Jarvis by Mark Zuckerberg: Released December 20, 2016

Jarvis is a simple AI, a digital butler that can be use to control Zuckerberg’s home. This is by simply talking to his phone through the application that he built, it can control everything including lights, appliances, music and security, temperature. It can learn his tastes and patterns, learn new words and concepts, and can entertain his daughter Max.

It uses several artificial intelligence techniques, including natural language processing, speech and face recognition, and reinforcement learning, written in Python, PHP, and Objective C.

One best thing about Jarvis is that it can recognize people at the gate, it lets them in automatically and tell Mark about it. If Mark doesn’t want to talk, he can text Jarvis through the Messenger bot and it will follow his command.

While coding, Mark encountered some predicaments like connecting and communicating with all the different systems in his home. He need to write a code to connect all his systems, which all speak different languages and protocols. Another thing is that most appliances are not yet connected to the internet and it is impossible to control some of these using internet-connected power switches that lets you turn the power on and off remotely.

He wrote a code so his computer could control his home which he can use as his foundation in controlling his home. He made this code so he can exchange a few words using text messages, and afterwards he added the ability to speak and translate his voice into text for Jarvis to read.

The keywords is somewhat important here, like “living room”, “lights”, and “on” to determine that he was telling Jarvis to turn the lights on in the living room. It must be clearly stated, like if it’s in “ family room” or “bedroom” or “living room” or “Max’s room”. For instance if Mark asked Jarvis to play a song, he should include where because it might end up playing music in Max’s room while she was taking a nap.

Music to some extent is appealing but more complex because there are too many artists, songs and albums to hold, unlike in lights that can only be turned on or off. One example of this was a request connected to Adele: “play someone like you”, “play someone like adele”, and “play some adele”. If you will take a look at it, these three sounds alike, but each one is a totally dissimilar category of request. “Play someone like you” is to play a specific song titled Someone Like You. “Play someone like adele,” is suggests an artist like Adele, and “Play some Adele” is to play music of Adele something like a playlist.

One difficult part of this project is the face recognition but very useful when your friends are at your door and your AI needs to find out who they are through the cameras that were installed.

Jarvis was programmed on a computer, but in order for it to be at its use, you have to exchange a few words with it anywhere you happen to be. Along with that, Mark can also use his phone. How? Mark created an App called Messenger Bot, working across both iOS and Android. In that way he can communicate with Jarvis by texting it and it will instantly be relayed to the Jarvis server and process it.

Texting is as important as voice and speech recognition. Jarvis could listen endlessly to what he says because of the dedicated Jarvis app that he built. This app allows him to situate his phone on a desk and just have Jarvis pay attention.

Morgan Freeman is a popular American actor with an authoritative voice and calm behavior. He is lending his iconic sound to Mark Zuckerberg and will record only limited set of commands.

For Mark, this project was fun and rewarding and he is looking forward to sharing what he learn and he is open to suggestions to make it more effective.

Imagine every home sitting on this planet to have artificial intelligence embedded onto its walls. Could this be our future? What outcome can these phenomena have?