My last blog entry, “What my broken down car taught me about the future,” has compelled me to write a new essay that shows how some of its insights will apply more generally in the future, and not just to cars and related industries. Due to several factors, manufactured objects will generally last much longer in the future, and sudden catastrophic failures of things will be much less common.
Things will be made of better materials
Better computers that can more accurately mimic the atomic forces and chemical reactions will be able to run simulations that lead to the discovery of new types of alloys and molecules. Those same computers will, perhaps with the aid of industrial and lab robots, also find the best ways to synthesize the new materials. Finally, the use of machine labor at every step of this process will basically eliminate labor costs, allowing the materials to be produced at lower cost than they could be with human workers today.
This means in the future we will have new kinds of metal alloys, polymers and crystals that have physical properties superior to whatever counterparts we have today. Think of a bulletproof vest that is more flexible and only half as heavy as Kevlar, or a wrench that is lighter than a common steel wrench but just as tough. And since machines will make all of these materials at lower cost, more people will be able to afford them and they will be more common. For example, if carbon fiber were cheaper, more cars would incorporate them into their bodies, lowering their weight.
Things will be designed better
In my review of the movie Starship Troopers, I discussed why the fearsome assault rifle used by the human soldiers was flawed, and why it would never come into existence in the future:
It wouldn’t make sense for people in the future to abandon the principles of good engineering by making highly inefficient guns like the Morita. To the contrary, future guns will, just like every other type of manufactured object, be even more highly optimized for their functions thanks to AI: Just create a computer simulation that exactly duplicates conditions in the real world (e.g. – gravity, all laws of physics, air pressure, physical characteristics of all metals and plastics the device could be built from), let “AI engineers” experiment with all possible designs, and then see which ones come out on top after a few billion simulation cycles. I strongly suspect the winners will be very similar to guns we’ve already built, but sleeker and lighter thanks to the deletion of unnecessary mass and to the use of materials with better strength-to-weight ratios.
That same computer simulation process will be used to design all other types of manufactured objects in the future. Again, as computation gets cheaper, companies will be able to run simulations to find the optimal designs for every kind of object. Someday, even cheap, common objects like doorknobs will be the products of billions of computer simulations that stumbled on the optimal size and arrangement of components through trial-and-error experiments with slightly different combinations.
As a result, manufactured objects will be more efficient and robust than today, but most won’t look different enough for humans to tell they’re different from today’s versions of them. The difference will probably be more apparent in complex machines like cars.
Things will be made better
Even if a piece of technology is well-designed and made of quality materials, it can still be unreliable if its parts are not manufactured properly or if its parts aren’t put together the right way. Human factory workers cause these problems because of poor training, tiredness, intoxication, incompetence, or deliberate sabotage. It goes without saying that advanced robots will greatly improve the quality and consistency of factory-produced goods, as they will never be affected by fatigue or bad moods, and will follow their instructions with perfect accuracy and precision. As factories become more automated, defective products will become less common.
Things will be used more carefully
As I noted in the essay about cars, most cars have their lifespans cut prematurely short by the carelessness of their owners. Gunning the engine will wear it out sooner, speeding over potholes will destroy shocks, and generally reckless driving will raise the odds of a car accident that is so bad it totals the vehicle.
Every type of manufactured object has engineering limits beyond which it can’t be pushed without risking damage. Humans lack the patience and intelligence to learn what those limits are for every piece of technology we interact with, and we lack the fine senses to always stay below those limits. While trying to unscrew the rusted bolt, you WILL put so much torque on the wrench that you snap it.
On the other hand, machines will have the cognitive capacity to quickly learn what the engineering limits are for every object they encounter, the patience to use them without exceeding those limits, and the sensors (tactile, visual, auditory) to watch what it’s doing and how much force it is applying. No autonomous car will ever overstress its own engine or drive over a pothole so fast it breaks part of the suspension system, and no robot mechanic will ever snap its own wrench trying to unscrew a stuck bolt. As a consequence, the longevity of every type of manufactured object will increase, in some cases astonishingly. The average lifespan of a passenger vehicle could exceed 30 years, and a simple object like a knife might stay in use for 100 years (until it had been worn down by so many resharpenings that it was too thin to withstand any more use).
Things will be maintained better
Even if you have a piece of quality technology and use it carefully, it will still need periodic maintenance. A Mercedes-Benz 300 D, perhaps the most reliable car ever made, still needs oil changes. Your refrigerator’s coils need to be brushed clean of debris periodically. You hand tools need to be checked for rust and hairline cracks and sprayed down with some kind of moisture protectant. All of your smoke alarms must be tested for function once a month. It goes on and on. If you own even a small number of possessions, it’s amazing to learn how many different tasks you SHOULD be undertaking regularly to keep them maintained.
Needless to say, few people take proper care of their things. Usually they didn’t read the user manual, memorize the section on maintenance, set automatic digital reminders to themselves to perform the tasks, and then rigidly follow them for the rest of their lives. So sue them, they’re only humans with imperfect memories, limited personal time, and limited self-discipline.
Once advanced robots are ubiquitous, these human-specific factors will disappear. Your robot butler actually WOULD know what kind of upkeep every item in your house needed, and it would do it according to schedule. Operating around the clock (they won’t need to sleep and could plug themselves into wall outlets with extension cords for indefinite duration power), a robot butler could do an enormous amount of maintenance work for you and could devote itself to truly minuscule tasks like hunting down and finding tiny problems you never would have known existed.
I’m reminded of the time I noticed a strange sound in the bathroom of my house that I seldom use. It was the toilet, and the water was flowing through it continuously, making a loud trickling sound. After removing the lid, I immediately saw the problem existed because the flush lever–which was made of plastic–had snapped in half, causing the flapper to jam in the open position.
Upon close inspection I noticed something else wrong: The two, metal bolts that held the toilet tank to the bowl were so badly rusted that they had practically disintegrated! In fact, after merely scraping the left bolt with my fingernail, it fell apart into an inky cloud of rust that spread through the water. It was a small miracle that the heavy tank hadn’t slid off already and fallen to the floor (this would have flooded the house if it had happened when I wasn’t home).
I went to the store, bought new bolts, a new flapper, and a new flush lever, and installed them. The toilet works like new, and its two halves are tightly joined again as they should be. Inspecting the inside of your toilet tank is another one of those things every homeowner should probably be doing once every X years, but of course no one does, and as a result, some number of tragic people suffer the disaster I described above. However, thanks to house robots, it will stop. And of course the superior maintenance practices will not be confined to households. All kinds of businesses and buildings will have robots that do the same work for them.
People also commonly skip maintenance because they lack the money for it. As I wrote in my essay about cars and the car industry, this will be less of an issue in the future thanks to robots doing work for free. Without human labor to pay for, the costs of all types of services, including maintenance, will drop.
Problems will be found earlier
A beneficial side effect of more frequent preventative maintenance will be the discovery of problems earlier. Putting aside jokes about scams, consider how common it is for mechanics to find unrelated problems in cars while doing an oil change or some other routine procedure. Because components often gracefully, rather than abruptly, fail, machines like cars can keep working even with a part that is wearing out (e.g. – cracked, leaking, bent). The machine’s performance might not even seem different to the operator. That’s why the only way to find many problems with manufactured objects is to go out of your way to look for them, even if nothing seems wrong.
Again, once robots are ubiquitous and put in charge of common tasks, they’ll do things humans lack the time, discipline, and training to do, like inspecting objects for faults. Once they are doing that, problems will be found and fixed earlier, making sudden, catastrophic failures like your car breaking down on the highway at night less frequent.
Repairs will be better
Just because you find a problem before it becomes critical and fix it doesn’t mean the story is over. Some catastrophic failures of machines happen because they are not repaired properly. As robots take over such tasks, the quality and consistency of this type of work will improve, meaning a repair job will be likelier to solve a problem for good.
Machines will be better-informed consumers, which will drive out bad products
My previous blog essay was about my quest to find a replacement for my old car, which had broken down. It was a 2005 Chevrolet Cobalt, which I got new that same year as a birthday present. Though I’d come to love that car over the next 19 years, I had to admit it wasn’t the best in its class. I drove it off the lot without realizing the air conditioner was broken and had to return a few days later to have it fixed. After a handful of years, one of the wheel bearings failed, which was unusually early and thankfully covered by the warranty. My Cobalt was recalled several times to fix different problems, most notoriously the ignition switch, which could twist itself to the “Off” position while the car was driving, suddenly locking the steering wheel in one position and leaving the driver unaware of why it happened (this caused 13 deaths and cost GM a $900 million class-action lawsuit, plus much more to fix millions of defective cars). Whenever I rented cars during vacations, I almost always found their steering and suspension systems to be more crisp and comfortable than my Cobalt, which felt “mushy” by comparison.
The 2005 Honda Civic was a direct competitor to my Cobalt, and was simply superior: the Civic had better fuel economy, a higher safety rating, better build quality, and the same amount of internal space. Since the Civics broke down less and used less gas, they were cheaper to own than Cobalts. When new, the Civic was actually cheaper, but today, used 2005 Civics actually sell for MORE than 2005 Cobalts! With all that in mind, why were any Chevy Cobalts bought at all? I think the answers include brand loyalty, the bogus economics of trading an old car for a new one, aesthetics (some people liked the look of the Cobalt more), but most of all, a failure to do adequate research. Figuring out what your actual vehicle needs are and then finding the best model of that type of vehicle requires a lot of thought and time spent reading and taking notes. Most people lack the time and skills for that, and consequently buy suboptimal cars.
Once again, intelligent machines won’t be bound by these limitations. Emotional factors like brand loyalty, aesthetics and the personal qualities of the salesperson will be irrelevant, and they will be unswayed by trade-in deals offered by dealerships. They will have sharp, honest grasps of what their transportation needs are, and will be able to do enormous amounts of product research in a second. Hyper-informed consumers like that will swiftly drive inferior products and firms out of the market, meaning cars like my beloved Chevy would go unsold and GM would either shape up fast or go bankrupt fast (which they actually did a few years after I got my car).
If companies only manufactured high-quality, optimized products, then the odds of anything breaking down would decrease yet more. Everything would be well-made.
In conclusion, thanks to all of these factors, sudden failures of manufactured objects of all kinds will become rarer, and their useful lives will be much longer in the future than now. This will mean less waste, fewer accidents, and fewer crises happening at the worst possible time.
Related: