Yahoo Finance 
Facebook adds speech recognition to Oculus Rift, Gear VR
Facebook today announced the launch of a voice search feature for its Oculus Rift virtual reality (VR) headset and Samsung’s Gear VR headset.
It’s available now in beta for people who use English as their primary language, the Oculus team said in a blog post. “This feature lets you perform voice searches from Oculus Home to intuitively navigate games, apps, and experiences,” the Facebook subsidiary said.
I tested out the feature on an Oculus Rift just now, and it does work. Once you enable it, while you’re in the Home portion of the Oculus app, you can say the phrase “Hey Oculus” to get the app to listen for your next voice command. If Oculus doesn’t know how to respond to you, it will let you know that with a line of text below an orb floating directly in front of you. Look down and you’ll see a tablet with a few suggested actions. It notes that you can search for apps in the store, launch apps from your library, or step away from the speech recognition area by saying “cancel that.” You can even turn off Oculus’ ability to listen to the “Hey Oculus” wakeword. If you can access Oculus Rooms for the Gear VR, voice search is there, too.
But there’s a lot that Oculus’ voice search can’t do. It won’t let you look at your friends, your notifications, your profile, your avatar, app updates that you haven’t installed, or individual settings. And Oculus certainly isn’t up for answering random questions, like “Who is Mark Zuckerberg?” You’re better off asking Siri, Alexa, Cortana, or Google. And if you want to talk to Oculus but you have no internet connection, you’re out of luck.
I guess I shouldn’t be too disappointed about Oculus’ speech shortcomings, because the team does point out in today’s blog post that it will be adding more functionality over time.
But I can’t help it. As far as I can tell, this is the first time that Facebook itself as a company is unveiling a speech recognition capability. The company has been developing artificial intelligence technology for years, particularly since the formation of the Facebook Artificial Intelligence Research (FAIR) lab in 2013. I’ve been wondering how good Facebook’s underlying speech technology is. Amazon, Apple, Google, and Microsoft make a big deal about their speech-activated virtual assistants, while Facebook has been silent all this time. Now that changes.
It turns out that Facebook’s system does appear to be good at understanding what I’m saying. And that’s a good start. Triggering the right action based on voice input is the next step, and Facebook will be able to perfect that part over time.
If engagement is good and people like it, the speech technology might well end up becoming a part of Facebook’s more popular services, like WhatsApp, Messenger, Instagram, and even good old Facebook itself.
