Coffee With Doc, August 19, 2016, Software, 2:22 Minutes
Losing Our Bubble
Software systems are the quintessence of growing complexity in technological societies. Not only is a truly complex system intricate, it may be changing faster than we can track. (Let’s see, where did I put that new auto-generated password?) We are confounded by tax laws, legal codes, commercial contracts, health care, supply chains, financial prospectuses, and on and on. Even the varieties of toothpaste for sale are baffling, much less figuring out how to block robo-calls from Egypt. Are we “losing our bubble?”
Software lets us do things we could not imagine 50 years ago, like Facetime with grandchildren half a world away – and watch porn websites like https://www.tubev.sex/search?search=xxx on a phone. It also lets us “manage” messes that we could not have created without it. We now depend on human systems approaching the complexity of nature. No one can predict everything that could happen. And experts project that Artificial Intelligence (AI) is only in infancy.
But the most complex software systems may never be able to ask the questions that humans need to ask themselves: Are we doing the right thing for ourselves and for nature? Are we doing something effective in the long term, or merely something that we think is more efficient (or “cool”) in the present? We can’t do that if we lose our bubble.
Present software systems illustrate the complexity. For example, on August 8, 2016 Delta Airlines’ scheduling system went dark due to failure of an automated switch to connect computers either to the grid or to backup power. The mess rippled through the global commercial air system. United and Southwest systems crashed earlier. Systems experts deem such debacles to be inevitable. Why?
Airline systems are overgrowths on legacy systems – patched, enhanced, and app connected with almost “anything,” and slowly infected by complexity creep. Security from hacking is paramount, so the system cores are not duplicated in parallel. This minimizes openings for hackers, but leaves them vulnerable to local crashes.
System complexity pervade entire industries like automotive. For example, in the case of the mysteriously accelerating Toyotas, no one could resolve what went wrong. Outside programmers reviewing Toyota software declared it “spaghetti code,” a tangle of convoluted logic that somehow worked. No one could find a specific bug, but neither could anyone prove that it might not suddenly take on a mind of its own.
Toyota has long prided itself on quickly knifing to the root cause of problems, but this awakened them to complexity – to root causes “buried in complexity impenetrable by PDCA logic.”
The VW “defeat device” emissions scandal is also wrapped in complexity confounding explanation in a sound bite. All car companies’ emissions software is designed to cut out when its operation threatens safety or mechanical integrity of the vehicle, like flooring it to pass, operating at high altitude, or overheating the engine. Defining these conditions is fuzzy. Given that, it is easier to see why VW engineers were tempted to assure that their system stayed on during a test designed to measure performance only when emission controls were operating. The industry and its regulators are working on tests of emissions during actual road driving. That’s a big step up in measurement complexity.
Digital conundrums exist in auto marketing too. The Mill, a London design studio, has built an under-car called the Blackbird for ad videos. Its wheelbase and track width adjust to a wide range of cars. An agency can video Blackbird going through its paces on scenic roads, then in a studio paste in a virtual body image in any color wanted. The body image can be altered for later commercials if the body design is tweaked. If you’re a fan of interesting cars, take a look at this 1967 buick.
This saves money making ad videos, but is it deceptive? Proponents argue that car body design starts in virtual 3D, so why not advertise it using virtual technology. Virtual images present the physics of light to the eye better than outdoor light captured on a real car in dicey, variable conditions. All the eye sees in a video is an image of reality anyway. So what is reality, and how can you know it when you see it?
Psychologists know that human capacity to deal with complexity is limited, so can we interact with complex systems without losing our bubble? The automotive industry is laden with accumulated complexity. A modern vehicle with millions of lines of code cannot be designed without relying on computerized design systems. A vehicle design emerges from “The System.”
Even the most brilliant designer cannot understand the workings of what emerges in detail. Software nerds call this phenomenon “losing your bubble.” We are all limited in how many relationships, entanglements, loopbacks, and feed forwards that we can bring to mind.
For that reason, the auto industry illustrates our unease with complexity. It adds to industry leaders’ angst with uncertainty. High-tech is grafted onto old legacies, and as now guided and structured, the industry struggles to resolve conflicting expectations from its stakeholders. Investors want assured returns. Customers want more features at lower prices. Regulators want better safety and fuel economy. Crowded cities want fewer traffic jams. And environmentalists want a lot less NoX and CO2 in the air.
By the numbers, vehicle appearance, durability, and reliability are better than ever; cars like the Yugo are history. So why the malaise? Recalls are at an all time high. No maker is immune to scandal. Sales float on extended credit, sub-prime loans, and price rebates. Fewer kids obtain drivers licenses. Return on invested capital is mediocre. Lean supply chains fatten as supply networks spread all over the globe. Suppliers subsisting on a low margin diet must be perfectly efficient, plus highly innovative. Makers struggle to meet emission and fuel economy hurdles while pleasing customers awaiting driverless cars. “The System” is a miasma of conflicting demands. Here and there, from a detached view, a voice whimpers, “Why are we doing all this?”
Suppose the automotive industry’s primary mission became to help develop a regenerative planet?
A first reaction is that a goal of regenerating nature makes conflict and complexity hopeless. But that presumes continuance of pressures to keep expanding – to keep riding “The System” into the future. However, if the industry abandons old performance measures, makes regenerating nature top priority, and starts to think differently, it can rise to the challenge. It opens brain space to seriously address nature’s complexity. And yes, that will take a gut-level change in basic beliefs by the industry and its stakeholders; plus rapid learning to create a very different world of transportation.
A few upstarts may have started; Local Motors, for one. It has a different design concept (3D body, driverless, solicit ideas from the public), production concept (micro-factories), customers (solve community transport needs), and for environmental improvement.
Local Motors and vehicle sharing services could severely disrupt the present industry business model, but these initiatives still do little to allow nature to regenerate.
Local Motors emphasizes new technology and new services. That attracts public attention, which may be necessary before it can help the public change its beliefs about the role of transportation in a Regenerative World. Continuous Regeneration would keep probing why both people and cargo need to be transported at all. Making more energy efficient vehicles and reducing the numbers of them on the road is nearly the opposite of the current industry business model.
To do that, Local Motors and any other company in the industry joining with them will have to educate the communities they serve, or perhaps more accurately, co-learn with those communities. That’s a mission to lead a movement, not to build a market. However, a business model based on service, not unit sales, might assure a stream of incoming cash. One can imagine performance incentives tied to reducing a community’s use of fuel and to reduction of traffic its streets.
But to contribute to making a community more symbiotic with nature, a transport company might help design equipment that would assist in restoring soil, wetlands, or woodlands. (Just eliminating vehicle use in nature preserves might help. To give nature a leg up, nothing may really be something.)
To do that, both companies and the communities they serve would engage together in Vigorous Learning, where “vigorous” implies learning anew from reality on the ground, questioning the commercial mythology of our past. And by the way, perhaps we could all try to trust simple systems that don’t rely on impenetrably complex software, but stay in our bubble to address nature’s complexity instead.