|
|
Musings Report 2019-31 8-3-19 Will Robots End Up in the Landfill?
You are receiving this email because you are one of the 500+ subscribers/major contributors to www.oftwominds.com.
For those who are new to the Musings reports: they are basically a glimpse into my notebook, the unfiltered swamp where I organize future themes, sort through the dozens of stories and links submitted by readers, refine my own research and start connecting dots which appear later in the blog or in my books. As always, I hope the Musings spark new appraisals and insights. Thank you for supporting the site and for inviting me into your circle of correspondents.
Welcome to August's MUS (Margins of the Unfiltered Swamp)
The first Musings of the month is a free-form exploration of the reaches of the fecund swamp that is the source of the blog, Musings and my books.
Thank You, Contributors!
Thank you and welcome, new subscriber-patron Boyd W.
Will Robots End Up in the Landfill?
Technology cannot repeal the laws of thermodynamics. Taking a pencil and extending the declining cost of solar panels to zero doesn’t negate the physical costs of mining and smelting the ore, shipping the metal to a factory, fabricating the photovoltaic cells, assembling and testing the panels, transporting them to the installation site on vehicles that are expensive to manufacture and maintain, installing the panels, wiring them to inverters and other equipment, testing the system onsite, and returning to perform maintenance and possibly repairs. Since the expected life of the installed panels is 20 to 25 years, the entire expense must be repeated, plus the additional expense of removing and recycling the worn out panels.
The cost of manufacturing, installing, maintaining, repairing and replacing the panels will never be close to zero due to the intrinsic costs in mining, smelting, refining, milling, transporting, assembling, testing, installing, maintaining and repairing the panels.
Even if robots perform all the work, robots are themselves resource and energy intensive. Robots are less like a computer chip (with declining marginal costs), and more like an auto, an immensely complex and costly assembly of intrinsically resource-intensive components, electronics, computer chips and millions of lines of software coding. Autos cost more than they did a generation ago, not less: as cheap-to-access resources such as metals and minerals are depleted, the remaining ores are more costly to extract; regulations require additional safety features and extremely complex software is increasingly prone to unanticipated errors.
All of these realities apply to autos, robots and every other complex, resource-intensive machinery.
To become more capable, machines become more complex and therefore more expensive to manufacture, test, maintain and repair. In a very telling edit of reality, those extolling the idea that robots will perform all of humanity’s work in the near future overlook these intrinsic costs, and overlook the expensive realities of fixing even simple machines when they fail or break down.
Consider two examples: a Japanese auto manufacturer known for the reliability of their autos encountered a semi-random flaw which they were unable to diagnose: the accelerator in one model of their cars would become stuck on, pushing the car to high speeds. At first, they attributed the problem to floor mats, then to faulty linkages and when the problems persisted, to faulty software.
A massive forensic investigation was required to locate the software bugs, a major undertaking given that software modules are added to existing coding; manufacturers don’t pay programmers to write millions of lines of code from scratch for each new model due to the high cost. As a result, routine testing did not reveal the fatal flaw; in complex software, intermittent or semi-random errors can go undetected even by extensive testing.
How will robots operating on millions of lines of software coding be any different?
Take a relatively simple machine like a clothes dryer. It is a metal box containing a heating element and a drum that spins. An electronic board with a digital display operates the machine’s cycles and controls. A dryer is far less complex than a robot, especially one that is capable of navigating the real world.
The dryer control board is relatively simple: a handful of low-cost commodity computer chips and a few circuit boards. Despite the relative simplicity, these boards fail with alarming regularity. This is also true of ranges, washing machines and other appliances. The replacement board for the dryer is one-third the cost of a new dryer. Labor adds another third, so replacing the board is two-thirds the cost of a new dryer.
The reliance on cheap commodity electronic components results in the lifespan of modern appliances being measured in years rather than decades, the lifespan of old purely mechanical appliances.
The ultimate cost of adding features, the functional value of which is very much in question, is far higher than the sticker price of the new dryer. In the real world of commoditized components and goods, technology has increased costs and consumed more resources for extremely marginal improvements (for example, ten choices of drying cycles rather than five).
Since advocates of robots claim robots will soon do all the work of humanity, consider the vast difference in cost between a robot that operates on a flat factory floor, repetitively attaching one part on a dryer assembly line and a robot that arrives onsite in the messy real world and is able to diagnose and repair a broken dryer.
The factory model operates on a flat floor; the repair robot has to navigate an irregular driveway and multiple changes in floor level. It also has to be powerful enough to lift the dryer off the washer (in a stacked configuration), move it to open ground, remove the top, perform the diagnostics, remove the defective board, retrieve the new board, install it correctly, re-assemble the case, test the repaired machine, then lift it back onto the washer.
To repair a dryer onsite, the robot will have to have the strength of a small forklift and a very high level of dexterity and precision motor control. The cost of adding each of these capabilities to a robot is extremely non-trivial, and it won’t ever drop to near-zero. Rather, it will only increase in cost even if commodity sensors and chips decline in price. The points of potential failure will proliferate with each new capability and each new level of complexity.
The robot itself is prone to the same kinds of failures that it is designed to repair, but due to its much greater complexity, repairing the repair robot could cost an order of magnitude more than repairing the simple dryer—and that repair costs two-thirds the price of a new dryer.
I’ve performed this exact repair on my own dryer (only a few years old), and other similar repairs on other appliances: a name-brand range that turned on the oven at random times due to a failed low-cost commodity electronic sensor (less than two years old), and an expensive name-brand heavy-duty washing machine, less than a year old, that failed due to a low-cost electronic sensor.
As I observed before: complex devices are only as reliable and durable as their lowest-quality component. This is as true of robots as it is for any other device.
I happen to be practical, frugal and handy. I’ve swapped out circuit boards in desktop computers and completed numerous repairs on houses, appliances, autos, etc. I’ve also learned simple markup coding (CSS, for example) as needed. In most cases, I’ve had to learn how to diagnose and effect the repair on the fly, i.e. without any previous training. As a result, I can make a more realistic assessment of real-world technology’s real-world costs than those who have limited experience with fixing complex technology when components fail.
Even with me performing the labor, the parts for all these appliance repairs were expensive. Many less handy people would have paid multiples of this already-high cost for labor, and others would have bought a new appliance and had the (still-functional other than the one failed component) appliance hauled to the landfill—a perfect example of our wasteful and expensive Landfill Economy.
Consider the hundreds of components in a so-called smart home designed to save energy and offer more convenience by networking sensors, cameras, appliances, locksets, servers, controllers, Wi-Fi chips and software. The projected advances in security and convenience, many of which are questionable—how much value is added by a refrigerator that can order a quart of milk delivered once it detects a low level in the carton?—come at a very high cost in components, installation, service and repairs as each of the hundreds of sensors, controllers, Wi-Fi chips are points of potential failure or unauthorized remote access. How long will these components last, how long before they must be replaced to function with a new software system, how functional will the system be if even one controller fails?
Given the extraordinary expense of installing and maintain this complex system and the marginal returns in convenience and security, how is this not another system destined for the landfill?
I have yet to find a true believer in robots will do all the work of humanity who has ever performed even a single repair of a complex system or device caused by a failed board, chip, sensor or software bug and done so not on a clean factory floor but in the messy real world.
I’ve never yet met an avid believer in robots will do all the work of humanity who has designed, prototyped, tested, manufactured, sold, maintained and repaired robots capable of climbing (or landing) on a roof, diagnosing the cause of a failed solar array, replacing the failed part and cleaning the panels, for a total system cost that’s less than the relatively modest cost of a human repairperson.
Here is the generalized conclusion: technology only lowers costs when it commoditizes tasks and products.
What happens to robots when the cost of diagnosis and repair exceeds the cost of a new robot? The (still functional except for the failed component) robot ends up at the landfill, alongside the dryer with the defective chip, the washer with the failed sensor, the vacuum with a broken plastic part, the car with a defective controller board and so on.
In the fantasy world of true believers, robots work perfectly essentially forever, until they’re replaced by robots that are better and cheaper. Maybe the fantasy extends to recycling the old robots. In the real world, recycling complex machines is expensive; the components must be separated, and not all parts can be recycled. It may not be possible to cost-effectively recycle the robots at all.
We always end up at the same place: the robots performing the recycling are themselves costly, complex machines that are only as reliable and durable as their lowest-quality component and buggiest module of software. When they fail, they’re as expensive to repair as the robots they’re recycling.
A system of complex devices that is inherently costly to manufacture and that are only as functional as the lowest-quality component and software in each complex device is a system optimized for waste and higher costs due to short product cycles / planned obsolescence, i.e. the Landfill Economy.
.
This is an excerpt from my newest book, which is still being edited.
From Left Field
Everyone Hates Customer Service. This Is Why. Technology lets companies see how badly they can treat consumers, right up until the moment they bolt
We’re No Longer in Smartphone Plateau. We’re in the Smartphone Decline.
The management of hardship: In an atmosphere of mistrust, conspiracy theories thrive.
Progress and Amnesia
That ‘Great Moderation’ moment again
Foreign Buying of U.S. Homes Suffers Record Drop: Pullback, led by Chinese buyers, is driving price cuts in coastal cities and causing new condos to sit empty
The backlash is growing against Xi Jinping in China and around the world
15 Habits Of People With Concealed Depression
The Cheapest Way to Save the Planet Grows Like a Weed
More Americans have died in car crashes since 2000 than in both World Wars
Die Loser: Why We Are Killing Ourselves More, and What’s the Cure? The use of SSRIs is soaring in the developed world—further evidence that we are making ourselves miserable by trying to live up to unrealistic standards.
AMERICAN SCHOOLS ARE TRAINING KIDS FOR A WORLD THAT DOESN'T EXIST -- amen....
These are the Winners of the 2019 iPhone Photography Awards
"Knowing your own darkness is the best method for dealing with the darknesses of other people." Carl Jung
Thanks for reading--
charles |
|
|
|
|
|
