What you're watching looks like a shiny new hardware company. the tech behind it are almost two decades old.

Before the App Store existed, before the smartphone boom, filings were already being made on the machinery of machine sight, careful diagrams of how a computer should break an image into gradients and decide what it's looking at.

The research underneath came from Navneet Dalal, the founder in this video, whose human-detection work out of INRIA Grenoble became the backbone of computer vision, cited over 50,000 times.

when this was drawn, there was no consumer market for machines that see. no product waiting. nothing to sell. a vision so early it had to live in a patent office for 20 years while the world caught up.

The thing it was waiting for is Matic. 5 cameras and a full computer inside the robot. it maps your entire home in 3D in about 20 minutes, and the map sharpens with every clean.

it knows the difference between a charging cable, a sock, a spill, and your dog, and acts differently for each one: cleans the spill, avoids the cable, gives the dog room. no app anywhere in the experience, you point at a mess, say "clean this," and it goes to that exact spot.

it cleans in total darkness, runs at conversation volume so it never wakes a napping baby, and because the entire brain lives on an NVIDIA chip inside the robot, none of the visual data ever leaves your home without your permission.

Most founders build on technologies that are already ready for the market. Navneet Dalal has been working on the underlying computer vision ideas for nearly two decades, and Matic is where that research finally turns into a consumer product.

https://t.co/Dsi4biisWE