The New Yorker Radio Hour - Don’t Worry, the Robots Can’t Do Your Job—Yet

Episode Date: December 12, 2017

The business reporter Sheelah Kolhatkar has recently written for The New Yorker about a wave of advances in robotic technology that will have dangerous implications for our economy and political stability. As more and more factories automate, many workers have found employment in warehouses, performing jobs where human dexterity and brains still hold a strong edge over clumsy robots that can’t recognize unfamiliar objects very well. But as robots advance in gripping skills, visual recognition, and problem solving, a dangerous wave of unemployment may loom. Kolhatkar speaks with a roboticist, an economist, and the C.E.O. of a robotics company, Symbotic, which is taking the people out of warehouses. Symbotic’s robots don’t earn pay, they don’t need health insurance—they don’t even need lights or heating to operate. Plus, Fabio Bertoni, The New Yorker’s lawyer, reveals what he does on the very rare occasions when he’s not at work. New Yorker Radio Hour listeners, we want to hear from you.  We have a few questions about the show and how you listen to it. The survey takes about twenty minutes, and your feedback will help us make our podcast better.  Take the survey here.

These are just anecdotes, but it's building up into something more coherent. I think it would be interesting to really try to unravel what his ties. There's a sort of country-city divide for their own convenient, and then it's not clear where it goes next. From One World Trade Center in Manhattan, this is The New Yorker Radio Hour, a co-production of WNYC Studios and The New Yorker. Welcome to The New Yorker Radio Hour. I'm David Remnick. Staff writer Sheila Cole Hadkar covers a lot of business stories for The New Yorker, and she's been looking at a tremendous economic change going on in our country right now. Sheila's been reporting on how automation is making its way into industries

that are still depending very much on human labor, with robots that are much more advanced than what you might imagine in an automotive factory, for example. And every time robotics makes another leap forward, economists start worrying with good reason. Here's Sheila Cole Hatcar. Well, I first became interested in this because income inequality is one of the defining issues of our time. And it's one of the major reasons that President Trump won elected office. It was one of the strong motivations of many of his voters. And it occurred to me that technology and automation might be partly to blame.

And this actually defies conventional wisdom because in the past, technology has eliminated certain kinds of jobs. But it was largely seen as creating new and new. better jobs. But that seemed to not be happening anymore. So I wondered what was going on. I have really interesting experiences over and over. I talk with Andrew McAfee, who's an economist at MIT. Because when I give my talks or engage in discussions about this phenomenon, very often afterward, I have fascinating hallway conversations with CEOs who will come up to me and say, hey, where all the jobs are going to come from? And I say, aren't you the job creation engine?

and they say, I can identify the technologies that we're investing in right now that are going to put very, very large numbers of people in my organization out of business. Do not quote me on this. So if it's true that robots and automated machines are replacing human workers in large numbers, we could be on the verge of one of the biggest waves of unemployment we've ever seen. So why are things changing now? Well, robots are getting smaller and faster and more precise in their movements, and, you know, most importantly, they're getting smarter. They can problem solve on the fly better than they ever could before. I talked with a professor named Stephanie Tellix.

She runs a computer science lab at Brown University. She and her students are at the cutting edge of research to make robots more dexterous and smarter. I'm personally more interested in manipulation scenarios where the robot can pick things up and put things down. It can move not only itself but other objects in the environment. Can I have the metal object over there? This one?

No, the other spoon. Final answer. You wanted object six. The way Stephanie sees it, robots are a long way away from actually replacing people. We know so much and we do so much intuitively, and robots have to be taught how to do absolutely everything. If you imagine that the robot, all it can do is deliver objects. So I'm saying get me the water bottle, get me the food, get me the Kleenex box, get me the phone. And my success rate is 90%, which is probably more or less where state of the art is at recognizing the object. That means 10 times in a day I'm going to fail to follow the command successfully.

It's not going to take very many drops before you're like, uh-uh, don't want you, robot. I'm going to go get my water bottle myself because I don't want you to drop it. And on a factory line, it's the same thing. Like if I'm a 90% pick success rate and I'm dropping 10, I am not a cost of effective robot. It's not just the physics of picking things up. Robots aren't great at recognizing objects they haven't encountered before. So what we've done is created an approach for a perception that allows the robot to automatically adapt itself to the environment and the objects that it's asked to manipulate. And so you can imagine that that factory robot, like if it's on assembly

line, it's doing an assembly task, there's like 10 things. Maybe it's got to pick up. You can imagine that by adapting itself, it can reach a much higher level of precision at recognizing objects and reach the rates of reliability that are needed to be successful in those tasks. Those kinds of advances are having a huge impact in places like warehouses, like the order fulfillment centers that service Amazon and other e-retailers. Warehouses have been a really bright spot in the economy for workers who don't have college degrees who perhaps couldn't find jobs in manufacturing. But I visited a robotics company called Symbotic, just outside Boston,

and they're working on making warehouses much more efficient and practically human-free. Welcome to the machine. Chris Gaghan is the CEO. So this is the, we call it the ITC integrated test center. They have all sorts of robots of different sizes, including a fleet of adorable little race cars that zoom around big cages. Each vehicle weighs about little less than 300 pounds. They travel at 25 miles a night.

hour. They kind of look like a little go-cart. We have a, it happened to be green. So in a symbolic warehouse, big stacks of goods come in, robots unpack them, the goods get stored all around this giant storage cage that looks like something out of the matrix. And little robot cars zoom around, gathering up the goods whenever there's an order, and they repack everything at the end onto these pallets, which are these tall kind of stacks of crates of goods. So here's where we can see the kind of research Stephanie Tellix is doing, playing out in the real world. The part that's revolutionary isn't just the robot's ability to quickly grab stuff. It's their brains.

It's the computer's ability to recognize and handle a wide variety of objects. That wouldn't have been possible without the advent of much more powerful computers, really inexpensive sensors. And a lot of those sensors were made because of what's happened with smartphones and other computer and consumer electronic devices that just wasn't possible 10 years ago. Most warehouses are built for human workers, and the pallets are usually stacked in the order that's best for humans. So taking into account worker safety, minimizing exhaustion, the potential for human error. So this stuff never arrives at the stores in the order that the stores would like.

The difference is when we build the pallet robotically, we build it in the order that the store wants it. So when the pallet arrives, instead of a human being having to downstack that pallet onto carts, they can take that pallet directly to the floor and start restocking the store. So what Chris Gahagan is talking about might sound like a minor point. But if you run a store, especially a chain of stores and you're waiting for stacks of goods and the robots have already arranged everything in exactly the order that you needed in, that's actually a really big deal. And what that allows a store to do is be more efficient with their labor,

but more importantly, reduce outages in the store. So they can start to replenish the store much faster because now when that, truck arrives, there's no work that's happening in the back of the store, literally that product is taken immediately from the truck out to the store for restocking. A symbotic system is really expensive to install, but afterward a company saves a lot of money. Not just in wages and benefits, but the company also saves on lighting and heat. People in the field talk about dark factories. Well, this would be like a dark warehouse. Lights are only on in the front of the system where human beings are running the system,

and we've done kind of initial analysis compared to a manual-based warehouse that has fork trucks and people in it, we are about 30% more energy efficient. So what's going on at Symbotic reflects what's going on all around the country. It eliminates the need for many types of jobs, but it does create one new class of jobs,

that of the system operator, which is a higher-skilled, higher-paid job that requires more education. But there's only one of them. In this political climate, many of Symbiotics clients don't want people to know they're even installing this system. Every time the minimum wage goes up or might go up, the company gets more calls. Higher wages make automation more attractive. Andrew McAfee says that we can't look at America as one economy anymore.

They're different economies for different classes of workers. We also should keep in mind that we have added jobs month by month without fail in this country for well-over-a-lawful. over 80 months in a row. So we are not at peak jobs. The job creation engine of the American economy is still positive, is still kind of working. What I think is going on, though, is that that job creation engine has kind of kicked into a lower gear. And instead of turning out lots of really solid middle class jobs like it was doing back in the 60s and 70s, the job creation engine today is kicking out lots of lower middle class jobs that are lower paid, They're more precarious.

The hours are all over the map. And they're just not as good a job in a lot of different ways. So are you, did you see any signs of some of these tensions in the presidential election and the voting patterns? I'm sorry to ask you. I'm sorry. Is that a serious question? No. Okay.

No, nothing rang any bell at any point in time. All right. I have no idea what you're doing. But I'm sure, you know, it's hard to avoid the, well, does this explain Trump question? These job market challenges brought on by technology and globalization are not the full explanation for Donald Trump as our president. However, it is really hard for me to imagine that a message of isolationism and nationalism and really pretty naked ethnocentrism would have played to the electorate in 1996 in 2000. That just doesn't work for me. So I do think there is something about this feeling of diminished opportunity.

People suddenly getting a raw deal when they signed up for a fair deal. That brings us back to Stephanie Tellix, the computer scientist at Brown. For most of her career, she was really excited and optimistic about technology. She told me she was inspired to go into this from watching the Jetsons as a kid. Then, during the presidential election, Stephanie learned that her parents were planning to vote for Trump and that his argument that job losses in the U.S. were due to immigration had really resonated with them. She knew that income inequality was a huge factor in the election, and she started to wonder if her work was contributing to the problem.

If those problems are really caused by me, by robotics, that's bad. Like, it's bad they don't understand. It's bad they're being told lies. It's bad because they're not, like, if they're going to apply the wrong solutions. And I also felt like if also, if they figure out that it's, robotics, you know, if we don't tell the story right, someone else is going to tell it for us. And I'm not necessarily going to like how that story goes. One scenario people are worried about is the trucking industry and the question of what will

happen to truck drivers once self-driving technology really takes over the industry. Some economists estimate that over the coming years, 2 million truck drivers could lose their jobs. I don't worry about the raw numbers at all. The reason not to worry about that is that every month, in America, there are more than 1.5 million layoffs. More than 1.5 million people lose their jobs in America every month. So the notion that we can't absorb an additional two or three million layoffs is just dead flat wrong. We could absorb that with hardly a hiccup, especially if it occurs over a decently long period of time. However, it's the kind of people

doing those jobs that are a concern because they tend to be less well-educated, middle-aged, white working class, adding two or three million more of those people to the ranks of the unemployed is not a recipe for stability, I don't think. So what are the solutions? One idea that comes up is UBI, universal basic income. This is the idea that the government would give every citizen an amount of money that would cover basic living expenses. It's a controversial idea because it sounds sort of socialistic, very un-American, that you would get money to sit around and do nothing.

But recently, economists have started to talk about this more and more, and even the celebrity co-founder of Tesla, Elon Musk, has been promoting this idea. But Andrew McAfee is really skeptical. When I read about and when I go look around the communities where work has gone away, I don't see people starving. I don't think lack of money is the great problem. It's a lack of dignity, glue for a community,

meaning, purpose in life, stuff to do to fill up the hours of the week, things like that, things that a job is really, really good at giving to people. What we need to do is give the innovators and the entrepreneurs the best possible playing field to come up with new things for people to do. Job creation has always been a really decentralized activity. It's done by companies, by entrepreneurs, by innovators who want to go build, a good or a service or put a product out there, and they need people to do that. That game is not over. It has become very geographically. What's the word? Concentrated. Yes. I mean, that's it. I mean,

there was just a study released that I was saying that basically your zip code tells you everything you need to know about what your economic fate is. And if you're in the wrong place and those innovative types of jobs are not there or all the McDonald's in town of introduced their kiosks and there's nothing else for a service worker to do, that is a big problem. One of the challenges associated with that is, A, people just don't like to move willy-nilly, just chasing jobs all over, and B, the evidence is pretty clear that there are fewer moves now, that geographic mobility among American workers is going down, not up. One of the, I think, clearest lessons from history is that big technology changes, big surges in technology

necessitate big societal adjustments. For example, the last time we had a surge of technology this big was probably 100 years, about 100 years ago, when we had the one-two punch of electrification and the internal combustion engine. One of the things we did in response to that was set up universal compulsory, universal free K-12 education. That was extremely controversial when people first started talking about it. It was going to be too expensive.

Some kids were not smart enough to handle a high school education, but we came together and we decided as a country to do this thing. My huge frustration is I just don't even see the will or the desire or the awareness of the need to come together and to do big things. Automation doesn't ever seem to come up in Washington. Members of Congress never talk about it. You never hear the president mention it. It didn't come up at all during the tax debate. Automation isn't the kind of emotional issue that gets voters riled up. like immigration or outsourcing.

Automation just sounds kind of boring. But it's the future. There are a lot of things we could be doing to prepare for it. But the first step we need to take is to just acknowledge that it's happening. Sheila Cole Hadkar is a staff writer with The New Yorker, and you can find her article, welcoming our new robot overlords.

That's if you need a little pick-me-up at new yorkeradio.org. I'm David Remnick, and I'm here with Fabio Bertone who plays a key role as the New Yorker's lawyer. He's deeply involved in our reporting every week and every day, especially our investigative pieces, and he keeps us out of trouble about 20 times a day. And when I start to feel like my job is stressful, I just wander down the hall to Fabio's office, and I feel a little bit better. I wanted to find out how he takes his mind off of work on the rare occasions when he's not at work. So it is Tuscanini's 150th anniversary of his birth.

there's a new biography of Tuscany So I'm looking at the book that's here This is Harvey Saxes' biography of Tuscany I'm lifting it up and getting half a hernia Yes, it's gigantic And so what made you So dip in Well actually I was invited to a concert

Given by musicians from La Scala And it was just it's interesting to hear about him He's anti-fascist He left Italy who refused to conduct for Mussolini He came to New York and became the conductor for NBC. So it's an interesting story. These are the days when the networks had classical music.

And he had, you know, the 30% of the American public listening to classical music on the radio every week. That's amazing. I just made up that statistic. But it was a big number. So I'm reading this book and I'm watching YouTube videos. And there's one from 1948 of him conducting the NBC Symphony Orchestra. in Beethoven's Symphony No. 9, and he's just very dramatic. He has incredible facial expressions

as he conducts. He makes all these gestures. Some of them don't appear to be actually conducting gestures, but they're like Italian hand gestures that are remembered from my childhood. Obscenities? No, there's a point of why, he does. He does shake his head at the musicians a couple times. But there's a point where he touches the side of his nose, which is a gesture that, you know, in my experience, means like, be careful here, like, pay attention. So there's this quiet moment here. He touches it.

It's just a warning. He just wants to put them on notice. He looks fabulous with that salt and pepper mustache and this big arm movements of the flowing white hair. He's fantastic. Yeah. So my next pick is, I have a friend.

He recently had to get rid of his smoker. So he gave his smoker to me. And now I'm burdened with the grave responsibility of learning how to barbecue. And in one of those wonderful, fortuitous moments, I discovered on our book table here, a book called Franklin Barbecue. In Austin.

From Austin. Aaron Franklin, Franklin Barbecue, did a book. And it's a fantastic book. He's got a lot of history of barbecue there, the science of barbecue. So I am now learning how to barbecue. You know, I tried to get into Franklin's.

I got there at 11 o'clock in the morning in Austin, and the line was said to be it was two hours long, and I now regret it that I didn't do it, but I blew it. So while you're listening to Tuscany-conduct Beethoven, the smoke will slowly be... Exactly. Exactly. Embueing the meat with all sorts of good things.

Apple would smoke, apparently. You have more for us? One more, and it's Lucho Batiste. And Lucho Batiste was a pop star in Italy in the 70s. Well, I think we should play a little bit of this, but only if you sing over it. No, no, no. Absolutely, no. Absolutely not. As your lawyer, I advise you against.

All right, well, at least let's hear a little of music. Yeah, and, you know, you need to appreciate this. We will go full in making no other. Un-Ironic. That's the best way to do it. It's a beautiful song. Yeah. Got a little synthesizer.

A lot of synthesizer. A lot of synthesizer. And it gets more as it goes on. Wow, that's a sound that has disappeared from the earth a little bit. Here we go. Oh, what I wouldn't pay to hear you sing this in the car. And it gets loud at the end. With barbecue sauce all over.

Fabia, thank you very much. Fabio Bertoni, the New Yorker's Legal Council. I'm David Remnick, and you're listening to The New Yorker Radio Hour. Next week on the New Yorker Radio Hour, staff writer John Lee Anderson was granted a rare interview with the president of Venezuela, Nicholas Maduro, who could be on the way to making himself the dictator of Venezuela. What would it take for democracy there to end altogether? It's an essential story, and I hope you'll join us.

The New Yorker Radio Hour is a co-production of WNYC Studios, and The New Yorker. Our theme music was composed and performed by Meryl Garbus of Tune Yards with additional music by Alexis Quadrado. This episode was produced by Alex Barron, Emily Boutin, Ave Cario, Rianne Corby, Jill Duboff, Karen Frillman, David Krasnow, Sarah Nix, Michael Rayfield, Mithelie Rowe, and Stephen Valentino with help from Corey Shrepple, Johnny Vince Evans, Eric Malinski, Emily Mann, and Jessica Henderson. Our story on automation was produced with help from the Frontline Dispatch, a new podcast from Frontline. The New Yorker Radio Hour is supported in part by the Cherina Endowment Fund.