a place with a lot of background noise (TV room, kitchen with a fan, etc.). One thing I doubt they’ve solved yet (we’ll see) is cross-talk, and picking out a particular voice amongst several.
- Are they going to get to the point of voice recognition, able to serve up different information depending on the voice that’s asking? (“Hey Alexa, what’s my schedule look like today?”)
- Is this going to be cheap enough to eventually put in many different rooms? Will they be aware of each other? Can it be an intercom too?
- Will they make a version that is simply a speakerless puck (a la Hiku)..I could see putting it in many more places if it was smaller and less obtrusive.
- I saw immediately on the web that someone was worried about Amazon listening constantly to my home environment. It’s valid, but no more so that iPhone (when connected to power), etc.
Cool stuff. It’ll be interesting to see whether this gets traction, and whether Apple has plans with homekit etc for things like this.