Technology

our jobs. But there's another view that says robots aren't a threat but a liberation, an opportunity to get away from the daily grind and focus on what's really special about being human. It's a nice idea. I think it's particularly appealing to Silicon Valley types who are programmed to always believe that technology will make the world a better place. Later on, I'm going to talk to Bloomberg columnist Noah Smith about whether that's a realistic vision for the whole economy.

But first, blob Like Economy report to Craig Torres went to a corner of Virginia where robots and humans do seem to be each playing to their strengths with not a dog insight. Every day we read stories about how automation and artificial intelligence will make humans less relevant. Many of these stories are alarmists suggesting robots will be doing what shop floor machinists or even doctors do every day.

The real story is more subtle. I visited a state of the art Rolls Royce facility outside of Richmond, Virginia to find out how automation is changing human work. I asked the plant manager, Laurence O'Dell, to take us on a tour and explain how humans and machines shake hands here to produce something very valuable and very precise. This plant makes jet engine parts. I have pulled into the

parking lot of Rolls Royce. The plant stands kind of in a forest, and as you drive down the road you see this modern looking facility with the very familiar insignia. Now I have to watch a safety video as I go into the plant. Once I am inside, I asked Lauren about what work was like when he was a young engineer in the night and an automotive engine plant. How close were men and women too, metal very close, very very close. In many ways. There were a lot

of hand assembly operations. If you were assembling an enginengine. It wasn't what we would recognize in any way today as smart automation. It was electrically speaking, following some some pain to dance steps on the floor one to three four, one to three four. When I look at the parts Rolls Royce produces here, they are finely cut like pieces of ornate industrial jewelry. One part, called a fan disc, starts out as a big cylinder of titanium, about as round as a tire you might see on a tractor

trailer on the highway. This disc is integrated with the fan blades you see on the front of a jet engine as you board a plane. The raw disc costs as much as fifty thousand dollars, so it has kept in a big suitcase until it has loaded onto the machining process to protect it from scratching and damaged. After it's machined, it's about the size of an all weather tire on my pickup truck, and it's just as articulated and shiny and deeply grooved. There their ultra precise parts.

They're made out of materials that are right at the ende of right at the at the cutting edge of of what engineers can create so these are hearts that are spinning at between six thousand and twelve thousand revolutions per minute, and they're they're extremely highly stressed, uh physically as well. You don't just walk on the shop floor here, you have to dress for it works pretty good. So

we're all dressed in steel toe slippers. I have my belt covered with a velcrow pad, my wedding ring is off, and I have something over my eyes, so I'm pretty scratch proof, i'd like to think. So if you were wearing earrings, we let me go with those work wears. Who I haven't. I haven't been out here since about eight o'clock this morning, so I have no idea what's going on. The rolls Royce plant is cavernous, albeit not gigantic.

There is lots of natural light. I feel like I'm in a workshop, but already you can tell there's something different about what we mean by work here. There aren't rows of men and women standing at stations doing an activity as in an auto plant. Instead, there are a series of large pods about half the size of a subway car with a window. Parts are placed within them, and this is where the magic happens. Inside a robotic arm picks out its own cutting tool, It measures itself

and stores this data along the way. Lauren explains, this whole process is about turning a mathematical description of a physical thing, a jet engine part, into reality. The car by drill bit trims cuts amid a splashing milk colored lubricant. The plant is relatively quiet and very clean. Where are the humans. They're not running this, They're making sure this automated process is repeating and repeating and repeating again without

much variant. There are essentially no manual operations here anymore. Though. UM we're using in our machines what you would recognize as big data, okay, to understand not only the parts and how the parts are being manufactured, but also the basic condition of the machine itself. Seems to be accelerating a little bit every week, where every week we're either implementing or we're taking a step toward implementing, UM a new kind of predictive technology. I have never seen anything

like the shop floor at Rolls Royce. This plant is less than eight years old. Something entirely new and different is happening here. Humans are not on the clock like pace of a typical manufacturing plant. The robots operating behind the windows of the big boxes are constantly self optimizing, that is, re measure ring themselves. I asked Lauren to explain this to me. What he describes is straight out of what twenty years ago would have been a comic

book kind of futurism. Humans are maintaining the process but also standing back from it in a way that seems unusual. This is how Lauren explains it. What it frees the humans up to do is to think about how can we make something better? So people to a significant extent aren't as tied to the equipment as they have been in the past, and they're really being freed up to work on higher order activities. Is that more human work? Would you say? Or not? Right now? I think it's

more human work. Okay, I'm starting to get it. In an ultra automated plant like this, humans are a step away from the whole process, less close to the metal, but still very familiar with exactly what is going on. So familiar, Lauren says that even the flex of titanium spinning off the cutting tool can tell them something is off. Lauren says he needs both skills people who know the guts of the process, but can think beyond it, troubleshoot it, and make it better. How do you train for something

like this? It sounds challenging to answer that question. I drove across the bottom of the state to Danville, Virginia, where Troy Simpson trains young men and women at a community college with a special third year program in advanced manufacturing. The location of Danville is highly significant. This city once hosted one of the South's largest textile companies, Dan River Mills. It's a bit like visiting an Athens of the South. There are old brick temples from the tobacco and textile

industry everywhere. It's some are falling into ruin, while others have been repurposed and restored. This area was hit hard by NAFTA and the China trade. The state and county are focused on restoring its prosperity by creating a new kind of workforce that modern manufacturers will find attractive. They've had some success. Here's Troy explaining how they are trying to create this new workforce as factories are looking towards a digital factory and pushing so much data, making intelligent

decisions through AI. Then we still got to have somebody to set those factories up and the troubleshoot those same problems that we have whether factor is digital or not. Right, But those technicians and now need to understand what is happening upstream and downstream of a process. So now we've got a technician that it's a whole plethora of things that could be causing those problems to maintain the process this so we have built on the facility to do

to train that technician in that very environment. So let's take a step back and figure out what we heard and saw. Rolls Royce has a perfect mathematical description of an engine part humans and machines joined forces to try and hit that perfection. Every day humans are looking at data as much as the things they work with. The machines and the metal work is becoming more conceptual, less heavy and dirty, plants are less hierarchical, and there's more

gender equality as work migrates in this direction. I like the way Lauren puts it. If we blindly believe data and we don't challenge it to say, does that make sense? Does what the machine told me makes sense? Can I challenge that? And I can can I verify it? Or am I just going to blindly believe it and make an offset to thinking Whatever it is, whether it's a physical machine or computer model or whatever, Am I just going to make that make that change on the basis

of of what it's told me. If it doesn't make sense, then why would you allow it to make the change? Or maybe we need to change our paradigm. Maybe we need to update our paradigm of how that little piece of the world works. And I think that for now, and I think that until artificial intelligence really becomes prevalent, I think that's where humans are always going to be taking a step up, not a step away. Don't think I don't think of it as a step away. I

think of as a step up the humans role. And there is no organism or machine better equipped to do this on the planet than a human being is to recognize, in the purest sense, a non standard situation. Yes, there is a quiet revolution happening in all trades, from medicine to making term both fan disks as artificial intelligence and automation creep into more of what we do. What does

it mean for humans? The stories I'm hearing are enormously positive yes, there will be fewer people working in a particular plant, but that doesn't mean there are going to be fewer people working in manufacturing. That depends on how quickly we can train people to work in this space where they understand a job as not producing one thing, but it's part of a larger creative process, full of data and feedback that aims at a more repeatable, more

perfect thing. I'm Craig Trusts, an economics writer for Bloomberg News. Well, I'm joined now by Noah Smith, Bloomberg opinion columnist on all Things interesting in economics. He's in our San Francisco studio. I mean, no, I think I guess the big question first coming out of that report from from Craig Torres, I mean, do you buy the idea that the robots and artificial intelligence could be the best thing that ever happened to humanity? I think the best thing that ever

happened to humanity is a little exaggerated. Um, But what's interesting is that most jobs can probably not be completely automated. Usually, what happens is that some tasks in the middle of the job get automated. In other words, a robot won't take your job it'll take half your job. And of course that really reminds me of what happened to menial labor and factory work during the Industrial Revolution. Before the Industrial Revolution, you had people putting together every bit of

every product themselves. Now you have these machines that can do some of the things, but they don't do all of the things. So you have assembly line workers doing some of the things, and you have um basically manufacturing workers doing some of the tasks. And that made the

manufactured goods a lot cheaper. Demand for the manufactured goods released biked, and therefore, even though half of craftsmen's jobs are being taken by these machines, you really had a lot more employment and a lot higher wages in the manufacturing sector during the Industrial Revolution. And I think this could be the same thing for service jobs. You know,

service sectors really have had low productivity growth. We've seen huge productivity growth and manufacturing very slow productivity and services. It's possible that this could change that and could make services behave more like manufacturing used to. I mean, it's interesting you talk about productivity. I do always find it quite funny when you're in these conversations where people are talking.

You know, what should we most worry about, you know, the end of the terrible things happening in the global economy. And of course, you know, one of the big things we talk about is the very slow rate of productivity growth, the fact that we're not getting more out of the same number of people, are not at the same rate we have in the past. And that's obviously hugely important because it drives wage growth and standard of living in growth.

But the people who are really gloomy about productivity growth are often also the ones who are really gloomy about everyone losing all their jobs because of technology, and you sort of you can't really have you know, if technology does get rid of all those jobs, it is inevitably going to raise productivity. I guess the question is whether the benefits of that get um distributed equally. But I guess when people think about this what they tend to come back at and say, well, we know that's true,

we know that it can't be completely automated. But if it can be automated, I mean, you know, if you had all of the picking for you know, in a warehouse, for grocery shopping or whatever else, if it all gets done by robots, but it's not accurate, and you need to have one or two people wandering around, you know, putting the robots back up when they fought, when they bump into something, or you know, correcting it when they've

when they've gone wrong. You know, that is still a loss of jobs that have been lost in that process. And this just feels like it's going to be on a greater scale than in the past. You buy that it could happen. The entire question is where or not demand expands. So what was interesting about manufacturing is that you suddenly had um people having so much more stuff,

physical stuff, manufactured things. People now had cars, and they had TVs, and they had much larger houses, and they had many more bulls and many more dresses and blah blah blah blah blah. And because this demand expanded, even though each each dress, for example, that was made didn't require as much human labor as before, overall people were getting paid more for doing human labor involved in dressmaking than they were before, much more, much much more so.

Because demand expanded in this way, UM, we had increase in employment, increase in wages, increase in productivity. All at the same time the question is could a similar thing happened with services the The dissatisfying answer is that a lot of different things could happen. But I believe people

aren't paying enough attention to the good possibility. People are fixating on this extremely simple story of like you get replaced by a robot and now you have nothing to do, and people are completely ignoring that upside of the story, you know, focusing less on trying to bring about the good future because they're not even trying to visualize it.

Your argument is it goes back to what people point out, even with the adjustments that you've had so far, you know, since the introduction of the a t M. I know, lots of bankers in the US like to point out that the number of factual cashiers in the U. S economy has increased dramatically since the introduction of a t M S. So you don't end up necessarily with the replacement. But I think even your own examples highlight that there's going to be adjustment costs and a transition where the

jobs that are created. You know, even on the certainly on the old version of the way it happened in the past, you had a lot of jobs created that were very different from the jobs that have been lost.

And if you're thinking about what happened to those individuals in those particularly in sectors that were really affected by technology, and in the sort of manufacturing technology, you know, the story of those people and their ability to get into the newer sectors is not not really a very happy one.

I'm certainly thinking of large parts of the manufacturing sector in the UK where people lost their jobs and the industries moved on, and those people did not necessarily ever work again, or they certainly spend a long time unemployed. So you know that has to be a question mark. Sure, Nor can I say that you know we're going to have a happy future where absolutely everybody has higher wages

and better jobs, etcetera. We're not, um, We're going to have a complicated future where some people lose out and some people win. And the question is on balance what happens, is it good or bad? What can we do to make sure that there's more people out there who are gaining than losing a lot more, What can we do to sort of steer this CATEC universe generally in the right direction. With our limited powers of government and are limited understanding of what's going on and the you know,

the results of our actions and the current situation. We have all this limited understanding. We're operating in the dark, and we have limited control, and we don't really know what our control does so much. How do we steer things in a better direction? And that is the question. I mean, that was interested. It seems to me we don't need to just it's not satisfying to say we're just going to wait and see why where the cards fall, and wait and see if a meteor lands on your

head or not. I mean, you kind of want to know whether you're government actually is willing to give you an umbrella or something stronger to prevent the meteor hitting knocking you out. What I sometimes think is when I look at the US, but it doesn't seem like there's a much appetite or capacity for government or for other institutions to help people manage the human consequences of this

kind of dislocation. I mean, we've got I was looking at some of the numbers the spending on public spending on sort of training and labor market stuff as one of the lower in the O E c D. It's gone down by as a share of GDP since the eighties in the US trade unions, who might have been the kind of natural institution that could have sort of stood between kind of workers and companies and helped people up skill. You know, they're clearly not and on the

horizon at all in the US now. So what do you when you think about what tools government has, even with all this uncertainty about the impact, You know, are there things if you hadn't if you had a different kind of appetite political appetite, are the things that you would say would make a real difference. Absolutely, Now we

don't really know what will work. By the way, we tried a lot of government retraining programs in previous decades and all the evaluations say that they were completely ineffectual. They didn't help anybody get jobs in long term. They were a giant waste, and that's why we cut back. The government was terrible retraining people. UM companies are really good at training people because they know what to train

people for, they know what they need. The thing is, you know, American companies have just been in this orgy of offshoring and outsourcing, especially outsourcing, so offshoring isn't as big of an issue as people think it's it's more outsourcing to contractors, you know, within the United States. Because of that, companies don't spend a lot of money and time training people anymore, and you get this equilibrium where it's very easy for people to jump to other companies,

unlike in a lot of other countries. But then because it's so easy, companies don't want to spend the time and effort training workers. And so we need to shift back to this equilibrium where companies give people more job security and and more training and more investment in the workers, and um, we need to have a more harmonious labor management relationship. And I think the most interesting proposal for how to do that has come from Elizabeth Warren, who

proposed codetermination. I don't know if they're familiar with this, but it's yeah, so it's I mean, Germany does it. It's it's corporatist. It's not the sort of socialist model of the benevolent, all knowing government will do everything for you,

but it is. It definitely makes companies more democratic and give work, gives workers more of a saying company policy, and of course workers are going to want more investment in workers and her proposal didn't have the sort of workers councils that Germany has, but I think that would be a good addition as well. So I think that policy has the largest chance of anything that we've thought of are proposed of shifting back to this good equilibrium

where companies invest a ton in training workers. But that policy is a million miles away. I mean, although it's been proposed by a US politician, it is a long way from the the culture that we've seen of American companies and American capitalism really for for many, many years. I mean, there's not that the few bits of worker code inmidentation that I think you did have, or we started to have in the US, I think in the thirties got um overwhelmed by efforts against it and never

really got established in the US. I mean your it is certainly true. There's a lot of this kind of thing in Europe. Jeremy Corbyn and the Labor Party has has talked about it. I was out of something where there was a lot of people from different parts of Europe talking about the challenge of technology and digitalization of the workforce, and the Swedish trade unionists said, you know, we've with our members. We don't think that the new machine opposes a threat to our members. It's the old

machine that poses a threat. And it's it's a completely different psychology, which comes from the social partnership they have and the kind of funds that they've come together. Employeed and employers contribute to retraining programs that, as you say, are closer to the work. But it feels just culturally, it's not just the current administration. It feels a million miles away from the from the traditions of the US. Is it really something that could easily start to be

effective in the US? Uh So, I don't know, um, but I do know that it's closer to our culture and traditions than a government managed thing, and it would be more effective as well. We have a history of doing corporatest things. We have a very corporatist health insurance system. We say, you know, in America, if you're not old or you know, your your health insurance goes through your employer. So we have a history of having things go through

people's employers. I think there's much more hope for that kind of of corporatist you know, retraining and worker investment system. Then there is for a government system, even before you take into account the fact that the government system would probably be ineffectual and sale. So, Um, the answer is is it a long shot? Maybe so, I don't really know. Uh. Is it a longer shot than other things? Probably not?

And if a if a third option, if a third option is to somehow have you know, because we can't do any of these other things, we should instead have the government try to disincentivized adoption of automation technology. That would be dumb and a disaster. Um. We should not. That is what we should absolutely not ever do, and we should firmly resist anyone who brings that idea up.

And the reason is because our companies are competing with companies from Asia and Europe that are not going to do that, that are going to do something a lot smarter than that. If we simply limit technology, this one technology because we're sort of afraid of it, and we spin these sci fi scenarios that haven't come true at all yet, Um, then we'll just be, you know, turning

ourselves into a backwater while other countries take over. We are going to have some interesting conversations about this with all the movers and Shakers and the New Economy Forum in Singapore. No Smith, thanks very much, Thank you, thanks for listening to The New Economy. Today's episode was reported by Quaig Toras with editor Scott Lamon and produced by Magnus Hendrison. Special thanks to Noah Smith. Francesco Levy is the head of Bloomberg Podcast h