The camera is the new keyboard
Snap filed to go public and calls itself a camera company. The phrase sounds like marketing, but it describes a real shift in how people put things into computers.
Snap filed for its IPO last week, and the first line of the business description in the S-1 is “Snap Inc. is a camera company.” They’ve been saying this since they renamed themselves last fall and launched Spectacles, the sunglasses with a camera. Most coverage treats it as positioning, a way to avoid being compared to Facebook.
I think it’s more literal than that, and it’s worth taking seriously.
For most of the history of computing, the way you put information into a computer was typing. The web was built on text boxes. Search was typing. Social networks started as status updates. Snapchat’s app opens straight into the camera, not a feed, and the default way to say something is to take a picture or a video of your face, your food, or whatever’s in front of you. For their users, the camera is the input device, and text is something you add on top.
That makes the camera a much more important piece of software than it used to be. It now needs to understand what it’s pointing at. Snapchat’s Lenses are a good example, since I spent years in exactly this area. They find your face, track dozens of landmarks at video frame rate, fit a rough 3D model, and deform or decorate it in real time, on a mid-range phone. When I was doing this in 2013, it was a research problem. Snap bought Looksery in 2015 and made it a daily habit for over a hundred million people.
Once the camera is the keyboard, a lot follows. Search starts from images: point at a plant, a product or a sign. Messages become visual. Ads become things you wear on your face. And the companies with the best computer vision running on phones get a real advantage, because every interaction starts with the camera understanding the scene.
Whether Snap itself wins is a different question. Instagram copied Stories last August and seems to be doing very well with it, and Facebook can copy features faster than Snap can invent them. But I’m fairly sure the idea outlasts the company. In five years the camera will be the main way a lot of people start interacting with their phone, and it will understand much more of what it sees than it does now.