In a lengthy post published to his nearly 13-year-old pet project, Facebook, social network wunderkind Mark Zuckerberg has officially unveiled the ins and outs of his built-from-scratch home artificial intelligence system. Dubbed Jarvis — a nod to the AI system Tony Stark employs in Marvel’s Iron Man comics — Zuckerberg designed his personalized home automation network to make accomplishing such tasks as playing music or dimming the lights easy while also allowing the AI to personally learn his “tastes and patterns.” Though he’d personally given updates (and even run public tests) on the system in the past, his recent briefing serves as the project’s most in-depth explainer.
Built using Objective C, PHP, and Python, the Jarvis AI system is essentially an ubertalented version of Amazon’s Alexa or Google’s Home device. Capable of operating a number of smart products in Zuckerberg’s home — such as his Crestron lighting system, his Sonos and Spotify setup, and the Nest cam he uses to check in on his daughter — Zuckerberg acknowledged that one of the hardest parts of the entire process was figuring out a way for each smart home system to communicate effectively with each other.
“Before I could build any AI, I first needed to write code to connect these systems, which all speak different languages and protocols,” Zuckerberg wrote. “We use a Crestron system with our lights, thermostat and doors; a Sonos system with Spotify for music; a Samsung TV; a Nest cam for Max; and, of course, my work is connected to Facebook’s systems. I had to reverse engineer APIs for some of these to even get to the point where I could issue a command from my computer to turn the lights on or get a song to play.”
After overcoming this initial hurdle, Zuckerberg turned his attention to allowing Jarvis to heed and interpret communication as similarly as possible to that of a human. To accomplish this, the system had to have a method for receiving text messages while maintaining the ability to translate actual speech. Zuckerberg felt the text message and speech combination was an incredibly important decision, admitting he’s surprised at how much more he texts Jarvis instead of talking to it.
“One thing that surprised me about my communication with Jarvis is that when I have the choice of either speaking or texting, I text much more than I would have expected,” he continued. “This is for a number of reasons, but mostly it feels less disturbing to people around me. If I’m doing something that relates to them, like playing music for all of us, then speaking feels fine, but most of the time text feels more appropriate.”
What may be most impressive about Jarvis’ ability to receive texts is the fact Zuckerberg built a dedicated Messenger bot that allows him to chat with the system no matter where he’s actually located. Not only does this allow him to preemptively turn on his lights but it also provides a supremely effective way to keep up with what’s going on at his house. For instance, with the inclusion of an advanced facial- and speech-recognition feature, Jarvis keeps Zuckerberg in the loop should someone stroll up to his house — known or unknown. Furthermore, if the system recognizes the person, it automatically lets them in — after taking their photo and adding it to the Messenger conversation, of course.
Though the system isn’t perfect — during an exclusive Fast Company demo, it took Zuckerberg four tries to turn off his lights — he says he’s committed to staying focused on continuously improving Jarvis since he uses it every day. Be it the addition of more voice terminals or the creation of an Android app (currently, Zuckerberg uses an iPhone to control Jarvis), there’s still plenty to be accomplished. However, from the looks of the schematic Zuckerberg published, it appears as though his system is essentially perfect. How so? Because he’s programmed Jarvis to be compatible with a T-shirt cannon that fires a new gray tee to Zuckerberg when he asks for one. If you ask us, that’s peak artificial intelligence.