What iRobot Taught Us About Home Robots

Hint: Don't expect a Rosie anytime soon

What’s more likely to service your home in the not-so-far future: a do-everything housekeeper robot like Rosie from The Jetsons, or a swarm of specialized bots that are more like Mrs. Potts and her household of anthropomorphized appliances in Disney’s The Beauty and the Beast?

This is the question that iRobot’s Roomba posed back in the early 2000s. Seems like it’s still to be answered even as iRobot recently declared bankruptcy.

I first met a Roomba robot vacuum cleaner in January of 2003, just a few months after it had been introduced by iRobot. I interviewed iRobot’s CEO and co-founder, Colin Angle, and the company sent me one of those first Roombas to try out at my house.

From the start, I thought the most interesting thing about the Roomba wasn’t the device itself, but the vision it teed up for how robots will infiltrate our lives. TV and movies liked to imagine robots as mechanical humanoids: Rosie on The Jetsons, C-3PO in Star Wars, The Iron Giant, Ex Machina. Lots of companies and labs are chasing this concept today.

But a walking, talking, do-everything robot is a highly complex technological lift. So far, we’ve mostly gotten Russian robots that fall on their faces on stage and soccer robots that couldn’t beat a team of blotto drunk 100-year-old nursing home residents.

Maybe it’s misguided to try to get robots to be like a human, versus getting robots to just be robots and do repetitive boring crap that robots can do really well.

The founding iRobot team, all out of MIT’s Artificial Intelligence Lab, came to that kind of thinking the hard way. The company was founded in 1990, a dozen years before the Roomba. As Angle told me in 2003, iRobot first tried to build more complex robots, like one that could seek out and explode underwater mines, and an AI-driven pet ball that would do things like cry if left alone in a dark place. None of it caught on.

But, Angle said, when he’d casually chat with someone and say he runs a robotics company, they’d often half joke that he should make a bot that would clean their floors. And then iRobot got a contract from S.C. Johnson – maker of Pledge, Windex and other cleaning products – to build an industrial floor-cleaning robot for, like, warehouses and factories. Put all that together, and iRobot’s team got thinking it could make an inexpensive bot that would take over a single repetitive boring duty from humans: vacuuming floors.

Once the company started down that path, “We did lots of things wrong,” Angle told me. “We got the robotic part nailed early thinking the cleaning part would be easy, then we spent 18 months making it clean.” (It actually took them longer. That first Roomba that I got did OK with dust, but couldn’t suck up a dropped piece of Cap’n Crunch.)

When the Roomba first came out, no one – including me – had seen anything like it. I described it in my column as looking “like a cross between a bathroom scale and a Belgian waffle iron.” At first it was sold in gimmicky catalogs like Sharper Image and Brookstone. The Roomba soon became a pop culture sensation. Go on YouTube, and you can find a ton of home videos of people’s cats joy-riding on Roombas. More than 1 million Roombas were sold in its first two years on the market.

If I put on my category design hat, iRobot did an amazing job of creating and dominating an entirely new category of product. It set the rules for robot vacuums, so that when competitors appeared, they all looked and operated much like Roombas. The company was in position to change our minds about home robotics. We weren’t going to come home to Rosie cooking and cleaning like a human maid – we were going to open the door to a swarm of mini-bots all doing their special thing.

Angle told me then that Roomba was going to be the first of many of these appliance bots. “My consumer group has a mission to make housework something people do by choice. We’re focused on tasks that we can accomplish robotically as good or better than the average person.” Other experts had similar outlooks. Sebastian Thrun, director of Stanford’s AI lab, in 2004 told me: “There will be robots that pick up dishes from the table and put them in the dishwasher within five years.”

It’s 20 years later, and if anyone has a robot picking up dishes, they are constantly replacing a lot of broken dishes.

I kept in touch with Angle and iRobot for a few years. I was at the iRobot booth at the Consumer Electronics Show in January 2004 when Bill Gates came by, wanting to know all about the Roomba and iRobot’s future plans. In May 2005, Roomba unveiled the Scooba, a floor mopping bot.

Angle told me then: “We aim our products at the axis of ‘hate doing it’ and ‘have to do it often.’ Mopping is near the top of people’s ‘hate doing it’ list.” He reiterated iRobot’s goal of making a suite of robots that do housework.

The company went public in November 2005. The Roomba remained a hit. By 2013, iRobot had sold more than 10 million.

And then, iRobot’s innovation seemed to stall. It introduced a window-washing bot and a gutter-cleaning bot. But they didn’t catch on, and it looks like IRobot doesn’t sell them anymore. Most housework, apparently, is really hard to automate. (The two most impactful housework automation machines – the electric dishwasher and electric clothes washer – were invented about 125 years ago. Since then…just Roomba.)

In 2022, Amazon said it would buy iRobot. Around that time, cheap robot vacuums from China started cutting into Roomba sales. Then the Amazon deal collapsed thanks to regulatory scrutiny from the European Union and the U.S. Federal Trade Commission. iRobot cut nearly one-third of its staff and Angle resigned.

This month, iRobot filed for bankruptcy and will be taken private by Shenzhen Picea Robotics, a lender and key supplier. Angle told CNBC: “This is nothing short of a tragedy for consumers, the robotics industry and America’s innovation economy.”

What about Angle’s robotic vision? It doesn’t look like he’s given up on it. He’s started a new company still in stealth mode, called Familiar Machines & Magic. He wrote on LinkedIn: “After decades in this field, I remain incurably optimistic about the future of people and robotics.” He added: “We are working to build the next generation of intelligent, teaming and purpose-first robots.”

“Teaming and purpose-first robots” is the key phrase. I really don’t see a day in any near future when I’m sitting here writing and a wise-cracking robot maid bustles around cleaning and bringing me martinis. (I don’t actually drink martinis when I write, but I like to imagine I would if a robot maid brought me one.)

More likely, I’ll be writing while one little bot vacuums, another scrubs the bathroom, one more folds and puts away laundry, and all of them terrify my cat, should I get another one.

—

This is the first column I wrote about iRobot, in January 2003, as accessed on Newspapers.com.