September 4, 2023 • 49 mins
Are robots and automated systems going to replace human workers? And if so, what then? The fear about automation has a long history and it's not an irrational fear.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:04):
Welcome to Tech Stuff, a production from iHeartRadio. Hey there,
and welcome to tech Stuff. I'm your host, Jonathan Strickland.
I'm an executive producer with iHeartRadio. And how the tech
are you? Here in the United States, we are celebrating
Labor Day, a federal holiday here in the States. Also
(00:26):
one that I find really interesting because it's all about
celebrating the American labor movement. But if you pay attention,
especially in the tech space, there are a lot of
entities out there, a lot of companies that are eagerly
opposing the labor movement and trying to do things like
prevent workers from organizing and forming unions. Yeah, a complicated
(00:51):
thing that we have a holiday to celebrate it, and
yet we have plenty of examples of companies and organizations
out there dedicated to venting more labor organization from happening.
But let's put aside all of that. I thought, since
we're on holiday today, I wanted to make sure that
you had an episode anyway. So we're going to actually
(01:13):
listen to an episode that came out a few years ago,
back in twenty twenty July thirteenth, twenty twenty. In fact,
it is titled The Robots Are Coming for Your Job,
and it's all about robots and automated systems and the
anxiety that exists around this idea of automation eliminating jobs,
(01:33):
which I think has only become even more of a
talking point in the wake of things like generative AI,
for example. So let's listen to this episode from twenty twenty,
the Robots Are Coming for Your Job, and I'll chat
with you again at the end. I read a lot
of tech news and sometimes that ends up inspiring me
(01:56):
to do an episode of tech stuff. That happened to
me recently when I read this headline off of the
website tech Spot. Sony factory assembles PS four in thirty seconds,
only four humans involved in the process. A PS four,
in case you're not aware, is a PlayStation four video
(02:19):
game consoles. So this factory can build a video game
console from parts in half a minute and only four
human beings touch the ding dang thing in the process.
Those four humans, by the way, are involved in the
beginning and the end of the process. Two of them
load motherboards onto the assembly line, and a motherboard is
(02:41):
the primary circuit board for a computer system, and the
other two human beings are at the end of the
assembly line, and their job is to package the completed consoles.
All the actual assembly work is done by robots. Now
you may be experiencing a couple of different responses to
this information. I know I did. One of those was
(03:02):
a wow, that's seriously impressive. The PS four, like many
computer systems, has a lot of components, many of which
attached to one another by wire or cable. So these
robots have to be able to take these flexible components
and to join them in their proper anchor points with
the appropriate amount of pressure and precision to make a
(03:24):
good connection. Now, if any of you out there have
ever built your own PC, you know that plugging cables
in can get a little tricky depending on the layout
of the motherboard and the various components. And if you're
someone like me, you're likely putting stuff together only to
realize that maybe you should have done some of that
before you mounted them in a computer case, because now
(03:46):
you just don't have the space to work in properly.
So it's pretty darn impressive that robots can do this
consistently and correctly at that level of speed. Another response
I had was kind of scary. I mean, typically you
would have dozens of people employed on the assembly line
to do this sort of work, but in this factory
(04:09):
it's been stripped down to thirty two robots and four
human beings. The article in textpot points out that twenty
six of those thirty two robots are just attaching flexible
components together inside the console. Now, I have no idea
how much these robots cost, but I wager that they
(04:30):
are expensive enough to equal the salary of a standard
human employee on the assembly line. However, you don't pay robots.
You do have to spend money to maintain and repair them,
but assuming whatever you're making is going to be around
for a little while, they'll pay for themselves because eventually
you'll get to a point where the salaries you'd be
(04:51):
paying for humans would be more than the purchase and
maintenance cost of the robots. And the increase in efficiency
means you can produce a whole lot more stuff in
a given amount of time than you would with a
human centric assembly line, so you'll have more product to
sell in a shorter amount of time. When you start
(05:14):
crunching numbers, you discover your robotic assembly line can make
more stuff at a lower cost over a given period
of time, like you know, over a couple of years,
than what you would accomplish with human beings on that
assembly line. So you don't have to worry about the
robots taking a vacation. They don't take sick time, they
don't even take the night off. They can work around
(05:35):
the clock. They don't need health insurance, though I would
guess that most companies ensure the heck out of these
things just in case one breaks down. But from a
financial point of view, they make sense if you're building
stuff at a large enough scale, stuff like video game
consoles for the PlayStation four. It's a no brainer because
(05:55):
that console has sold around one hundred and ten million
units so far. That's a number large enough that I
can't even imagine what it would look like if you
had all those consoles together in one place. So if
there's enough demand for you to sell one hundred and
ten million whatever it is you want to sell, you
(06:16):
need to have a way to make those as efficiently
as possible, and that will help maximize your profits. And
the more efficient the process, the more competitively you can
price your product and still make a profit. But the
idea of robots performing jobs far more effectively, consistently, and
efficiently than humans raises a lot of questions. And these
(06:38):
are not new questions either, but they are questions like
if more factories rely on robots for production, particularly if
those robots can be programmed to produce new products once
older ones go obsolete, what happens to the job market,
What happens to the millions of people who work in
manufacturing on assembly lines? They go? What will this do
(07:02):
to economies around the world. Lots of people have tried
to answer these questions, sometimes giving drastically different answers. And
we're going to take a look at the history and
evolution of industrial robots in this episode and explore the
ramifications of automated manufacturing. And this is where I dive
(07:22):
into history. I've talked about the history of robots before,
so I'll try to restrict my focus to industrial robots.
And before I get into that, let's just address the
fact that the use of machinery to increase efficiency has
been a controversial subject since long before there ever was
such a thing as a robot. Generally speaking, machines are
(07:45):
meant to make work easier or in some cases make
the work possible just to begin with, they are labor
saving devices, requiring humans to put forth less effort to
get the same or better results. This applies to the
simplest of machines, I mean stuff like levers or pulleys
or an inclined plane, and it applies to very complex
(08:09):
machines as well. Before the Industrial Revolution, most stuff like
textiles was made by crafts people out of their own homes.
This was literally the cottage industry. Trades people would travel
and become the lifeline for the cottage industry, supplying raw materials,
buying finished products, and selling those products off at a
(08:31):
profit elsewhere. Many tradespeople built a good deal of wealth
working this way, and they had the means to look
at alternatives to this decentralized cottage industry approach. An idea
began to form. If you brought together crafts people to
a centralized location, and if you simplified the process of production,
(08:54):
you could make way more stuff, which in turn means
you could sell way more stuff, which in turn means
you can make way more money, and money makes the
world go round. This thought process helped fuel a similar
line of thinking. If you could design machines that could
(09:15):
do a lot of work that typically fell to skilled
crafts people. You wouldn't need the crafts people at all.
You could train anyone, even if that person had no
experience with the process, just to work the machine. And
while it might take years of dedication to go through
the process of being an apprentice to learn a trade
(09:35):
well enough so that you can actually make a living
at it, with a machine, you can skip right over that.
As long as the machine's end product was good enough.
It didn't have to be better than the stuff crafts
people were making. It just had to be good enough
and cheap enough and fast enough to produce. Then you
(09:55):
could sell the finished product at a lower cost than
what crafts people would charge because not as much time
and effort went into making the thing. Now, I guess
it's clunky to talk about this while using an example,
so let's go with a poster child for the Industrial Revolution, weaving.
The weaving trade is an ancient one, and it requires
(10:16):
a good deal of skill to do it well by hand.
In the late seventeen hundreds, a man named Edmund Cartwright
patented a loom powered by a water wheel. The loom's
operation was such that a person who had no training
in weaving could operate the machine and produce finished textiles.
Cartwright's design would be built upon by other inventors who
(10:39):
had turned to steam power and other means to operate
the power loom. Many cottage industry weavers found themselves out
of work. They could potentially opt to work in the
textile factories, as those were popping up all over the place,
particularly in England, but the wages were low. As you
can imagine, this didn't sit well with the weavers. There
(11:04):
were protests, including some that incorporated violence and destruction. Ultimately,
the factory process won out, and along with it some
really awful working conditions followed, including stuff like child labor
and ridiculously low wages and dangerous working conditions. This led
(11:25):
to more protests, including the type that would give us
the word sabotage. And let's get a quick side note
on that one, as it is the source of a
little mythology or misinformation. See The apocryphal story goes that
the word sabotage comes from the word sabo, which describes
(11:45):
the wooden shoes worn by laborers, mainly Dutch laborers, but
also laborers in France. And according to the story, these
laborers wore those shoes and used them to great effect.
In an effort to protest the conditions and factories, they
would toss their wooden shoes into the machinery to break
(12:07):
the various gears and literally grind production to a halt,
as it were. But this story, while compelling, isn't really
the truth. Sabotage does stem from the word sabo, but
in French there is a verb sabote. This verb means
to make a loud noise with wooden shoes. Now isn't
(12:29):
it great that there's a verb for that? And it
makes sense wooden shoes would make a great deal of
racket as people would walk around. Heck, if a toddler
wore wooden shoes, I think it would probably sound as
though the world were shaking apart. I don't know how
toddlers manage to sound like they weigh eight hundred pounds,
(12:50):
but they do it. And if you have a toddler,
you know what I'm talking about. And in the culture
of France, the idea of a clumsy slow worker was
often linked to someone who wore wooden shoes because they're
awkward to wear anyway, The reason sabote led to sabotage
is because factory workers who were protesting their work conditions
(13:14):
and wages would purposefully work more slowly and less efficiently
in order to affect the overall output of a factory.
It was related to a similar strategy that British laborers employed,
and their version was called kakani. It was a saying
from Scotland which essentially means don't do so much man now.
(13:34):
I would argue this also feeds into a strategy that
we see to this very day in certain government offices,
where the idea is there's no need to do too
much too quickly, as it doesn't result in increased compensation
and it also sets a really high bar of expectation,
so why not just take it easy? You know, I
don't have a coffee break now. In the early twentieth century,
(13:55):
people began to use the word sabotage to really refer
to a purpose full approach to undermining the output of factories,
and it had nothing to do with tossing wooden shoes
into machinery, though it did also pertain to instances in
which workers purposefully damaged equipment and tried to slow down
(14:16):
the production that way. While this isn't directly tied to
the idea that machines themselves are displacing workers, it is
related to the effect of moving toward a manufacturing based
economy and how that allows for the exploitation of workers.
The machines themselves aren't really at fault, but they facilitate
(14:36):
the system of operations that leads to exploitation. Now that's
something that'll be a theme in this episode, and we
can't ignore the social aspect of what's going on here,
or else we miss the whole point. But let's skip ahead.
I've spoken about this before, but we get the word
robot from a check author named Kyl Kopek. He wrote
(14:59):
a pl called Rossum's Universal Robots or R you Are
in nineteen twenty. Kopek took an older word rabota, which
means forced labor in Europe. This concept was tied to
that of the old system of serfdom, in which people
would do work on behalf of a landowner. In return,
(15:21):
those people would be allowed to live on part of
that landowner's land. In ru are factory owners devise a
way to build laborers from raw materials. Now in the play,
they are indistinguishable from humans other than they have no
inner desires. But in the course of the play, these
(15:43):
laborers eventually take over all the jobs that humans previously held,
and humans themselves become a threatened species as these laborers
begin to understand the power that they hold by occupying
all the positions of employment, including as soldiers in the military.
And so with the introduction of the concept of robot,
(16:04):
we actually get the very first robotic uprising all the
way back in nineteen twenty. See I told you it
was an old idea. It's important to remember that in
the play, the robots are nearly identical to humans. They
aren't mechanical the way our robots of today are, but
the idea of creating machines that can do work without
(16:25):
a will of their own is a part of robotics
in general, and industrial robotics in particular. When we come back,
we'll talk about the earliest industrial robots and what they did,
But first let's take a quick break. It's interesting to
(16:49):
me that the tech world adopted the term robot when
we think about the origins of that word. In Cope's work,
robots were sentient slaves. They could perform the work humans
would otherwise do, but they lack the emotions that humans have,
and the whole idea is that these devices could do
(17:10):
our work for us without question or protest. They would,
in theory, endure conditions that people wouldn't or couldn't, But
in the play they ultimately lead to the destruction of
the human race and potentially they become the new dominant
species on the planet. Now. I say potentially because part
of the play's plot involves the destruction of the formula
(17:34):
that scientists use to produce the robots in the first place.
That is an important plot point. The robots are not
sure how to make more robots, so they might just
die out. Now, it seems to me as though that's
a pretty emotionally charged term to adopt for an entire
discipline of technology, right, robots, Especially if you are actually
(17:57):
aware of that play, and by the way, I recommend
people read it. It's a good play. But then a
lot of people are not aware of the origins of
the word, or at least not beyond knowing that it
came from a play in the nineteen twenty So I
guess for them, it's just, you know, a word. A
robot by any other name would smell as sweet as
(18:18):
it were. And we've definitely seen the themes of rure
serving as an undercurrent for stuff that's happening in robotics
in general. But let's move ahead. In nineteen fifty four,
an engineer named George Duvault designed an industrial robot. He
was nine years old when Kopek coined the term robot.
(18:39):
He called his design the Programmed Article Transferred Device, for
which he received a US patent in nineteen sixty one.
This machine was a robotic arm, and it was capable
of picking up something and then transferring it a short
distance away, just within reach of the arm. The arm
itself couldn't move, it was anchored in place. It could
(19:02):
also follow. In fact, this is the important part, It
would follow a pre program series of instructions to do this.
Deval's argument for his device was that up to this point,
mechanical handling of objects fell into two broad categories. Either
stuff got moved by humans, typically operating a powerful machine
like a crane or a forklift, or stuff got moved
(19:26):
by a device that operated under cam control. Now, manual
control is self explanatory, so let's talk about cams. A
cam is a rotating component in machinery. Typically a cam
has some variation in its surface. So let's start with
a wheel. Just imagine a wheel that is spinning on
an axle. Well, you wouldn't typically have a perfectly smooth
(19:50):
wheel as a cam. Part of that surface might be flat,
or it might have dips in it, And when the
cam rotates, these variations apply force to some other mechanical
component that is held against the cam, and it causes
that mechanical component to move in specific ways. A cam
operating system can work on its own, but it will
(20:13):
always repeat the exact same motions As long as everything
is working, it'll just repeat those steps. Once the cams
complete a full systematic rotation, you can't really adapt it
to do anything else. The movements depend entirely on the
cams themselves, so if you wanted it to do something else,
you would at first have to swap out the cams,
(20:35):
and even then you would be under whatever the limitations
of the device was itself, like it wouldn't have full
range of motion. Moreover, this level of specialization also means
that it's typically really expensive to rely upon cam based systems,
so it was really only useful if the application had
to do with mass manufacturing or else you're looking at
(20:56):
an economic loss. The cost of the system was just
too much. So Devaal was proposing a machine that could
be programmed to do operations. And this would let a
programmer create different processes using the same machine. Or you
could get a whole bunch of the same basic machine
and program each one to do a particular job. Meanwhile,
(21:18):
you'd free people up to work on other stuff in
the manufacturing process, and you could take the most dangerous
stuff and give it to the robots. Now, the story
goes that Daval was at a party in nineteen fifty
six and he got into a conversation with a man
named Joseph Engelberger. Joseph was a scientist and an entrepreneur,
and when the subject turned to Deval's programmed article transferred device,
(21:41):
as well as the work of a science fiction author
known as Isaac Asimov, you know, the father of robotics.
He famously incorporated a concept of the laws of robotics
in his works. We won't really go into that in
this episode, but the laws of robotics still play a
big part in the discipline of robotics in general. But
(22:02):
it's kind of outside the focus of this episode. Ingelberger
used his connections to get funding for devauld to create
a more advanced version of the programmed article transfer machine,
and it would be a robotic arm capable of making repeated,
precise movements while holding very heavy objects. They called it
the Unimate Unimate, and the first prototype, Unimate zero zero one,
(22:29):
would go to General Motors to work on a die
casting assembly line. Now, according to the company robot Works,
that's a WRX. This robot cost around sixty five thousand
dollars to produce, and Ingelberger sold it off at a
tremendous loss. General Motors only paid eighteen thousand dollars for
(22:50):
a sixty five thousand dollars machine. But Ingelberger really wanted
to establish that robotics were a way to perform repetitive,
dangerous functions at a lower risk to humans. Welding die
cast components on autobodies was a great first application of
industrial robots for a few reasons. Die Casting is a
(23:10):
process involving molten metal. You take that molten metal and
you force it into steel molds, and these are water
called dyes. The molten metal cools in the exact shape
of the mold. So this is a way to make
or cast a bunch of identical parts out of metal
and get consistent quality out of it rather than forging
(23:33):
each piece and then fitting them together. A dye can
have complex shapes in it, such as external threads, which
means you don't have to make a pipe, for example,
and then do a secondary process on that pipe to
get the result you want, So you wouldn't have to
carve those threads into a otherwise smooth pipe. You could
(23:54):
just cast the pipe with the threads incorporated on it already.
But welding die cast parts onto autobodies is hard work.
The components are really heavy, so you're at risk of
immediate injury if something goes wrong, like let's say you
drop a weighty component on your foot, or you might
develop a repetitive stress injury after going through the same
(24:17):
welding motions over and over again. In addition, the fumes
given off while welding were sometimes toxic still are, so
it's not great to have people exposed to them for
very long. So a robot was a great substitute for
a person. The robot could handle much greater weight than
people could The robot didn't breathe, so there was no
(24:37):
respiratory issue there, and it didn't get tired. I mean,
it would wear down over time, but you could repair
it in fairly short order. The Unimate worked with computer
controlled hydraulic systems. A hydraulic system uses a liquid that's
under pressure in order to do work like pushing against
a piston to power an actuator of some sort like
(24:58):
lift a platform. The UNIMATE zero zero one weighed twenty
seven hundred pounds or about one two hundred and twenty
five kilograms, and it could work twenty four hours a day,
placing components with a precision of within one fifty thousandth
of an inch. Now I'm not going to do the
conversion on that, because I think it's sufficient to say
(25:20):
that it was just really precise. According to a charmingly
dated newsreel from Britain, complete with swinging sixties music that
sounded like it came straight off an Austin Powers movie,
the robot could operate for five hundred hours without the
need for a human to check in on it. Engelberger,
a savvy businessman and promoter, would arrange for Unimate to
(25:43):
show what it could do at trade shows and on
TV appearances, including one on The Tonight Show with Johnny Carson.
If you don't know who that is, ask your parents,
and if they don't know, ask your grandparents. Nineteen sixty nine,
General Motors had jumped on board the robot train, as
(26:05):
it were. They rebuilt a manufacturing plant in Lordstown, Ohio,
and they installed unimate robots to perform spot welding on
car bodies, and the results spoke for themselves. The plant
was capable of producing one hundred and ten cars per hour,
which was more than double the speed that the plant
could manage before the installation of the robots. The business
(26:29):
case for the robots seemed clear. After a hefty upfront cost,
you could produce way more stuff per day, and as
long as the demand for that stuff is high enough,
it could mean greater revenue. You could also bring the
cost of production for an individual unit down. Then you
could pass savings on to customers and get really competitive
(26:49):
with your pricing, or you could just keep everything priced
the same and try to increase your profit margin. The
key to all this was that you had to be
sure the thing you were producing would bring in enough
money to offset the cost of automation, so it would
not make sense to spend millions of dollars building out
a factory staffed with robots. If you were making something
(27:10):
that had a very small market to begin with, yes,
you'd be able to produce way more Watching My colle
its than you could before. But if the demand for
Watchma collets is really modest, that doesn't do you any good.
In fact, you might end up flooding the market and
devaluing your product. So while robots were taking on jobs
(27:31):
that were previously held by humans, there was no real
danger of a massive upheaval where everything would be automated.
The limitations in the technology were just too great and
the cost was too high for most companies to go
that route. And this also became the starting point for
something that would become really important that the main goal
of developing industrial robots wasn't to displace humans. It was
(27:55):
meant to offload duties that were dull, dirty, or days.
You'll often hear those terms being used with robotics if
it is a job that carries with it a significant
risk to the person performing it, or a job so
demanding that you can only expect a person to stick
with it for a short while before they need to
do something else. Then building a robot to do that job,
(28:17):
or at least that list of tasks makes sense. The
robot is just a thing. It can endure conditions that
humans can't, and it doesn't get sick, and it doesn't
get hurt. If something breaks down, you can typically repair
it pretty quickly. We humans don't have that luxury. Now,
I'm not going to go and run down a full
(28:38):
history of all industrial robots because that would mostly involve
me talking about model numbers with slight differences like the
number of axes of movement or points of articulation for
one robot versus another, and that's not really interesting. But
I do want to hit a couple of highlights. One
is that in nineteen seventy five, the ASEAB robot would
(29:02):
be the first fully electrically driven robot. It also used
Intel's first chip set as processors. Now, this was not
a super strong robot because those electrically driven limbs just
can't pack the same punch as a hydraulic system, which
typically moves much more slowly but can handle much heavier payloads.
(29:24):
So this particular robot could only lift weights up to
around thirteen pounds or six kilograms, but the move toward
processors and electrically driven components marked a big technological step,
even if the arm's physical capabilities were much less impressive
than a hydraulic system. By the end of the nineteen seventies,
(29:47):
Japan was getting into the robotics game with arc welding
robots for assembly lines, and then it was off to
the robotic races, with the eighties seeing a surge in
advances with industrial robots. Soon, massive manufacturing facilities were installing
robots to take over elements of the assembly line process,
(30:08):
particularly in that dirty, dull, and dangerous category. The robots
became more sophisticated, which also added to their value. When
we come back, I'll talk more about why that's important,
but first let's take another quick break. By the mid
(30:32):
nineteen nineties, robotics companies were making machines that could coordinate
and synchronize the movements of more than one robot at
the same time, allowing for more complex manufacturing processes. By
the early two thousands, there were systems that could synchronize
the actions of up to four robots at a time,
further adding to the overall system flexibility. Now, I mentioned
(30:54):
earlier that a programmable robot is more versatile than something
like a cam operated system. Well, more sophisticated robots, with
more axes of motion and more points of articulation have
the potential to do lots of different types of jobs,
and this is of critical importance. If the robot is
(31:15):
too limited, if you can only do a small range
of motions, you can't necessarily repurpose it for new processes.
And as markets change, you may find yourself needing to
be flexible when it comes to the stuff you're manufacturing.
So let's use an extreme hypothetical example that would probably
never happen. So let's say that you run an auto
(31:38):
manufacturing facility, but then there's a massive market change and
it drastically affects the demand for your cars. There's just
not enough demand to support the production. So rather than
just you know, closing up shop and calling it a day,
your business decides to do an amazing pivot and you
(31:59):
begin to convert your manufacturing facility over to I don't know,
home appliances. Now, again, this is an extreme hypothetical example,
but let's just go with it. Okay, So here we go.
If the robots on your assembly line are powerful, but
limited in movement and function. You may find it impossible
to adapt them to your new line of business, which
(32:21):
would mean you need to either invest in new robots
or you'd have to hire human workers to put together
your appliances. And it would also mean that your old
robots would be a sunk cost. You would need to
either sell them off or put them in storage or something.
If the robots are really sophisticated, however, you might be
able to program them to do some of the operations
(32:43):
on the new assembly line, and that would keep them useful,
it would lower the cost of production. Or, for a
less extreme example, you introduce a new model of whatever
thing it is that you're producing. Anything new will require
adjustments in the assembly line process, and if the changes
are big enough, the robots may not be able to
(33:05):
make as big a contribution in the process. That's something
that could happen with the example of the PlayStation we
were talking about. Yeah, those robots can put together a
PS four and thirty seconds. There's no guarantee they'll be
able to do the same thing with a PS five,
at least not without a major overhaul of their assembly line. System.
While the manufacturing facility can churn out a finished PS
(33:27):
four and thirty seconds, we might not see them work
at all with PS five, at least not right away.
It would all have to be optimized. So for decades,
industrial robots were kept as separate from human workers as
was possible. You wanted to keep them well away from
all the people, or keep the people well away from
all the robots. Often the robots would operate within cages
(33:50):
specifically to limit the possibility of a human coming within range.
After all, these robots are large, they're heavy, powerful, and
many of them are incapable of sensing stuff in their
environment and whether or not a human is within their
range of motion. Instead, they're just going through that pre
(34:11):
programmed series of motions and they're not going to stop
unless someone turns it off. A robot is performing that
same series of steps over and over, and that can
mean that if a human in that area gets near
the robot, they could end up getting injured or worse.
And in fact, this has happened several times over the
course of the last few decades, and at least in
(34:32):
some cases it seems as though the robot might have
been at fault meaning it's not always a case of
human carelessness. For example, an engineer in twenty fifteen died
when a robot armed from one section of the factory
floor moved beyond its operating area and into the neighboring
section that the engineer was working in. This is something
(34:55):
that should not have happened. The robot arms should not
have moved that far into the neighboring section. The robot
arm hit the engineer on the head, and she later
died from her injuries. In the United States, the government
has listed thirty three workplace deaths due to accidents with
industrial robots between the years nineteen eighty four and twenty fourteen.
(35:18):
The investigations also found that the majority of those tragedies
was typically the fault of human error. There was a
person who was wandering into the operation zone of a robot.
That twenty fifteen incident was an outlier. Not that any
of this makes the thought of working around industrial robots
(35:39):
less scary or those other accidents any less tragic. They're
all terribly tragic. Moreover, we're seeing more robots that are
capable of roaming a work space. They are no longer
anchored to a specific spot on the floor. In some cases,
they also, unlike the first industrial robots, typically have external sense.
(36:00):
These not only help the robots navigate their environments, but
also hopefully avoid accidents with human workers. Let's take Amazon's
warehouse robots for example. These robots look like really big
robotic vacuum cleaners. They are designed to roll under shelves
and the shelves are just slightly larger than the dimensions
(36:23):
of the robot. And when an order comes in, a
robot from the warehouse rolls over to a shelf that
holds the respective item on it according to the inventory system,
and the robot goes under the shelf then lifts the
shelf by raising a platter like platform on the top
of the robot. Think of it as like a little
fork left, except it's more like a I don't know,
(36:45):
like a tray that a waiter would use to carry
drinks to a table. But it carries the whole shelf
up and over to the edge of a cage, where
a human operator will take the respective item off the
shelf and scan it and put it into a bin.
And then those bins go to other humans who further
scan those items and then put them into other bins
(37:06):
that ultimately go to the packing department. And if you
watch videos of these robots, it looks like they're doing
a complicated ballet as they maneuver through this warehouse, avoiding
other robots and shelves. As they bring those shelves to humans.
Markings on the warehouse floor tell the robots where they
are with respect to everything else in the warehouse, and
(37:29):
the robots even will position shelves that have items that
are being ordered a lot toward the edges of this
space so that they're easier to get to and move
them over to the human beings. So it's kind of
an interesting dynamic system. It's not like they pick up
the shelf and then bring the shelf immediately right back
to where it started. The shelf can end up in
(37:50):
a different location entirely. In addition, cameras give the robots
the ability to sense any obstacles that might block their path,
allowing the robot to come to a stop and a
wait further instructions and report that it has found something
unusual on the warehouse floor. Even so, typically humans are
(38:11):
not allowed to roam the area where the robots pick
up shelves if something has fallen on the warehouse floor.
A designated troubleshooter gets an alert, and that person must
use an interface to draw the path that they are
going to take from the entrance of the cage all
the way to the point of trouble. Like let's say
(38:32):
that a product has fallen out of a shelf and
has hit the floor, and a robot has reported it.
You would use a tablet. If you're the troubleshooter, you'd
use a tablet and you would draw, almost like a maze,
the path you would take to get to that particular item,
and you would follow that path out and back. In addition,
you'd wear a radio transmitter that would send a signal
(38:55):
out that the robots could all detect, and that would
alert the robots to the presence of you, the troubleshooter.
That helps prevent a situation in which the robots are
going to collide with you, right, you want to avoid that. Now,
there's a lot of work that goes into designing robots
that can interoperate in a space that's occupied by humans,
(39:15):
and it's a very challenging line of technology because it
takes more than just thinking about how the machines work
you also have to think about how people work, and moreover,
you have to think about how people change the way
they work when they're in the company of a robot.
It's kind of similar to the concept in quantum theory, right,
(39:35):
the idea that you change a thing you observe just
through the act of observation. Well, you can have a
workspace that humans had only been working in for a while,
and you could say, all right, well, I've observed how
the humans work, and I'm going to build a robot
that does this one task that the humans do, and
I'm just going to incorporate right into their workspace. But
(39:58):
then you find out that when you do that, the
humans all behave in a new way because there's a
new thing in the environment that you didn't account for,
and now the design of your robot doesn't work as well.
We humans are tricky like that. Moreover, we need to
get to that threat that weavers were worried about more
than a century ago. Is automation going to take our jobs? Now?
(40:22):
There have been a few studies, all using different methodologies,
and some of those studies coming under criticism for the
approaches that were used. But there have been a few
studies that suggest we'll see automation continue to impact jobs
in the near future and drastically so over the course
of the long run. The interpretation of those results have
(40:43):
been reported in ways that range from automation is going
to be disruptive that's on the light end to fifty
percent of all jobs are going to be taken by
the robots. So what's the actual truth. Well, the truth,
as it turns out, is complex. For one thing, automation
rarely takes over an entire job. What is far more
(41:06):
likely to happen is that automation will take over certain
tasks that are part of a job, or perhaps multiple jobs.
So if a job requires a wide variety of tasks,
some of which may require critical thinking, it's really hard
to design a robot that can do all of that.
It's far more likely that you would automate certain job responsibilities,
(41:30):
which would mean that those jobs themselves wouldn't go away,
they would just change. The repetitive responsibilities would be offloaded,
and you would focus on something else. You might have
to spend more time doing other duties rather than the
routine ones, which isn't necessarily a bad thing, but there
are cases where automation would likely take over an entire job,
(41:55):
for example, truck drivers in shipping trucks. Much of the
work in autonomous vehicles is really focusing not necessarily on
replacing passenger vehicles so much as commercial vehicles like shipping trucks.
The Bureau of Labor Statistics in the United States estimated
(42:15):
that the age of the average US truck driver is
fifty five and more than ninety percent of all truck
drivers in the US are mail and that will present
a challenge see Generally, the pro argument for automation is
that while robots and automated systems will eliminate some jobs,
(42:36):
they will create other jobs, presumably better jobs. And this
is true. At the turn of the twentieth century, forty
percent of all jobs in the United States were on farms.
So that means four out of ten people in the
US who had a job we're working on a farm. Today,
(42:57):
Agriculture and all the related food sectors make up just
eleven percent of all jobs in the United States. And
if we just limit this to the people who are
working on farms, you know, not all agricultural jobs and
food sector jobs, just the farm jobs. If we do that,
we're talking about only one point three percent of all
(43:17):
US employment, So going from forty percent to one point
three percent, that's a drastic change. Now, clearly automation has
transformed agriculture. It allows us to do a lot more
while relying on fewer people, and new jobs did come around,
so we didn't see an unemployment rate reaching levels higher
(43:38):
than forty percent pre COVID. The pro automation argument states
that new jobs, which again should ideally be better than
existing jobs, as in less strenuous and less dangerous and
more interesting, will emerge as older jobs are phased out.
Now that works fine on a macro scale when you're
(44:01):
taking a really big picture look at the overall trends,
but when you consider the particulars, like our truck drivers,
you start to see some obstacles. See this year, I
turned forty five, so I'm a lot closer to the
average age of a truck driver in the United States
than I am to someone who's just getting into the
(44:21):
job market for the first time. And I can tell
you that even as a relatively tech savvy guy, I
would find it really challenging to pick up the job skills.
I would need to go into a different line of work,
particularly one where I'm competing against people who already have
training and experience in that field. So imagine having to
(44:44):
tell a group of fifty five year old truck drivers
that they're out of a job. But good news, if
you just start taking classes, you can learn to code
and make less money than you did in your old job.
It's not great, is what I'm saying. That being said,
automation is clearly not going anywhere. It's going to continue
to play a big role in how we get work done,
(45:07):
and in our best case scenarios, it's going to augment
the work that humans do, leading to better, more efficient,
and more cost effective outcomes. It will free us up
to focus on the parts of our jobs that we
find the most fulfilling. We can handle the stuff that
requires flexibility and intuitive thinking, and the machines can handle
(45:27):
the routine and the dangerous. But in a worst case scenario,
we'll see an unprepared population of former workers who are
now out of a job and without the support system
there to help them transition into something new so that
they can continue to contribute to society and earn a living. Now,
(45:49):
this is why you will often hear conversations about automation
get tied into concepts like a guaranteed jobs program. This
is typically where something like a government creates a system
that makes certain every person who wants a job can
get a job. Or you'll hear about guaranteed basic income.
(46:10):
This is a strategy in which tax dollars go to
fund a standard income payout to all citizens so that
they can meet their most basic needs. Now, these are
big ideas, they aren't easy to implement or administer, and
they're not cheap, But it may be that they will
become necessary or some similar strategy will be needed to
(46:33):
make certain that we have a plan to move toward
rather than being caught in a world where a disproportionate
percentage of people can't find gainful employment. Heck, we're seeing
something like that right now due to the COVID crisis,
which is also underlining the importance of automation in a
world where it's not necessarily safe to have a bunch
(46:54):
of human beings all gathered in the same place at
the same time. Robots coming for our jobs, well, for
some of our jobs, definitely, many of those jobs come
with some pretty tough consequences for humans who are working
those jobs today. Those jobs may have high injury rates,
the people who work them may have lower life expectancies,
(47:17):
and they are a whole host of health issues that
can come along with certain jobs. So you could make
a strong argument that really this is for the best
because it will help save lives and reduce the chance
for injury or illness for a lot of people. But
for other jobs, the robots aren't likely to take over
in the near future. For a lot of jobs, automated systems,
(47:38):
not necessarily robots, but perhaps software based AI, will augment
what humans are doing. It's important we have conversations about
this stuff and to talk about how to address the
consequences of increased automation. There are ways we can enjoy
the benefits of automation, but only if we think critics
(48:00):
about it and create policies and procedures accordingly. Now I
gotta get going. I hear robo Jonathan is going to
host the next episode of Tech Stuff, and I have
to train them on how to make puns and pop
culture references. I hope you liked that episode from twenty twenty.
As I said, We've got a lot more to talk
(48:21):
about these days because of things like generative AI and
the concern that that could impact jobs that for a
long time people assumed we're safe from automation, especially from
things like robotics, you know, white collar jobs that people
just thought were kind of the domain of humans, And
now there's a real question as to whether or not
that's actually the case. So I think it's even more prevalent.
(48:44):
And again seeing the labor movement in the tech sector
in particular over the last few years tells us that
there's some very important issues still at the very heart
of technology and the way we do business that relate
back to the foundations of the labor movement here in America.
(49:04):
So I hope you enjoyed that episode. For those of
you in the United States, I hope you're having a
healthy and safe and fun Labor Day. For everyone else,
I hope you're having a great Monday, you know. I
hope your day is fantastic too, And I'll talk to
you again really soon. Tech Stuff is an iHeartRadio production.
(49:29):
For more podcasts from iHeartRadio, visit the iHeartRadio app, Apple Podcasts,
or wherever you listen to your favorite shows.
Welcome to Tech Stuff, a production from iHeartRadio. Hey there,
and welcome to tech Stuff. I'm your host, Jonathan Strickland.
I'm an executive producer with iHeartRadio. And how the tech
are you? Here in the United States, we are celebrating
Labor Day, a federal holiday here in the States. Also
(00:26):
one that I find really interesting because it's all about
celebrating the American labor movement. But if you pay attention,
especially in the tech space, there are a lot of
entities out there, a lot of companies that are eagerly
opposing the labor movement and trying to do things like
prevent workers from organizing and forming unions. Yeah, a complicated
(00:51):
thing that we have a holiday to celebrate it, and
yet we have plenty of examples of companies and organizations
out there dedicated to venting more labor organization from happening.
But let's put aside all of that. I thought, since
we're on holiday today, I wanted to make sure that
you had an episode anyway. So we're going to actually
(01:13):
listen to an episode that came out a few years ago,
back in twenty twenty July thirteenth, twenty twenty. In fact,
it is titled The Robots Are Coming for Your Job,
and it's all about robots and automated systems and the
anxiety that exists around this idea of automation eliminating jobs,
(01:33):
which I think has only become even more of a
talking point in the wake of things like generative AI,
for example. So let's listen to this episode from twenty twenty,
the Robots Are Coming for Your Job, and I'll chat
with you again at the end. I read a lot
of tech news and sometimes that ends up inspiring me
(01:56):
to do an episode of tech stuff. That happened to
me recently when I read this headline off of the
website tech Spot. Sony factory assembles PS four in thirty seconds,
only four humans involved in the process. A PS four,
in case you're not aware, is a PlayStation four video
(02:19):
game consoles. So this factory can build a video game
console from parts in half a minute and only four
human beings touch the ding dang thing in the process.
Those four humans, by the way, are involved in the
beginning and the end of the process. Two of them
load motherboards onto the assembly line, and a motherboard is
(02:41):
the primary circuit board for a computer system, and the
other two human beings are at the end of the
assembly line, and their job is to package the completed consoles.
All the actual assembly work is done by robots. Now
you may be experiencing a couple of different responses to
this information. I know I did. One of those was
(03:02):
a wow, that's seriously impressive. The PS four, like many
computer systems, has a lot of components, many of which
attached to one another by wire or cable. So these
robots have to be able to take these flexible components
and to join them in their proper anchor points with
the appropriate amount of pressure and precision to make a
(03:24):
good connection. Now, if any of you out there have
ever built your own PC, you know that plugging cables
in can get a little tricky depending on the layout
of the motherboard and the various components. And if you're
someone like me, you're likely putting stuff together only to
realize that maybe you should have done some of that
before you mounted them in a computer case, because now
(03:46):
you just don't have the space to work in properly.
So it's pretty darn impressive that robots can do this
consistently and correctly at that level of speed. Another response
I had was kind of scary. I mean, typically you
would have dozens of people employed on the assembly line
to do this sort of work, but in this factory
(04:09):
it's been stripped down to thirty two robots and four
human beings. The article in textpot points out that twenty
six of those thirty two robots are just attaching flexible
components together inside the console. Now, I have no idea
how much these robots cost, but I wager that they
(04:30):
are expensive enough to equal the salary of a standard
human employee on the assembly line. However, you don't pay robots.
You do have to spend money to maintain and repair them,
but assuming whatever you're making is going to be around
for a little while, they'll pay for themselves because eventually
you'll get to a point where the salaries you'd be
(04:51):
paying for humans would be more than the purchase and
maintenance cost of the robots. And the increase in efficiency
means you can produce a whole lot more stuff in
a given amount of time than you would with a
human centric assembly line, so you'll have more product to
sell in a shorter amount of time. When you start
(05:14):
crunching numbers, you discover your robotic assembly line can make
more stuff at a lower cost over a given period
of time, like you know, over a couple of years,
than what you would accomplish with human beings on that
assembly line. So you don't have to worry about the
robots taking a vacation. They don't take sick time, they
don't even take the night off. They can work around
(05:35):
the clock. They don't need health insurance, though I would
guess that most companies ensure the heck out of these
things just in case one breaks down. But from a
financial point of view, they make sense if you're building
stuff at a large enough scale, stuff like video game
consoles for the PlayStation four. It's a no brainer because
(05:55):
that console has sold around one hundred and ten million
units so far. That's a number large enough that I
can't even imagine what it would look like if you
had all those consoles together in one place. So if
there's enough demand for you to sell one hundred and
ten million whatever it is you want to sell, you
(06:16):
need to have a way to make those as efficiently
as possible, and that will help maximize your profits. And
the more efficient the process, the more competitively you can
price your product and still make a profit. But the
idea of robots performing jobs far more effectively, consistently, and
efficiently than humans raises a lot of questions. And these
(06:38):
are not new questions either, but they are questions like
if more factories rely on robots for production, particularly if
those robots can be programmed to produce new products once
older ones go obsolete, what happens to the job market,
What happens to the millions of people who work in
manufacturing on assembly lines? They go? What will this do
(07:02):
to economies around the world. Lots of people have tried
to answer these questions, sometimes giving drastically different answers. And
we're going to take a look at the history and
evolution of industrial robots in this episode and explore the
ramifications of automated manufacturing. And this is where I dive
(07:22):
into history. I've talked about the history of robots before,
so I'll try to restrict my focus to industrial robots.
And before I get into that, let's just address the
fact that the use of machinery to increase efficiency has
been a controversial subject since long before there ever was
such a thing as a robot. Generally speaking, machines are
(07:45):
meant to make work easier or in some cases make
the work possible just to begin with, they are labor
saving devices, requiring humans to put forth less effort to
get the same or better results. This applies to the
simplest of machines, I mean stuff like levers or pulleys
or an inclined plane, and it applies to very complex
(08:09):
machines as well. Before the Industrial Revolution, most stuff like
textiles was made by crafts people out of their own homes.
This was literally the cottage industry. Trades people would travel
and become the lifeline for the cottage industry, supplying raw materials,
buying finished products, and selling those products off at a
(08:31):
profit elsewhere. Many tradespeople built a good deal of wealth
working this way, and they had the means to look
at alternatives to this decentralized cottage industry approach. An idea
began to form. If you brought together crafts people to
a centralized location, and if you simplified the process of production,
(08:54):
you could make way more stuff, which in turn means
you could sell way more stuff, which in turn means
you can make way more money, and money makes the
world go round. This thought process helped fuel a similar
line of thinking. If you could design machines that could
(09:15):
do a lot of work that typically fell to skilled
crafts people. You wouldn't need the crafts people at all.
You could train anyone, even if that person had no
experience with the process, just to work the machine. And
while it might take years of dedication to go through
the process of being an apprentice to learn a trade
(09:35):
well enough so that you can actually make a living
at it, with a machine, you can skip right over that.
As long as the machine's end product was good enough.
It didn't have to be better than the stuff crafts
people were making. It just had to be good enough
and cheap enough and fast enough to produce. Then you
(09:55):
could sell the finished product at a lower cost than
what crafts people would charge because not as much time
and effort went into making the thing. Now, I guess
it's clunky to talk about this while using an example,
so let's go with a poster child for the Industrial Revolution, weaving.
The weaving trade is an ancient one, and it requires
(10:16):
a good deal of skill to do it well by hand.
In the late seventeen hundreds, a man named Edmund Cartwright
patented a loom powered by a water wheel. The loom's
operation was such that a person who had no training
in weaving could operate the machine and produce finished textiles.
Cartwright's design would be built upon by other inventors who
(10:39):
had turned to steam power and other means to operate
the power loom. Many cottage industry weavers found themselves out
of work. They could potentially opt to work in the
textile factories, as those were popping up all over the place,
particularly in England, but the wages were low. As you
can imagine, this didn't sit well with the weavers. There
(11:04):
were protests, including some that incorporated violence and destruction. Ultimately,
the factory process won out, and along with it some
really awful working conditions followed, including stuff like child labor
and ridiculously low wages and dangerous working conditions. This led
(11:25):
to more protests, including the type that would give us
the word sabotage. And let's get a quick side note
on that one, as it is the source of a
little mythology or misinformation. See The apocryphal story goes that
the word sabotage comes from the word sabo, which describes
(11:45):
the wooden shoes worn by laborers, mainly Dutch laborers, but
also laborers in France. And according to the story, these
laborers wore those shoes and used them to great effect.
In an effort to protest the conditions and factories, they
would toss their wooden shoes into the machinery to break
(12:07):
the various gears and literally grind production to a halt,
as it were. But this story, while compelling, isn't really
the truth. Sabotage does stem from the word sabo, but
in French there is a verb sabote. This verb means
to make a loud noise with wooden shoes. Now isn't
(12:29):
it great that there's a verb for that? And it
makes sense wooden shoes would make a great deal of
racket as people would walk around. Heck, if a toddler
wore wooden shoes, I think it would probably sound as
though the world were shaking apart. I don't know how
toddlers manage to sound like they weigh eight hundred pounds,
(12:50):
but they do it. And if you have a toddler,
you know what I'm talking about. And in the culture
of France, the idea of a clumsy slow worker was
often linked to someone who wore wooden shoes because they're
awkward to wear anyway, The reason sabote led to sabotage
is because factory workers who were protesting their work conditions
(13:14):
and wages would purposefully work more slowly and less efficiently
in order to affect the overall output of a factory.
It was related to a similar strategy that British laborers employed,
and their version was called kakani. It was a saying
from Scotland which essentially means don't do so much man now.
(13:34):
I would argue this also feeds into a strategy that
we see to this very day in certain government offices,
where the idea is there's no need to do too
much too quickly, as it doesn't result in increased compensation
and it also sets a really high bar of expectation,
so why not just take it easy? You know, I
don't have a coffee break now. In the early twentieth century,
(13:55):
people began to use the word sabotage to really refer
to a purpose full approach to undermining the output of factories,
and it had nothing to do with tossing wooden shoes
into machinery, though it did also pertain to instances in
which workers purposefully damaged equipment and tried to slow down
(14:16):
the production that way. While this isn't directly tied to
the idea that machines themselves are displacing workers, it is
related to the effect of moving toward a manufacturing based
economy and how that allows for the exploitation of workers.
The machines themselves aren't really at fault, but they facilitate
(14:36):
the system of operations that leads to exploitation. Now that's
something that'll be a theme in this episode, and we
can't ignore the social aspect of what's going on here,
or else we miss the whole point. But let's skip ahead.
I've spoken about this before, but we get the word
robot from a check author named Kyl Kopek. He wrote
(14:59):
a pl called Rossum's Universal Robots or R you Are
in nineteen twenty. Kopek took an older word rabota, which
means forced labor in Europe. This concept was tied to
that of the old system of serfdom, in which people
would do work on behalf of a landowner. In return,
(15:21):
those people would be allowed to live on part of
that landowner's land. In ru are factory owners devise a
way to build laborers from raw materials. Now in the play,
they are indistinguishable from humans other than they have no
inner desires. But in the course of the play, these
(15:43):
laborers eventually take over all the jobs that humans previously held,
and humans themselves become a threatened species as these laborers
begin to understand the power that they hold by occupying
all the positions of employment, including as soldiers in the military.
And so with the introduction of the concept of robot,
(16:04):
we actually get the very first robotic uprising all the
way back in nineteen twenty. See I told you it
was an old idea. It's important to remember that in
the play, the robots are nearly identical to humans. They
aren't mechanical the way our robots of today are, but
the idea of creating machines that can do work without
(16:25):
a will of their own is a part of robotics
in general, and industrial robotics in particular. When we come back,
we'll talk about the earliest industrial robots and what they did,
But first let's take a quick break. It's interesting to
(16:49):
me that the tech world adopted the term robot when
we think about the origins of that word. In Cope's work,
robots were sentient slaves. They could perform the work humans
would otherwise do, but they lack the emotions that humans have,
and the whole idea is that these devices could do
(17:10):
our work for us without question or protest. They would,
in theory, endure conditions that people wouldn't or couldn't, But
in the play they ultimately lead to the destruction of
the human race and potentially they become the new dominant
species on the planet. Now. I say potentially because part
of the play's plot involves the destruction of the formula
(17:34):
that scientists use to produce the robots in the first place.
That is an important plot point. The robots are not
sure how to make more robots, so they might just
die out. Now, it seems to me as though that's
a pretty emotionally charged term to adopt for an entire
discipline of technology, right, robots, Especially if you are actually
(17:57):
aware of that play, and by the way, I recommend
people read it. It's a good play. But then a
lot of people are not aware of the origins of
the word, or at least not beyond knowing that it
came from a play in the nineteen twenty So I
guess for them, it's just, you know, a word. A
robot by any other name would smell as sweet as
(18:18):
it were. And we've definitely seen the themes of rure
serving as an undercurrent for stuff that's happening in robotics
in general. But let's move ahead. In nineteen fifty four,
an engineer named George Duvault designed an industrial robot. He
was nine years old when Kopek coined the term robot.
(18:39):
He called his design the Programmed Article Transferred Device, for
which he received a US patent in nineteen sixty one.
This machine was a robotic arm, and it was capable
of picking up something and then transferring it a short
distance away, just within reach of the arm. The arm
itself couldn't move, it was anchored in place. It could
(19:02):
also follow. In fact, this is the important part, It
would follow a pre program series of instructions to do this.
Deval's argument for his device was that up to this point,
mechanical handling of objects fell into two broad categories. Either
stuff got moved by humans, typically operating a powerful machine
like a crane or a forklift, or stuff got moved
(19:26):
by a device that operated under cam control. Now, manual
control is self explanatory, so let's talk about cams. A
cam is a rotating component in machinery. Typically a cam
has some variation in its surface. So let's start with
a wheel. Just imagine a wheel that is spinning on
an axle. Well, you wouldn't typically have a perfectly smooth
(19:50):
wheel as a cam. Part of that surface might be flat,
or it might have dips in it, And when the
cam rotates, these variations apply force to some other mechanical
component that is held against the cam, and it causes
that mechanical component to move in specific ways. A cam
operating system can work on its own, but it will
(20:13):
always repeat the exact same motions As long as everything
is working, it'll just repeat those steps. Once the cams
complete a full systematic rotation, you can't really adapt it
to do anything else. The movements depend entirely on the
cams themselves, so if you wanted it to do something else,
you would at first have to swap out the cams,
(20:35):
and even then you would be under whatever the limitations
of the device was itself, like it wouldn't have full
range of motion. Moreover, this level of specialization also means
that it's typically really expensive to rely upon cam based systems,
so it was really only useful if the application had
to do with mass manufacturing or else you're looking at
(20:56):
an economic loss. The cost of the system was just
too much. So Devaal was proposing a machine that could
be programmed to do operations. And this would let a
programmer create different processes using the same machine. Or you
could get a whole bunch of the same basic machine
and program each one to do a particular job. Meanwhile,
(21:18):
you'd free people up to work on other stuff in
the manufacturing process, and you could take the most dangerous
stuff and give it to the robots. Now, the story
goes that Daval was at a party in nineteen fifty
six and he got into a conversation with a man
named Joseph Engelberger. Joseph was a scientist and an entrepreneur,
and when the subject turned to Deval's programmed article transferred device,
(21:41):
as well as the work of a science fiction author
known as Isaac Asimov, you know, the father of robotics.
He famously incorporated a concept of the laws of robotics
in his works. We won't really go into that in
this episode, but the laws of robotics still play a
big part in the discipline of robotics in general. But
(22:02):
it's kind of outside the focus of this episode. Ingelberger
used his connections to get funding for devauld to create
a more advanced version of the programmed article transfer machine,
and it would be a robotic arm capable of making repeated,
precise movements while holding very heavy objects. They called it
the Unimate Unimate, and the first prototype, Unimate zero zero one,
(22:29):
would go to General Motors to work on a die
casting assembly line. Now, according to the company robot Works,
that's a WRX. This robot cost around sixty five thousand
dollars to produce, and Ingelberger sold it off at a
tremendous loss. General Motors only paid eighteen thousand dollars for
(22:50):
a sixty five thousand dollars machine. But Ingelberger really wanted
to establish that robotics were a way to perform repetitive,
dangerous functions at a lower risk to humans. Welding die
cast components on autobodies was a great first application of
industrial robots for a few reasons. Die Casting is a
(23:10):
process involving molten metal. You take that molten metal and
you force it into steel molds, and these are water
called dyes. The molten metal cools in the exact shape
of the mold. So this is a way to make
or cast a bunch of identical parts out of metal
and get consistent quality out of it rather than forging
(23:33):
each piece and then fitting them together. A dye can
have complex shapes in it, such as external threads, which
means you don't have to make a pipe, for example,
and then do a secondary process on that pipe to
get the result you want, So you wouldn't have to
carve those threads into a otherwise smooth pipe. You could
(23:54):
just cast the pipe with the threads incorporated on it already.
But welding die cast parts onto autobodies is hard work.
The components are really heavy, so you're at risk of
immediate injury if something goes wrong, like let's say you
drop a weighty component on your foot, or you might
develop a repetitive stress injury after going through the same
(24:17):
welding motions over and over again. In addition, the fumes
given off while welding were sometimes toxic still are, so
it's not great to have people exposed to them for
very long. So a robot was a great substitute for
a person. The robot could handle much greater weight than
people could The robot didn't breathe, so there was no
(24:37):
respiratory issue there, and it didn't get tired. I mean,
it would wear down over time, but you could repair
it in fairly short order. The Unimate worked with computer
controlled hydraulic systems. A hydraulic system uses a liquid that's
under pressure in order to do work like pushing against
a piston to power an actuator of some sort like
(24:58):
lift a platform. The UNIMATE zero zero one weighed twenty
seven hundred pounds or about one two hundred and twenty
five kilograms, and it could work twenty four hours a day,
placing components with a precision of within one fifty thousandth
of an inch. Now I'm not going to do the
conversion on that, because I think it's sufficient to say
(25:20):
that it was just really precise. According to a charmingly
dated newsreel from Britain, complete with swinging sixties music that
sounded like it came straight off an Austin Powers movie,
the robot could operate for five hundred hours without the
need for a human to check in on it. Engelberger,
a savvy businessman and promoter, would arrange for Unimate to
(25:43):
show what it could do at trade shows and on
TV appearances, including one on The Tonight Show with Johnny Carson.
If you don't know who that is, ask your parents,
and if they don't know, ask your grandparents. Nineteen sixty nine,
General Motors had jumped on board the robot train, as
(26:05):
it were. They rebuilt a manufacturing plant in Lordstown, Ohio,
and they installed unimate robots to perform spot welding on
car bodies, and the results spoke for themselves. The plant
was capable of producing one hundred and ten cars per hour,
which was more than double the speed that the plant
could manage before the installation of the robots. The business
(26:29):
case for the robots seemed clear. After a hefty upfront cost,
you could produce way more stuff per day, and as
long as the demand for that stuff is high enough,
it could mean greater revenue. You could also bring the
cost of production for an individual unit down. Then you
could pass savings on to customers and get really competitive
(26:49):
with your pricing, or you could just keep everything priced
the same and try to increase your profit margin. The
key to all this was that you had to be
sure the thing you were producing would bring in enough
money to offset the cost of automation, so it would
not make sense to spend millions of dollars building out
a factory staffed with robots. If you were making something
(27:10):
that had a very small market to begin with, yes,
you'd be able to produce way more Watching My colle
its than you could before. But if the demand for
Watchma collets is really modest, that doesn't do you any good.
In fact, you might end up flooding the market and
devaluing your product. So while robots were taking on jobs
(27:31):
that were previously held by humans, there was no real
danger of a massive upheaval where everything would be automated.
The limitations in the technology were just too great and
the cost was too high for most companies to go
that route. And this also became the starting point for
something that would become really important that the main goal
of developing industrial robots wasn't to displace humans. It was
(27:55):
meant to offload duties that were dull, dirty, or days.
You'll often hear those terms being used with robotics if
it is a job that carries with it a significant
risk to the person performing it, or a job so
demanding that you can only expect a person to stick
with it for a short while before they need to
do something else. Then building a robot to do that job,
(28:17):
or at least that list of tasks makes sense. The
robot is just a thing. It can endure conditions that
humans can't, and it doesn't get sick, and it doesn't
get hurt. If something breaks down, you can typically repair
it pretty quickly. We humans don't have that luxury. Now,
I'm not going to go and run down a full
(28:38):
history of all industrial robots because that would mostly involve
me talking about model numbers with slight differences like the
number of axes of movement or points of articulation for
one robot versus another, and that's not really interesting. But
I do want to hit a couple of highlights. One
is that in nineteen seventy five, the ASEAB robot would
(29:02):
be the first fully electrically driven robot. It also used
Intel's first chip set as processors. Now, this was not
a super strong robot because those electrically driven limbs just
can't pack the same punch as a hydraulic system, which
typically moves much more slowly but can handle much heavier payloads.
(29:24):
So this particular robot could only lift weights up to
around thirteen pounds or six kilograms, but the move toward
processors and electrically driven components marked a big technological step,
even if the arm's physical capabilities were much less impressive
than a hydraulic system. By the end of the nineteen seventies,
(29:47):
Japan was getting into the robotics game with arc welding
robots for assembly lines, and then it was off to
the robotic races, with the eighties seeing a surge in
advances with industrial robots. Soon, massive manufacturing facilities were installing
robots to take over elements of the assembly line process,
(30:08):
particularly in that dirty, dull, and dangerous category. The robots
became more sophisticated, which also added to their value. When
we come back, I'll talk more about why that's important,
but first let's take another quick break. By the mid
(30:32):
nineteen nineties, robotics companies were making machines that could coordinate
and synchronize the movements of more than one robot at
the same time, allowing for more complex manufacturing processes. By
the early two thousands, there were systems that could synchronize
the actions of up to four robots at a time,
further adding to the overall system flexibility. Now, I mentioned
(30:54):
earlier that a programmable robot is more versatile than something
like a cam operated system. Well, more sophisticated robots, with
more axes of motion and more points of articulation have
the potential to do lots of different types of jobs,
and this is of critical importance. If the robot is
(31:15):
too limited, if you can only do a small range
of motions, you can't necessarily repurpose it for new processes.
And as markets change, you may find yourself needing to
be flexible when it comes to the stuff you're manufacturing.
So let's use an extreme hypothetical example that would probably
never happen. So let's say that you run an auto
(31:38):
manufacturing facility, but then there's a massive market change and
it drastically affects the demand for your cars. There's just
not enough demand to support the production. So rather than
just you know, closing up shop and calling it a day,
your business decides to do an amazing pivot and you
(31:59):
begin to convert your manufacturing facility over to I don't know,
home appliances. Now, again, this is an extreme hypothetical example,
but let's just go with it. Okay, So here we go.
If the robots on your assembly line are powerful, but
limited in movement and function. You may find it impossible
to adapt them to your new line of business, which
(32:21):
would mean you need to either invest in new robots
or you'd have to hire human workers to put together
your appliances. And it would also mean that your old
robots would be a sunk cost. You would need to
either sell them off or put them in storage or something.
If the robots are really sophisticated, however, you might be
able to program them to do some of the operations
(32:43):
on the new assembly line, and that would keep them useful,
it would lower the cost of production. Or, for a
less extreme example, you introduce a new model of whatever
thing it is that you're producing. Anything new will require
adjustments in the assembly line process, and if the changes
are big enough, the robots may not be able to
(33:05):
make as big a contribution in the process. That's something
that could happen with the example of the PlayStation we
were talking about. Yeah, those robots can put together a
PS four and thirty seconds. There's no guarantee they'll be
able to do the same thing with a PS five,
at least not without a major overhaul of their assembly line. System.
While the manufacturing facility can churn out a finished PS
(33:27):
four and thirty seconds, we might not see them work
at all with PS five, at least not right away.
It would all have to be optimized. So for decades,
industrial robots were kept as separate from human workers as
was possible. You wanted to keep them well away from
all the people, or keep the people well away from
all the robots. Often the robots would operate within cages
(33:50):
specifically to limit the possibility of a human coming within range.
After all, these robots are large, they're heavy, powerful, and
many of them are incapable of sensing stuff in their
environment and whether or not a human is within their
range of motion. Instead, they're just going through that pre
(34:11):
programmed series of motions and they're not going to stop
unless someone turns it off. A robot is performing that
same series of steps over and over, and that can
mean that if a human in that area gets near
the robot, they could end up getting injured or worse.
And in fact, this has happened several times over the
course of the last few decades, and at least in
(34:32):
some cases it seems as though the robot might have
been at fault meaning it's not always a case of
human carelessness. For example, an engineer in twenty fifteen died
when a robot armed from one section of the factory
floor moved beyond its operating area and into the neighboring
section that the engineer was working in. This is something
(34:55):
that should not have happened. The robot arms should not
have moved that far into the neighboring section. The robot
arm hit the engineer on the head, and she later
died from her injuries. In the United States, the government
has listed thirty three workplace deaths due to accidents with
industrial robots between the years nineteen eighty four and twenty fourteen.
(35:18):
The investigations also found that the majority of those tragedies
was typically the fault of human error. There was a
person who was wandering into the operation zone of a robot.
That twenty fifteen incident was an outlier. Not that any
of this makes the thought of working around industrial robots
(35:39):
less scary or those other accidents any less tragic. They're
all terribly tragic. Moreover, we're seeing more robots that are
capable of roaming a work space. They are no longer
anchored to a specific spot on the floor. In some cases,
they also, unlike the first industrial robots, typically have external sense.
(36:00):
These not only help the robots navigate their environments, but
also hopefully avoid accidents with human workers. Let's take Amazon's
warehouse robots for example. These robots look like really big
robotic vacuum cleaners. They are designed to roll under shelves
and the shelves are just slightly larger than the dimensions
(36:23):
of the robot. And when an order comes in, a
robot from the warehouse rolls over to a shelf that
holds the respective item on it according to the inventory system,
and the robot goes under the shelf then lifts the
shelf by raising a platter like platform on the top
of the robot. Think of it as like a little
fork left, except it's more like a I don't know,
(36:45):
like a tray that a waiter would use to carry
drinks to a table. But it carries the whole shelf
up and over to the edge of a cage, where
a human operator will take the respective item off the
shelf and scan it and put it into a bin.
And then those bins go to other humans who further
scan those items and then put them into other bins
(37:06):
that ultimately go to the packing department. And if you
watch videos of these robots, it looks like they're doing
a complicated ballet as they maneuver through this warehouse, avoiding
other robots and shelves. As they bring those shelves to humans.
Markings on the warehouse floor tell the robots where they
are with respect to everything else in the warehouse, and
(37:29):
the robots even will position shelves that have items that
are being ordered a lot toward the edges of this
space so that they're easier to get to and move
them over to the human beings. So it's kind of
an interesting dynamic system. It's not like they pick up
the shelf and then bring the shelf immediately right back
to where it started. The shelf can end up in
(37:50):
a different location entirely. In addition, cameras give the robots
the ability to sense any obstacles that might block their path,
allowing the robot to come to a stop and a
wait further instructions and report that it has found something
unusual on the warehouse floor. Even so, typically humans are
(38:11):
not allowed to roam the area where the robots pick
up shelves if something has fallen on the warehouse floor.
A designated troubleshooter gets an alert, and that person must
use an interface to draw the path that they are
going to take from the entrance of the cage all
the way to the point of trouble. Like let's say
(38:32):
that a product has fallen out of a shelf and
has hit the floor, and a robot has reported it.
You would use a tablet. If you're the troubleshooter, you'd
use a tablet and you would draw, almost like a maze,
the path you would take to get to that particular item,
and you would follow that path out and back. In addition,
you'd wear a radio transmitter that would send a signal
(38:55):
out that the robots could all detect, and that would
alert the robots to the presence of you, the troubleshooter.
That helps prevent a situation in which the robots are
going to collide with you, right, you want to avoid that. Now,
there's a lot of work that goes into designing robots
that can interoperate in a space that's occupied by humans,
(39:15):
and it's a very challenging line of technology because it
takes more than just thinking about how the machines work
you also have to think about how people work, and moreover,
you have to think about how people change the way
they work when they're in the company of a robot.
It's kind of similar to the concept in quantum theory, right,
(39:35):
the idea that you change a thing you observe just
through the act of observation. Well, you can have a
workspace that humans had only been working in for a while,
and you could say, all right, well, I've observed how
the humans work, and I'm going to build a robot
that does this one task that the humans do, and
I'm just going to incorporate right into their workspace. But
(39:58):
then you find out that when you do that, the
humans all behave in a new way because there's a
new thing in the environment that you didn't account for,
and now the design of your robot doesn't work as well.
We humans are tricky like that. Moreover, we need to
get to that threat that weavers were worried about more
than a century ago. Is automation going to take our jobs? Now?
(40:22):
There have been a few studies, all using different methodologies,
and some of those studies coming under criticism for the
approaches that were used. But there have been a few
studies that suggest we'll see automation continue to impact jobs
in the near future and drastically so over the course
of the long run. The interpretation of those results have
(40:43):
been reported in ways that range from automation is going
to be disruptive that's on the light end to fifty
percent of all jobs are going to be taken by
the robots. So what's the actual truth. Well, the truth,
as it turns out, is complex. For one thing, automation
rarely takes over an entire job. What is far more
(41:06):
likely to happen is that automation will take over certain
tasks that are part of a job, or perhaps multiple jobs.
So if a job requires a wide variety of tasks,
some of which may require critical thinking, it's really hard
to design a robot that can do all of that.
It's far more likely that you would automate certain job responsibilities,
(41:30):
which would mean that those jobs themselves wouldn't go away,
they would just change. The repetitive responsibilities would be offloaded,
and you would focus on something else. You might have
to spend more time doing other duties rather than the
routine ones, which isn't necessarily a bad thing, but there
are cases where automation would likely take over an entire job,
(41:55):
for example, truck drivers in shipping trucks. Much of the
work in autonomous vehicles is really focusing not necessarily on
replacing passenger vehicles so much as commercial vehicles like shipping trucks.
The Bureau of Labor Statistics in the United States estimated
(42:15):
that the age of the average US truck driver is
fifty five and more than ninety percent of all truck
drivers in the US are mail and that will present
a challenge see Generally, the pro argument for automation is
that while robots and automated systems will eliminate some jobs,
(42:36):
they will create other jobs, presumably better jobs. And this
is true. At the turn of the twentieth century, forty
percent of all jobs in the United States were on farms.
So that means four out of ten people in the
US who had a job we're working on a farm. Today,
(42:57):
Agriculture and all the related food sectors make up just
eleven percent of all jobs in the United States. And
if we just limit this to the people who are
working on farms, you know, not all agricultural jobs and
food sector jobs, just the farm jobs. If we do that,
we're talking about only one point three percent of all
(43:17):
US employment, So going from forty percent to one point
three percent, that's a drastic change. Now, clearly automation has
transformed agriculture. It allows us to do a lot more
while relying on fewer people, and new jobs did come around,
so we didn't see an unemployment rate reaching levels higher
(43:38):
than forty percent pre COVID. The pro automation argument states
that new jobs, which again should ideally be better than
existing jobs, as in less strenuous and less dangerous and
more interesting, will emerge as older jobs are phased out.
Now that works fine on a macro scale when you're
(44:01):
taking a really big picture look at the overall trends,
but when you consider the particulars, like our truck drivers,
you start to see some obstacles. See this year, I
turned forty five, so I'm a lot closer to the
average age of a truck driver in the United States
than I am to someone who's just getting into the
(44:21):
job market for the first time. And I can tell
you that even as a relatively tech savvy guy, I
would find it really challenging to pick up the job skills.
I would need to go into a different line of work,
particularly one where I'm competing against people who already have
training and experience in that field. So imagine having to
(44:44):
tell a group of fifty five year old truck drivers
that they're out of a job. But good news, if
you just start taking classes, you can learn to code
and make less money than you did in your old job.
It's not great, is what I'm saying. That being said,
automation is clearly not going anywhere. It's going to continue
to play a big role in how we get work done,
(45:07):
and in our best case scenarios, it's going to augment
the work that humans do, leading to better, more efficient,
and more cost effective outcomes. It will free us up
to focus on the parts of our jobs that we
find the most fulfilling. We can handle the stuff that
requires flexibility and intuitive thinking, and the machines can handle
(45:27):
the routine and the dangerous. But in a worst case scenario,
we'll see an unprepared population of former workers who are
now out of a job and without the support system
there to help them transition into something new so that
they can continue to contribute to society and earn a living. Now,
(45:49):
this is why you will often hear conversations about automation
get tied into concepts like a guaranteed jobs program. This
is typically where something like a government creates a system
that makes certain every person who wants a job can
get a job. Or you'll hear about guaranteed basic income.
(46:10):
This is a strategy in which tax dollars go to
fund a standard income payout to all citizens so that
they can meet their most basic needs. Now, these are
big ideas, they aren't easy to implement or administer, and
they're not cheap, But it may be that they will
become necessary or some similar strategy will be needed to
(46:33):
make certain that we have a plan to move toward
rather than being caught in a world where a disproportionate
percentage of people can't find gainful employment. Heck, we're seeing
something like that right now due to the COVID crisis,
which is also underlining the importance of automation in a
world where it's not necessarily safe to have a bunch
(46:54):
of human beings all gathered in the same place at
the same time. Robots coming for our jobs, well, for
some of our jobs, definitely, many of those jobs come
with some pretty tough consequences for humans who are working
those jobs today. Those jobs may have high injury rates,
the people who work them may have lower life expectancies,
(47:17):
and they are a whole host of health issues that
can come along with certain jobs. So you could make
a strong argument that really this is for the best
because it will help save lives and reduce the chance
for injury or illness for a lot of people. But
for other jobs, the robots aren't likely to take over
in the near future. For a lot of jobs, automated systems,
(47:38):
not necessarily robots, but perhaps software based AI, will augment
what humans are doing. It's important we have conversations about
this stuff and to talk about how to address the
consequences of increased automation. There are ways we can enjoy
the benefits of automation, but only if we think critics
(48:00):
about it and create policies and procedures accordingly. Now I
gotta get going. I hear robo Jonathan is going to
host the next episode of Tech Stuff, and I have
to train them on how to make puns and pop
culture references. I hope you liked that episode from twenty twenty.
As I said, We've got a lot more to talk
(48:21):
about these days because of things like generative AI and
the concern that that could impact jobs that for a
long time people assumed we're safe from automation, especially from
things like robotics, you know, white collar jobs that people
just thought were kind of the domain of humans, And
now there's a real question as to whether or not
that's actually the case. So I think it's even more prevalent.
(48:44):
And again seeing the labor movement in the tech sector
in particular over the last few years tells us that
there's some very important issues still at the very heart
of technology and the way we do business that relate
back to the foundations of the labor movement here in America.
(49:04):
So I hope you enjoyed that episode. For those of
you in the United States, I hope you're having a
healthy and safe and fun Labor Day. For everyone else,
I hope you're having a great Monday, you know. I
hope your day is fantastic too, And I'll talk to
you again really soon. Tech Stuff is an iHeartRadio production.
(49:29):
For more podcasts from iHeartRadio, visit the iHeartRadio app, Apple Podcasts,
or wherever you listen to your favorite shows.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.