Why Amazon’s home robot sounds like a terrible idea

Wrong company. Wrong product. Robotics companies should be focusing on robots for businesses, not homes.

closeup of retro robot artificial intelligence

Amazon’s working on a robot.

Details about Amazon's home robot project leaked recently. What we learned is that the project, code-named Vesta, is in the "late-prototype stage."

The robot will be "powered" by Amazon's Alexa voice assistant and will feature cameras, a screen and a small compartment for carrying things. It might also come with a camera on the end of a retractable pole for helping users find stuff under the bed.

The company has more than 800 employees working on it.

The publication Insider has seen an internal Vesta document and has spoken to six anonymous sources directly involved in the program.

The robot may cost more than $1,000. Amazon is reportedly considering rolling it out, so to speak, in a limited, invitation-only beta.

The robot is being developed in Amazon's Silicon Valley R&D organization, called Lab126, which also developed the Kindle line, Amazon Echo and other successful Amazon consumer products, plus the Fire Phone.

The Fire Phone was a horrible product and a bad idea. And the company is haunted by its failure.

Amazon staff are reportedly worried about its prospects for success. Sources complained about "shifting strategies and delayed launch dates" and wondered whether Vesta will be a hit.

They’re right to be nervous. Amazon’s Vesta is very unlikely to be successful.

The trouble with mobile home robots

For centuries, futurists, writers, scientists and filmmakers have driven home the notion that, in the future, mankind would be served by people-like robots — machine servants which, among other things, move about in our homes to help us with our chores. Rosey, the robot, from the Jetsons cartoon, was the perfect embodiment of this idea. She wore a modified maid outfit, and rolled around dusting and vacuuming.

Amazon is discovering (the hard way) what’s wrong with the idea of mobile robots as we reflexively envision them.

The mobility itself is a big part of the problem. In order to navigate any given house, with floors of wood, linoleum, tile, or shag carpet upon which the robot may encounter rugs, laundry, toys, dust, pets, pet hair, wires and cabling, spilled liquids, furniture — anything, really. Any kind of robotic locomotion scheme that can reliably navigate all this is complicated, expensive and battery intensive.

Mobility requires navigational systems, presumably lidar, which uses cameras to map the room and the floor and the walls and the objects on the floor. To buy a robot that scans the interior of your house for navigation is to trust that the company won’t abuse the data harvested through that camera-based system, and the public does not and should not trust Amazon on this. Amazon is notoriously secretive about what it does with private consumer data, and its privacy policies are vague about it. Consumers should not trust Amazon to keep private data private.

Mobility in a home is an expensive, privacy-invasive problem, and the benefit of this mobility is unclear. For example, if you want a smart display all over the house, it would be cheaper and more efficient to buy 10 of them and place them everywhere. If these did not contain cameras, then the public would be more accepting of them.

Yes, a mobile robot can bring you things — like drinks. But that’s a novelty use, not a practical one. If the robot can’t open the refrigerator door and identify and grasp the requested beverage, then what you’re talking about is a human placing a drink on the robot’s tray, or whatever, and the robot rolling 12 feet to deliver it. Such an application would be entertaining — once. But it’s not practically useful.

Other uses for home robotics include robotic toys and appliances. Window-cleaning robots can reach where humans can’t. Build-it-yourself robot kits for kids can be fantastic learning tools. Robot vacuums and mops do their job, and cats love sitting on them, according to YouTube. Robots have all kinds of uses in the home. But a general purpose mobile robot like Vesta is not one of them.

A robot that’s basically an Amazon Echo on wheels makes zero sense for the home.

Robotic elements — intelligent machines that move for limited purposes make sense — robotic door openers, automatic electric car charging arms, robotic lawn mowers, coffee making machines, and other single use-machine also make sense.

Space robots, security robots, farm robots, hospital robots — these are good ideas. Amazon’s own factory robots are legendary. There is a million different applications for robots. But a home robot that rolls around to tell you the weather and find your car keys is in reality one of the worst applications for robots imaginable. But we’re working on it because the futurists told us all our lives that we’d have mobile robot servants in our homes.

Contrast Vesta with Amazon’s other mobile home robot — the upcoming Ring Always Home Cam, which is an indoor security drone that costs $249. By flying, all the complications of navigating floors is avoided. By focusing on responding to security events, including motion, the flying “robot,” sticks to a single, high-value task, and doesn’t try to emulate a sentient being. (To the best of my knowledge, Amazon still hasn’t promised to avoid using the cam’s video streams to catalogue items in the home for contextual advertising and amazon.com promotions. I cannot currently endorse the Ring Always Home Cam on privacy concerns.)

Why businesses need robots

While homes are not great places for mobile robots, offices are fantastic. And the applications are immeasurable. We need robots for telepresence, so the many remote workers can participate in meetings. We need robots for delivery of print jobs, mail, coffee and other stuff. We need robots for office security.

Offices have predictable floor surfaces. IoT sensors can be built into different spaces in the office to help with navigation. And offices are semi-public, not private spaces, and so using cameras for navigation and other uses is not an invasion of privacy on the same level as home robots.

An office is a great place for mobile robots. So are factory floors, delivery trucks. Warehouses, and elsewhere.

But the worst place for a general-purpose mobile robot is the home. Homes are too hard to navigate. Homes are private spaces. And there’s no good reason to have a robot roaming around the house, other than that SciFi writers and Hollywood have let us to believe that we’d have home robots “in the future.”

Amazon’s Vesta robot product just sounds like a terrible idea.