Today's products are much more complicated and reliable than their predecessors, in many cases, but when it comes to fixing them, well. . . .that's another story
I spent a day recently at a local antique-auto show ("antique" is defined as 25+ years old) and came away with a few observations.
First, and this is a subjective, esthetic comment: many of those older cars have a timeless beauty. I wonder which of today's cars will look so good in retrospect, if any.
Second, the only thing electronic in those older cars was the radio and tape player. In contrast, today's cars are replete with sensors, microcontrollers, high-end processors, actuators, small motors, drive-by-wire subsystems, and millions of lines of firmware.
Third, those older cars were—and still are—very repairable and maintainable. There's an ironic aspect to this since, frankly, the reliability of many of them ranged from lousy to mediocre.
The reality is that today's cars are both much more complex and yet much more reliable than many of those older vehicles, in contrast to the "simple is better" rule that is we usually trust. For example, those old-time carburetors, with their linkages, choke, internal floats, single or multiple barrels, and various compensation mechanisms, and could be downright cranky, compared to our sophisticated, closed-loop, fuel-injected systems.
At the same time, though, if you had (or have) a problem with those non-electronic cars, diagnosing it in most cases requires only moderate skills and tools: voltmeter, pressure gauges, and some measuring instruments, along with careful observation. What you can see is pretty much what you need to see. There's no on-board diagnostics (OBD II)), no internal LAN to debug, no "system design" which fights you, or interposes itself, when you try to push and poke it to see what is going on.
Further, if you need a replacement part today for these older cars, you can almost always get one quickly from the numerous aftermarket suppliers (often small, specialty operations and shops). And if that doesn’t work, you can even get a new part made to order at a local machine shop, using the original as a guide and template.
Compare this to today's cars, and other sophisticated consumer products such as TVs, computers, and various embedded devices. On one hand, they are far more reliable, despite their increased functionality and complexity. On the other hand—and it's a big "other hand"—they are harder to diagnose and repair, and after just a few years, that situation becomes "almost impossible to repair."
Even if you can figure out what's wrong (a non-trivial exercise in many cases), you can’t get the parts, and you couldn't do anything with them even if you could. Are you going to de-solder and replace that 100+ lead processor on the main board, or the microscopic SMT components? For most of today's products, the least-replaceable unit (LRU) is an entire circuit board—or the whole product.
Does it have to be this way? Is this the tradeoff we make for the increased functionality and form-factor that users expect, with the initial reliability they demand, and all at a price they will pay? Do product designers have to eliminate connectors, sub-modules, sockets, and other non-essential aspects, to meet these requirements?
As in most things, the answer is "it depends." It is not a question with a clear-cut, unambiguous answer. But seeing those older cars, and their reliability versus repairability balance, does make you stop and think.