The Digital Transformation of Cars Is Just Beginning
mule.substack.com
mule.substack.com
Similar to how introducing new tech and new dependencies, complicated architectures etc. in a software project seems to be rewarded as long as you jump ship quickly enough before the mess you created blows up.
Also saves on component costs, and pushes manufacturing and design issues into software.
It's terrible, and I'm hoping there is a reaction against it. This is the #1 thing that stops me from having any interest in buying a Tesla.
A few months ago I took a rental car - I don't drive much any more so I don't own one. On the (German) Autobahn there always is a lot of construction going on. There are yellow stripes overriding the white ones for where the lanes are, because often they reroute one or two lanes to the opposite side so the lanes get a lot more narrow for all. You still have four lanes but only 2/3rd of the width, while they work on the now free and unused part.
Unbeknown to me my car had active lane keeping assistance and I had a really hard time when it tried, with not insignificant force especially when it surprises you, to keep the car within the white stripes. I could have collided with the car next to me with a little bad luck there. It was not easy to force the steering wheel, against the power of the active system, with only one hand while using the other one to navigate the menus to turn off that darn dangerous system. Or find one of the 50 buttons to do it.
The most dangerous part was the initial seconds when I crossed over to the lanes, following the yellow stripes, and suddenly somebody, some force, took the steering wheel and steered in the opposite direction. It took me a few seconds to understand what was going on, I did not expect that there would be a turned-on active system. I had only had experience with passive systems that would not interfere with steering, cruise control and distance-to-car-in-front control as the only active systems. But taking my steering wheel from me???
Also, in my experience it doesn’t really matter if it’s a touch screen, a touch control, or a tactile button. People don’t know their dash controls, even if they’ve owned a vehicle for years (and I’m talking about systems that are 100% physical switches).
As an exercise, try drawing out what the center stack of your vehicle is, where all the switches are and what they do. Alternatively, the next time you drive try setting the HVAC to something specific without taking your eyes off the road, at all. How long does it take? How much of your attention has shifted from driving?
once, to turn up the temperature in a car, all that was needed was a twist of a wheel, now, you are required to go through every single degree celsius to do the same.
Modern vehicles are a network of ~20 different independent computer systems. The UI (regardless of touchscreen) is only one of them.
Eventually a car maker will produce an economy car without a rear window. Glass is expensive. People just end up using the camera system anyway. It'll do better in a roll-over test, without the rear glass.
The answer to the problem of you life depending on complex high-maintenance parts that do not operate under normal environmental and operation conditions is not to throw increasingly more complex and high-maintenance parts at it. Safety is not only for those willing to pay extra for it. The 737 MAX 8 should not be held up as an exemplar of good practice but as a stern warning.
At highway speeds they add drag to the car. Car manufacturers are trying to make their MPG seem low.
It also requires electronics, a screen, wires and electricity to operate. A rube goldberg contraption that can fail in any number of ways. While my existing boring low tech mirrors just reflects existing photons without any aforementioned nonsense. Just why?
Like you know if you take a look at the innards of a 80s pc (say 286) it will be chock full of discrete ICs. Caches, bus controllers. Now its just a handful of chips plus tons of firmware.
I find it strange that the number of features and sensors would so directly drive the number of semis. Maybe at first, but in the end a car will be a ”SoC” (or few) like everything else.
What am I missing?
A car being one chip makes as little sense as trying to hook up a whole office to one server, running all the display cables and USB for peripherals to one spot. Sure, it maybe could be done, but who wants to spend the money dragging all those wires vs one run of cat6? Same thing with cars. For instance, it isn't uncommon for the automatic window roll down feature to be implemented in the door itself. It's just easier to have more localized controllers that communicate over a network (CANBUS) than it is to run physical wiring to send every signal to where it needs to go.
Sure, it isn't uncommon to have very underpowered machines on people's desks that don't do much more than drive a monitor and process keyboard input, but there still needs to be some sort of machine to connect those peripherals and send info over the network. This isn't that different from a car, this is exactly how there end up being 30+ different modules in a vehicle.
Think of the automatic window roll down feature in the front passenger door. Thus relies on either sensing motor current or timing. It's way easier to throw a small microcontroller in there to accomplish this than run 20-30 feet of wiring in total to send all the signals back to one master controller. Wiring is way more expensive than most people realize, the wiring harness on a car is tremendously expensive to produce. Reducing wiring complexity saves enough money that it is easy to justify the engineering and BOM cost of an additional module.
- Would this drive up the cost of these vehicles?
- Is there likely to be a fundamental change in the way that these components are sourced, a move away from just in time?
- Is this mirrored in other industries?
The obvious parts are the progression of technology being added to cars over time, and the corresponding increase in cost. Some of the connectivity wasn't explained (V2I, V2V, V2P) but I assume Vehicle to Internet, Vehicle to Vehicle... OK I looked it up, and here's a full reference[1].
[0] https://www.eetimes.com/level-5-avs-unlikely-before-2035/
Actually it's already on going, with less interest in car ownership in the 21st century, the advent of Uber and similar services, and renewed interest in modern public transportation, at least from countries that progress to the future (like China, India, and so on), not from countries that cling to the past and let their infrastructure deteriorate (US but also common in Europe, where several countries are basically shells of their 19th/20th century self, infrastructure, geo-politically, and demographics wise).