Roomba: Hitting the Wall | Fast, Cheap, and Out of Control

Business Wars

The episode chronicles the rise of iRobot, beginning with Rodney Brooks' radical philosophy that robots should mimic simple insect behaviors

Key takeaways

• Rodney Brooks' foundational robotics theory was inspired by observing how mosquitoes navigate environments using simple rules.
• The early iRobot team, including Colin Angle and Helen Greint, struggled to find a consumer product, relying instead on military and research grants.

Main topics

• Robotics innovation
• Business pivot strategies

Notable quotes

Robots dumber than mosquitoes.

Conclusion

The episode highlights how business survival often depends on leveragi

Transcript preview

Speaker 2 (0:00) Hey, Business Wars fans, have you heard? Audible subscribers can listen to new episodes of Business Wars early and ad-free, right now. Join Audible in the Audible app or by subscribing on Apple Podcasts. A listener note, Business Wars uses dramatizations based on true events. Some elements, including dialogue, are recreated, but everything is grounded in research. Speaker 2 (0:33) It's December 1984 in Thailand. 30-year-old Rodney Brooks, an Australian-born robotics professor, wanders aimlessly through the backyard of his in-law's house while his wife and her Thai parents are out running errands. Brooks pauses beneath a mango tree and looks up into the branches. He studies the way the sunlight dapples through the trees. He listens to the birds. Then something catches his eye. Mosquitoes. Tiny ones. They zip through the humid air. Brooks swats at one. He misses. The insect dodges and lands on a leaf next to a bead of water. Another mosquito buzzes. Brooks swats again. Another miss. Now the professor is interested. He keeps watching as one mosquito after another flies around, hovering over the garden, darting between plants and trees. and he's struck by a thought. Brooks teaches and builds robots at an MIT lab in Cambridge, Massachusetts. His colleagues have been loading their robots with complex artificial intelligence, with the goal of making them think and react like people. But lately, Brooks has questioned that approach. He knows people think and react through a network of billions of neurons. Robots can't match that level of processing power. Now, standing in the backyard, Brooks realizes his robots can't even match the intelligence of a bug. These mosquitoes are adapting on the fly, literally, to whatever is around them. But robots have to crunch reams of data to respond to even the slightest change. Brooks runs inside the house and finds a pen and paper. He starts scribbling a note. Robots dumber than mosquitoes. Speaker 2 (2:33) That's the problem. We need to simplify them. We just need to give the robot basic rules. Brooks keeps jotting down notes. Rule number one, move your legs. Rule number two, don't run into stuff. Rule number three. explore. Brooks figures these kinds of basic rules are all his machines really need. They don't need detailed maps of the world or computer simulations of the human mind. They need to do simple things in simple ways, just like a bug. For Brooks, this approach breaks with the traditional ways he's worked with robots so far in his career. It's also a break with the mainstream. Back in Cambridge, Brooks shares his idea with some of his colleagues at MIT. He tells them that robots should think like insects. The idea lands with a thud. Brooks' colleagues are trying to build machines that can reason and then act. They're building mechanical people, or at least trying to. They tell Brooks it sounds like he wants to build toys. But Brooks isn't deterred. In fact... He's convinced the future of robotics isn't humanoid. It's buggy. Brooks doesn't know it yet, but his conviction will become the foundation for a lucrative robotics company called iRobot. They'll put robots in millions of homes. But they'll also discover that while building robots is hard, building a successful business is even harder. Speaker 1 (4:19) Hi, I'm Lindsey Graham, the host of American History Tellers. In our latest series, a police raid on a New York City gay bar in 1969 ignites a series of demonstrations lasting six days and marks a turning point in the fight for LGBTQ rights. Listen to American History Tellers on Audible or wherever you get your podcasts. Hi, I'm Lindsey Graham, the host of the podcast American Scandal. Our latest series tells the story of college counselor Rick Singer, who thought he'd found a way to rig the admissions process for his wealthy clients until an unrelated FBI investigation inadvertently exposed the largest college admission scandal in American history. Listen to American Scandal on Audible or wherever you get your podcasts. Speaker 2 (5:05) From Audible Originals, I'm David Brown, and this is Business Wars. Speaker 2 (5:36) Hey, do you have a robot in your home? Chances are, if you do, it's a Roomba, or a copycat Roomba. One of those low-slung, disc-shaped devices humming across your living room floor, bumping into chair legs, getting stuck under the couch, and quietly eating dust bunnies while you eat dinner. In fact, you might have even named your Roomba. Perhaps your cat is joyriding on it right now. That little household helper. started with Rodney Brooks' breakthrough shift in approaching robotics. In the mid-1980s, Brooks argued for something almost heretical in the field. Robots don't need to be that smart. They just need to follow basic commands. Move forward. Avoid obstacles. Go back home. Brooks and his company, iRobot, launched the Roomba in 2002. Since then, 50 million Roombas have been sold worldwide. It was the first affordable robot of its kind that actually lived in your home, not in a factory, not in a research lab, not in a science fiction movie. And it became so dominant that even if you own a competitor, you probably still call it a Roomba. But then the Roomba hit a wall. Cheaper, smarter rivals, many from China, started showing up. And at iRobot, the wheels started falling off. For iRobot, the question is simple. Can a company built on a radical, bug-brained philosophy survive when the competition starts to outsmart it? This is Episode 1. Fast, cheap, and out of control. Speaker 2 (7:25) It's summer 1988 in Cambridge, Massachusetts. MIT professor Rodney Brooks watches intently as one of his students carefully lowers a strange-looking machine onto the tiled floor of Brooks' robotics lab. The machine is about the size of a loaf of bread and looks like a giant bug. Six spindly legs are attached to a rectangular metal torso. Two long wire whiskers stick out from what would be its face, if it had a face. What it does have is a name, Genghis. 21-year-old Colin Angle, a nerdy bespectacled student, has spent the past three months working day and night to build Genghis with his lab mates and Brooks. Today will be the first time Angle will ask Genghis to do what it was built to do. Walk. Engel's Bachelor of Science thesis is riding on this. Genghis doesn't have a central brain or master computer telling it what to do. The robot only has four tiny onboard processors programmed with short bits of software code. The code tells each leg how to react when sensors embedded in the legs encounter something. If the programming works, Genghis will sense when it's on the ground and will move its legs forward. If not... Well, Angle's thesis is in the toilet. Angle switches the robot on. Speaker 1 (8:55) Come on, Genghis. Come on, Genghis. Speaker 2 (8:59) The robot's legs move forward one at a time. It's not graceful. It skitters on the floor like a stunned roach. But it is walking. Brooks wants to see more. He grabs a textbook off a nearby table and tosses it down in Genghis's path. Genghis' whiskers hit the book first. The robot pauses. Angle holds his breath. Slowly, one leg lifts. It swings the leg forward. Then it steps back down. Another leg follows. And another, and another, until the robot is fully on top of the book. Then, Genghis moves forward. Brooks grabs two more books and stacks them on top of each other. Genghis climbs right over them. Brooks gives Angle a high five. As Genghis moves around the lab, it looks like it's thinking when it encounters obstacles. But it's not. It's just reacting. And yet, in 1988, there's no other robot in the world that walks and adjusts to a changing environment on the fly. Brooks has been building all kinds of robots for the past three years, but Genghis is a massive breakthrough, one that proves his simplified approach works. Brooks and Engel believe this is the future. But outside this lab, many peers just think they're delusional. Speaker 2 (10:37) It's July 1990 on the MIT campus. Inside Rodney Brooks' lab, strange creatures are crawling around. A colony of wheeled robots zip along in a single-file line like an army of ants. A palm-sized robot does a moonwalk across the lab floor. A robot that looks like a trash can speeds away any time students get close. Speaker 2 (11:00) In his office, Brooks is reading a Boston Globe feature about his work. It says his simplified architecture has deeply divided the artificial intelligence community. Detractors claim his robots are a huge step backward for AI. Brooks grabs a pair of scissors from his desk drawer. Speaker 2 (11:22) He cuts out a couple of the most stinging quotes from the article. One says Brooks makes cute bugs, but his software should never be used to do something complex, like drive a car. Brooks gets up and tapes the quotes to the door in his office, which is already covered with newspaper clippings predicting his work will fail. He treats it like locker room motivation for his students. As Brooks hangs up the latest takedown, his grad student Colin Angle arrives for a meeting with him. Speaker 2 (11:53) Colin, despite what everyone else is saying, you and I both know we've really got some good tech here. You know, I've been thinking. I want to start a robotics company. Angle grins. That's great. Then I want to be CEO. Oh, okay. But I'll be chairman. And chief technical officer. Now, what are we going to make? Well, I don't know exactly. But it should be something useful, right? Not factory robots, not something only the government can afford. Something ordinary people would actually want, Speaker 1 (12:26) right? I like that. Angle grabs the phone off Brooks' desk. Speaker 2 (12:32) He dials the number for Helen Greiner, a British-born MIT graduate and a good friend of his. Greiner's working for a California company, doing robotics research for the government. Angle knows she has the business experience he and Brooks lack. Helen, it's Colin. Speaker 7 (12:51) Colin! Speaker 2 (12:52) Listen, Rod and I are starting a robotics company. You should join us. Exactly. Which means you know what you're doing. Tell her she'll be a co-founder. Rod says you'll be a co-founder. Come on, Helen. You don't want to be a government contractor for the rest of your life. You want to make R2-D2. Or, I don't know, something like Speaker 1 (13:15) that. Something that matters to people. Come on. We can do this together. Speaker 2 (13:20) There's a long pause on the other line. Speaker 7 (13:23) Okay. I'm in. Speaker 2 (13:26) In the days that