Bragi's CEO wants to know how fast I type. Nikolaj Hviid, founder of the (very) smart earbud startup, figures out that if I'm at 60 words per minute, with an average word length I can input just four to five bits of data into a computer per second. Pretty slow. Especially when you compare it to the gigabits of data a second we get from our phones and computers.
That's where Bragi come in.
"The purpose of wearables is actually not delivering heart rate or steps," he says. "The purpose of wearables is to increase the data rate from the human to the machine."
Still with us? What Hviid is talking about is the Bragi Dash & Dash Pro's new Alexa integration and the 'ear computer's' ability to let the voice service know you're tapping your cheek to wake the assistant via its gesture control '4D menu'.
He sees Bragi's core role as not only picking up your voice commands and delivering Alexa's responses into your ears ‚Äď as it has been doing with Siri and Google Voice/Assistant for a while now ‚Äď but also supplying Alexa with more contextual information about well, you.
"Let's talk about voice as part of natural interfaces," he says. "You just nodded because you agreed. That's part of your expression towards me. If you take voice as a monolithic input, that is not going to work very well ‚Äď it has to be combined with the way you express yourself anyway. When you get a phone call, we can accept the phone call by nodding."
A library of expressions
His ambition is for Bragi, via its 27 sensors, to build an infrastructure or "library of expressions" personalised to both your cultural norms and to you as an individual. The aim is to go beyond detecting nods and shakes of the head to whether you're sitting, walking, cycling or running, when interacting with Alexa, "or that you are panicking, or that you're calm, or that you're falling over" so that both the voice assistant and Bragi's audio transparency and noise cancellation features can respond accordingly.
Amazon already has a team dedicated to making Alexa's personality more human and more useful, based on things like user feedback and unanswered questions. The potential uses for contextual cues from the individual users is huge.
"When we look at these types of devices [like Bragi Dash], they are sensor filled and can start delivering more insight, more context back to those agents," explains Hviid. "It is understanding the entire spectrum of how a human communicates through body and voice and emotion. For that we have built an embedded AI so we're capable of picking up all that stuff as a combination. And that's where we will make the most advances on our side in the near future."
The 4D menu, which is still a work in progress as per Hugh's Dash Pro review, could even facilitate multiple personalities from one voice assistant in future. So depending on whether you're listening to music, out for a run or looking at something in particular, you could essentially select a plumber AI, surf trainer or a doctor AI.
Added context could come from who you are speaking to, so a plumber AI wouldn't answer questions on Taylor Swift, for instance, and with maybe more than one voice at a time: "You might want to have both your doctor and your coach [on a workout], those are the two people you might want to have accompanying you and maybe the third one is actually a friend who can virtually talk with you."
Alexa on the Dash now
Hviid tells me that because of Bragi's control over its embedded system and OS, Alexa will work very smoothly on Bragi Dash earbuds, more so than its existing voice assistants. "You will get instant assistance," he promised, with a click of the fingers. It's all controlled from the Bragi companion app, aside from Skills which you can add in the Alexa app.
I get a quick demo of Alexa on the Dash Pro. There's no need for the "Hey Alexa" command which might not be comfortable in public. Instead I tap my earbud (you can also tap your cheek with the virtual menu) and the voice assistant registers the request on the app, but it's too noisy on the IFA showfloor to get a full demo. The Alexa rollout is set for October, for all Bragi Dash and Dash Pro owners, so not too long to wait until we test it out properly.
The partnership can be traced way back to Bragi's Kickstarter campaign in 2014 when Hviid began to talk to the Alexa team, before it was such a big deal. Bragi's CEO commends both Amazon's commitment to "voice first" (as opposed to Apple which has a considerable stake in keeping users swiping around touchscreen apps) as well as Amazon's lead in entertainment, e-commerce and third party services.
Then there's the new Cortana collaboration to take care of the productivity angle: "Alexa now collaborates with Cortana from Microsoft, which is huge for us because when they've done that ‚Äď you can start sending emails, receiving emails from Office 365."
As for the competition, Hviid contends that the Dash family is a "different beast" to non-smart wireless buds from Bang & Olufsen and even what Samsung is doing with its Gear IconX 2018 series. I ask him about Samsung's form factor, Bixby and fitness features. "In an Asian environment, you're usually very proud if someone copies you," he says. "We're not in Asia though."
Hearables have landed
- Pilot on course to break down language barriersWaverly Labs' CEO Andrew Ochoa on building the ambitious hearable
- Harman talks the future of hearables and AIWe speak to Harman's VP of wearables about a future with Samsung
- What hearables can learn from AirPodsThey‚Äôre far from perfect, but they can teach smarter earbuds an important lesson
In the near future, Hviid tells me Bragi will be working on adding Bragi Skills to Alexa "so you can get more access to the device", improving its super hearing ‚Äď audio transparency ‚Äď feature as well as taking on these larger challenges around context and natural voice interactions.
"You can use Alexa to turn on lights or open the door and I find that very inspiring but I know people very close to me who could have had a lot of use for that," says Hviid. "What is Alexa good for? What can it do? It's not visual or audio, it's how it co-exists in use cases. It's a shift in nature of how you've done things before. But when the shift comes, it comes quite fast."