Voice as a software interface.

Humans like to talk. It’s the way we’ve been communicating as a species for a long time. So it’s not surprising that science fiction writers have always assumed that we should be talking to our computers.

But as consumers we’ve all been frustrated by the attempts at voice recognition technology. We’ve yelled our bank’s voice-prompt telephone system, we’ve been embarrassed by lewd voice-to-text messages we’ve accidentally sent, and we’ve been tempted to throw our Siri-enabled phone off a bridge.

It turns out that getting a computer to understand speech is a hard problem to solve. It involves mapping audio patterns to known words, understanding the subtleties of context required for natural language processing, then how that maps to actionable instructions for the computer. Each one of those is a hard problem to solve – aggregate them together and you’ve got a really hard problem.

But I predict that 2017 will be the year that voice as a software interface really starts to take hold. Apple opened-up Siri to developers in June this year, and now they started shipping MacOS Sierra with Siri built in. Meanwhile, Amazon has opened-up their powerful Amazon Voice Service (AVS), available on AWS to all developers as a voice recognition and natural language understanding engine that can be used on any device that has a microphone and a speaker.

With Siri and AVS now available to independent software developers, I believe we’ll see a huge proliferation of high-quality use cases for voice as a software interface. There’s nothing like open source frameworks and millions of really smart engineers when it comes to finally getting technology to solve problems in useful ways.

So remember, you heard it here first. Open the pod bay doors, Hal.