When Google's Scott Huffman announced the release of an SDK for Google Assistant at the keynote kicking off this year's Google I/O, he urged developers to put the voice-powered assistant to unique uses on all sort of devices. "Speakers, toys, drink-mixing robots, whatever crazy device you can think of," the Google executive said.
Huffman wasn't kidding about the drink-mixing robots.
Wander around the grounds at Google I/O 2017, and you'd run into the Mocktails Mixer, a device that taps into Google Assistant's speech recognition and AI tools to whip you up some liquid refreshment.
The Mocktails Mixer is the brain child of Pittsburgh-based Deeplocal, which bills itself as an innovation study. It's certainly an innovative use of Google Assistant and one that didn't require a big hardware investment. The drink mixer is built off a Raspberry Pi and an Arduino microcontroller. You've got a power supply and some pumps, with Google Assistant supplying the rest of the magic.
All told, the systems costs a few hundred dollars to make, according to the Mocktails Mixer if you ask it nicely enough.
The result is a machine that can mix drinks on your vocal command. Say to the Google Assistant inside the machine "Let's talk to Mocktails Mixer," and the familiar Assistant gives way to a new AI-powered voice. Ask the Mocktails Mixer what's on the menu, and it will rattle off your choices. Pick a drink, and the AI leaps into action. You can even prod it to tell a joke, if you like one-liners about Java programmers.
It wasn't long before I was enjoying a Mang-I/O, which is a combination of mango, cucumber, honey syrup, fresh lime juice, and non-alcoholic ginger beer. While not the more adult beverage I was hoping for after the thirsty work of covering Google I/O, it was still plenty tasty. And more importantly, it illustrates that Google Assistant can do more than just give you the weather report or rattle off ballgame scores if you put it in the right developers' hands.