Uncubed

Upgrade |

SoundHound’s revolutionary new tech was 16 years in the making

Image courtesy SoundHound

Listen Up

While the experience of Siri and Cortana is fun, a user is inevitably left with the sense that this a technology that hasn’t quite found its bearings just yet.

Hound, a remarkable new voice-recognition platform from the creators of SoundHound, feels like a major leap forward.

Among the questions CEO and founder Keyvan Mohajer successfully asked Hound during our conversation: What is the time in Tokyo when it is 2:30 pm in San Francisco? When is the sun going to rise three days before Christmas of 2021 in Tokyo, Japan? How many days are there between the day after tomorrow and four days before the second Thursday of November in 2022?

(In case you were wondering – 6:30 am the next day; 6:47 am; 2,374 days.)

Of course, the app also has its practical side. Demand, “Show me hotels in San Francisco for tomorrow staying for two nights, that cost between $200 and $300 per night, and are pet friendly and have a gym and a pool,” and Hound will return ten results in the blink of an eye.

It’s then possible to sort and refine those results, without rephrasing your original question. As you might expect from technology this magical, Hound and Houndify (the platform which allows anyone to integrate a voice-enabled interface to apps, cars, refrigerators, et cetera) have been in the works for the last 16 years.

“When I started my phD at Stanford in the year 2000,” Mohajer told us, “I made several observations, the first that there will be a time when we talk to the things around us and they talk back to us, this era of voice-enabled conversational interface. Now that’s obvious today, but it wasn’t obvious then. My second observation was that it is going to happen in my lifetime.

“Four years later, I went to investors and told them in 10 to 20 years we’re going to talk to everything, and I want to be the one doing it,” Mohajer said. “And the feedback was, ‘That’s a long time for VCs – what can you do in three years?’

“I went back to my dorm room, and we thought about what business and technology we could build in three years on our own. We realized we could do music recognition… After about a year working out of my dorm room, you could hum a song and we could tell you what it is, and nobody else could do that. Even today after 10 years, nobody else can do that. And that was fundable.”

The SoundHound app launched in 2009 on iPhone – today, the tech has reached 300 million global users on both iOS and Android.

“But from the moment we launched I told investors, I want to build our own speech engine, because one day when the world is ready, we’ll be ready, and that will be huge for us,” Mohajer said. “It took us about nine years to build the technology behind Hound. Now it’s a very hot space, everyone talks about conversational interfaces, but you won’t see a lot of new startups enter the space.”

That speech engine became Hound, which launched in private beta last year, before officially launching in March. According to Mohajer thousands of companies are now using the Houndify platform; Samsung announced earlier this month that they are integrating Houndify with their IoT platform.

The tech represents a significant leap beyond the abilities of Siri or Cortana, allowing users to engage in incredibly complex conversations and receive highly specific answers.

“We believe we’re the only company that has been able to combine speech recognition and natural language understanding, so when you speak to Hound, or anything that has been Houndified, we have the meaning of what you’re saying, which makes it faster and more accurate. It’s a zero-latency experience.”

You can download Hound and SoundHound for iOS and Android.

Share