Flux: Robot Queen Helen Greiner on robots, drones and the self-aware Roomba

“If nobody else is going to invent a dishwashing machine, I’ll do it myself.” That proclamation from Ohioan engineer Josephine Cochrane — who was, incidentally, granddaughter of steamboat inventor John Fitch — started a revolution in homemaking that 60-some years in the making resulted in the popularization of the dishwasher.

Cochrane filed a patent in 1886 for the mechanical dishwasher and set up a factory to produce the machines. The Hobart Corporation later bought her business and went on to release the dishwashers under a spin-off brand, appliance giant KitchenAid. It may have taken more than 60 years, but by the 1970s, the dishwasher had become one of the most common home appliances.

Fast-forward a century and the “let a robot do your dirty work” dream has stood the test of time. Yet household penetration beyond dishwashing and laundry machines has been slow. In an interview for Flux, I sat down with today’s modern Josephine Cochrane, Helen Greiner, the co-founder of iRobot.

The company behind the first automated and commercially successful home vacuum, the Roomba, iRobot’s appliance hit the market in 2002 and has now sold more than 16 million units worldwide.

We got into how founders should think about timing a market, navigating user adoption cycles and iterating on product. In 2008, Helen made the jump from terrestrial to aerial robotics, founding drone company CyPhy Works to focus on applications including public safety, construction and agriculture. Helen also shared her thoughts on why the sky is a natural superhighway for drone delivery, how to get more women into technology, her love of Star Wars and what happened when she tried to fly her drone on the White House lawn.

An excerpt of our conversation is published below. Listen to the full episode on iTunes.

AMLG: What got you interested in robotics  —  has this been a lifelong passion?

HG: I did see Star Wars when I was 11 and R2-D2, he’s my muse. He had an agenda. He was one of the main characters. He had a personality and he was really more than a machine. I’ve always wanted to build devices that are machines but also more than machines. And I could see the connection between a very early computer that my dad brought home  —  a TRS-80 from RadioShack  —  and robots. So I went to MIT to learn how to build robots. Although I learned a lot of wonderful things there, they didn’t really know how to build robots at that time. At least not commercially viable robots.

The TRS-80 (Model I) was released by Tandy in 1977 and sold through RadioShack. “TRS” stood for “Tandy Radio Shack” and the “80” referenced the Z-80 microprocessor. It used Level I BASIC, had 4K RAM and cost $600 (~$2,400 today.) It was one of the first PCs and sold better than Apple in the early 1980s. [Source]

I started iRobot with two business partners, Rod Brooks and Colin Angle. We started out more with a technology than a product idea and it took us a while to get going. We started iRobot in 1990 and didn’t put the Roomba on the market until 2002. In fact, we didn’t start building it until 1998. In that time we were doing a lot of different robot projects, many with Fortune 500 companies.

The idea was, here’s some ambitious MIT kids and we’ve got all this innovation and we can put robotic technology into your product. We did get a toy on the market with Hasbro and we did put some robots out into museum displays. But they were mostly demonstration systems. It wasn’t until we raised venture capital and put money into real markets that we got the first ground robots out with the military and put the Roomba on the market.

AMLG: It sounds like you tried out a bunch of different ideas. You mentioned the toy you built for Hasbro; what was it and what’s the story there. Did the toy do well and what did you learn?

In 2000 iRobot teamed up with Hasbro to create the “My Real Baby” doll.

HG: It was a robotic baby doll and it had an emotive engine inside, it was behavioral control. It did a lot of wonderful things and we put our hearts and souls into it. Then we put it on the market and it didn’t sell. We still think it’s a great product, but what we learned from that was to really take cost into account as the first design criteria, how to get products manufactured in the Far East and manage supply chain issues. All the things you need to get a consumer robot on the market. We were able to get that learning under our belt, including the manufacturer we used for the first Roomba — the same manufacturer we initially used for the doll with Hasbro.

AMLG: How did you go from producing a robotic doll to producing a robotic vacuum? What was the decision process? I’m guessing it wasn’t exactly a straight line.

HG: We worked on 18 different kinds of robots in those dozen years. There was a lot more experimentation and learning that had to be done because we were very very early into the robotics field. You know what they say  —  pioneers are the ones with arrows in their back. We had a lot of learning from the robotic doll. But the project that pushed us towards the Roomba was a large cleaning robot we were doing with SC Johnson Professional. Our engineers were looking at it and saying, “this is great and it’s got a lot of potential, but if we get anything wrong it could do real damage. It could take out cleaning stores at night. It could take out a whole shelf of groceries. It may not be the right thing to do first.” So our engineers actually proposed to do something smaller. There’s a motto in engineering: keep it simple stupid  —  KISS. Potentially the right thing is to do the simple thing first and get the experience under your belt. Even though there was obviously a great market in larger commercial cleaning robots, the right thing to do was get the technology working on a smaller scale first.

The Omnibot 2000 was a remote-controlled personal robot made by Tomy in the 1980s. It did not have a real onboard programmable computer, but could pour drinks from a serving tray and play back audio. His eyes also flashed.

AMLG: When I was little my brother and I had a remote-controlled robot, the Omnibot 2000. We used him to deliver drinks to our parents when they had guests over. But having a robot like that in the home was still entirely novelty value, it wasn’t practical. I’m curious, how did you think about getting consumers used to the idea of having a robot in their home, in the year 2000-2001 when that was still pretty bold? Did you think everyone was ready socially and culturally?

HG: Our mentality was why wouldn’t anyone want a robot in their homes! Of course everybody wants robots in their homes. But what we discovered is that people buy it as an appliance. They buy it as a cleaning device. They buy it because it does the job more efficiently and more effectively than a human. And it’s more effective because it does it more often. You set it on a schedule and it does it every day rather than you taking out the vacuum once a week.

AMLG: And having a laundry machine or a dishwasher was a big deal when they first landed, and now of course we take those machines for granted. But having an object moving around you constantly feels different, no?

HG: Absolutely. There was a lot of pushback on dishwashers at first. There were a lot of people saying, no way, it’s not efficient. It wasn’t until people saw that it does something new  —  that it actually sterilizes the dishes better because it can operate it at a much higher temperature  —  that people said, OK, this is a need that I can’t fulfill another way.

AMLG: So that’s the commercial side. But the other massive achievement was the PackBot, which identified and disposed bombs in Iraq and Afghanistan. You’ve shipped over 6,000 of these robots to the troops and saved thousands of lives searching for IEDs. Can you tell us a bit more about that side of the business?

510 PackBot performs reconnaissance for chemical, biological, radiological and nuclear (CBRN) detection, building and route clearance, explosive ordnance disposal (EOD), HazMat handling, improvised explosives device (IED) detection.

HG: One of the things we did along the way was take government research contracts to develop the technology. Those research contracts actually turned into a bomb disposal robot. When September 11th happened we were on site the next day, searching for survivors in the surrounding buildings. Then we went into Afghanistan where they were having terrible trouble with cave clearing. They would tie ropes around the soldiers and send them into the caves. That’s a tough job. The last occupant was an enemy combatant so it could have booby traps, tripwires, weapons caches. The idea to send robots in went to the highest levels of the military and they actually started the Rapid Equipping Force at the Pentagon to get technology into the hands of the soldiers quickly.

One of the things I’m very proud of was getting those robots onto the quick reaction force, getting bombs disposed of, being able to save the lives of hundreds of soldiers and thousands of civilians. That made it easier to start programs to get other commercial technologies into the hands of soldiers without going through the regularly scheduled acquisition processes, which take decades.

AMLG: It’s incredible how fast you were able to do that  —  to get them into the hands of the troops. So how did that work having both a product in the consumer market, but also having these high-level defense discussions going on and shipping product to the military. How did you juggle doing both, was it symbiotic?

The Roomba 880.

HG: We were able to juggle it initially  —  they both hit at the same time. We deployed the PackBot the same year as we put the Roomba on the market. It was a very good year for us. iRobot wouldn’t have existed without both. Because the Roomba got great sales for the first Christmas but then it started going down. There’s been ups and downs along the way and now it’s going gangbusters because the adoption cycle has really taken place. But without the military  —  and we had contracts with hundreds of millions of dollars  —  without that revenue coming in, I’m not sure that iRobot would have made it.

AMLG: It’s fascinating that you used that side to fund the business until consumer adoption was ready. So was it 10 years between the founding and putting the Roomba on the market?

HG: It was really 12 years before we put the Roomba on the market. It took us about four years to develop it, starting in 1998.

Helen in Wired, “Why Robots are as Interesting as Humans”

`

AMLG: Wow, so 12 years. Of course, one of the common fears for investors and VCs around frontier technology is that mistiming is death, or as you put it, it can be arrows in the back if you’re too early. Many of the founders I work with in VR/AR, and blockchain, grapple with these questions, navigating what lines of revenue they can pursue earlier or how long to plan to be in cockroach mode. So can you expand a little on that  —  what did you learn about timing one’s product to hit the market? Are there any clues to figuring out when the adoption cycle will pick up? I mean, now you’re building a drone company. How do you think about timing the company?

HG: Well, it’s all timing, and I haven’t always got it right. We got it right with the Roomba, although it did take a little bit longer, but now it’s a large percentage of vacuum cleaner sales around the world. One of the things we tried to do is put an internet-connected robot on the market in 2000. That was a bit ahead of its time. The timing is about now. So we were maybe 17 years too early. With the drone market I really believe we’re in the right place at the right time. All the major industrial companies are considering their drone strategy and we have technology that helps them by providing persistent, rugged, reliable environmentally sealed drones.

AMLG: So what lessons are applicable from iRobot to CyPhy Works?

HG: What we learned from iRobot is to look at the market first. We knew that it had to be easy to use. They like to say you can set it up in freezing weather with gloves on in 5 to 10 seconds. That’s the kind of ease of operation that’s required. They need persistence. Battery-operated things constantly need recharging, but people are busy and they can’t pay attention to getting a system recharged. So all the learning  —  from the little things like what’s needed in the design, to the big picture and what kind of functionality is needed  —  came from experiences at iRobot.

AMLG: So your thinking is, we’ll start with military applications and if it can survive those harsh conditions and demands, it’s definitely going to be good enough for commercial applications?

HG: Absolutely. They have a lot of the same needs. Being able to work in all weather, being able to work in high winds, being environmentally sealed, being persistent, not needing operators constantly watching the system. No piloting required. All that is very similar to what we’ve built for the military market.

AMLG: On the commercial side I saw the announcement that you teamed up with UPS to do a test near Boston, a three-mile drone delivery of asthma inhalers to a kids’ summer camp. It sounded like the drones had to fly over the ocean for part of it. What did you learn from that test?

HG: What we learned? That it’s possible to do this kind of a mission completely autonomously with a drone. Everything from the take-off, to navigating to the island, to landing and drop-off. The timing, the amount you can carry, all those things seem to line up to make drone delivery an effective solution for those hard to reach places or times when you need something quickly and you don’t want to or can’t wait.

AMLG: When it comes to drone delivery, the elephant in the room is Bezos. I know you know Jeff, and that he was an investor in iRobot. And obviously drone delivery is a big part of Amazon’s plans for world domination. Some people even say you inspired him to think about drone deliveries. So any comments on Amazon’s drone plans?

HG: Well I don’t know about that, but I know I was speaking publicly about drone delivery before he went on 60 Minutes and announced Amazon’s plans.

AMLG: Ha! So a non-answer answer, funny nonetheless. Well, it’s clear you’ve got this visionary thing going where you see the markets long before others do. So why in 2008 did you decide the timing was right for drones?

Greiner, Brooks and Angle at the iRobot IPO in 2005.

HG: When I looked around at what to do after iRobot  —  I did iRobot, was president and chairman for 18 years, took it public, stayed on for three more years  —  I was thinking, when you’re flying you avoid a lot of the problems with ground robots. The problem with ground robots is the ground. If you hit an obstacle or a fence or a ditch or stairs that are just a little too high or rough terrain. That’s a big problem for a little robot on the ground. But when you’re able to hover and fly, all those obstacles kind of go away. So it’s almost a magic technology to be able to get things places. If you look around the room you’re in, just above head height you could imagine flying in and being able to get anywhere. It’s the same thing above the treetops and buildings. It’s almost like a superhighway made for drone delivery.

AMLG: But even highways in the sky have to have rules and regulations. Are you closely involved with those regulatory discussions?

HG: Yes we’re supporting the FAA’s effort to de-conflict the airspace, but I don’t think that’s going to be the major problem. In general, commercial planes don’t fly low to the ground and there are solutions like at Airmap to de-conflict particular drones from running into other drones. There’s ADSB to know where planes are flying. It’s a big air space and these are comparatively small drones. When you look at how many obstacles a self-driving car will have to avoid  —  other cars, pedestrians, bicycles, dogs, children, anything that could come onto a road, and everything’s trying to use the same space. Up in the air you can just stay far away from things that you have a technological beacon or a GPS position of.

AMLG: What about interior drones. I know that UPS has been exploring what they might do with their warehouses internally, are you involved with that?

HG: We’re building some interior flying robots, one we call the Pocket Flyer. The first ones we’re building are for reconnaissance. Going into buildings, going into tunnels, culverts, places that you can’t easily access. We’ve worked with the Air Force on those technologies. Places that are dangerous to go.

A CyPhy drone.

AMLG: So you revolutionized cleaning and introduced robots into the home. Do you think there’s going to be a similar learning curve for average people on the street getting used to drones flying overhead?

HG: I’m not sure that’s going to be an issue. Some people have said to me, “I’d hate to see the sun blacked out by the swarm of drones that come and deliver the packages to all the houses on my street.” I just don’t think that’s the way it’s going to work. A truck is a very efficient solution. I don’t think we’re going to replace that. We’re going to be delivering when somebody needs something in a timely manner. We’re talking about like one drone, a few drones here and there, not a large number. They’re pretty small and you can’t hear them. You can barely see them. People are used to helicopters and planes flying overhead and nobody seems concerned about that.

AMLG: What’s your feeling on how long this will all take? When is this going to be a reality at scale?

HG: For the delivery drones I think it’s going to be at least 2020, because, as you mentioned, we need to overcome some cultural hurdles but also regulatory hurdles and there’s still some technology that needs inventing.

AMLG: What are those remaining pieces of technology that we still have to tackle?

HG: Well, the airspace deconfliction. Sense and avoid. Also aerodynamic structures that fly longer and carry more payload.

AMLG: You’re also focused on STEM education and technological literacy for kids. How do you think we can have the most impact there?

HG: I’m in a unique position for that. STEM is not just a wonderful field to be in  —  to build and design things, it’s a creative enterprise. Being able to let children know that it’s working with people, it’s working with teams, it’s getting something impactful done. But it can also be an on-ramp to entrepreneurship. Some of the most powerful CEOs studied engineering because knowing what’s inside the products is extremely helpful. It really appeals to kids to say, hey maybe I can grow up and build a Roomba, maybe I can grow up and build a robot that saves lives. So I try and get out as part of this program to speak to a lot of children. I don’t have to say “oh and you can do it if you’re a woman,” I just need to show up.

AMLG: I mean you’re a living, breathing example of how to be not just a badass woman, but a badass human. If I were an 8-year-old kid and you came to my classroom I know I’d be sold on building robots.

The National Assessment of Educational Progress (NAEP) administered a technology and engineering literacy test in 2014. Out of 21,000 eighth-grade students surveyed 45% of girls scored proficient compared to 42% of boys.

AMLG: On your point about inspiring girls. I read a study the NAEP ran. The study was on technology and engineering literacy (TEL) of eighth-grade students. When you break down the results by gender the 14-year-old girls actually did a few points better than the 14-year-old boys. So the interest and ability is clearly there early on. But it drops off. How do we get that to translate into careers, especially with girls. How do we maintain that interest over time so they become inventors and entrepreneurs and engineers?

HG: Well at the risk of being stereotypical, a lot of girls that I talk to, what they want to do with their careers and their lives is help people. But they’re not given the information that when you’re an engineer you can really have an impactful career helping people and changing lives. Maybe you’re going to make better potable water or self-driving cars that saves lives or you’ll help the environment with solutions to global warming or better energy technology. There’s so many things that can have a big impact and relaying that back to children, that’s what’s needed. Because a lot of children in general and little girls are inspired to go out and change the world.

AMLG: That’s an interesting point. I hadn’t thought about it like that. It definitely feels like an important framing that’s been missed. So on this question of impact, if you had 30 billion dollars what would you do with it?

HG: There’s a lot of suffering in the world. I am reading about the starvation happening in South Sudan. It’s incredible to me in 2017 that things like that still happen in the world. The refugee crisis is near and dear to my heart because my dad was a refugee out of Hungary in 1956, going to England. He got a lot of support from the English system. Refugees were given a house. They were able to get great educations and go from being a refugee to having a fulfilled life. There’s so many refugees today there’s such an immense problem. I’d like to help with that.

AMLG: OK, into rapid fire. What would you say is the most surprising thing about you that people don’t know about you?

HG: Probably that I love sports. I try to pick up one new sport a year. Everything from kayaking to kiteboarding to wakeboarding to snowboarding. I just love getting out there.

The self-aware Roomba ponders the meaning of life [@selfawareroomba]

AMLG: Seems like you’re just good at everything; athlete, engineer, entrepreneur, presidential ambassador. Second question  —  thoughts on the self-aware Roomba? As in the famous twitter account.

HG: Oh yes, I definitely follow self-aware Roomba.

AMLG: It’s gotta be one of my all time favorite Twitter accounts.

HG: That and the sarcastic Rover.

AMLG: OK, third and final question. The latest Star Wars series. Did you love it or hate it?

HG: I love the new ones. It’s old school, back to like the original. So yeah I absolutely loved it. I loved the new robots. I love that the robots are still the major characters. Of course I would love to see more R2-D2, but you know.