Just how many people do we need doing that job, anyway?
Think of your favorite "web 2.0" service or app. Maybe it's something that sends dog walkers around to your house regularly. Perhaps it can be used to deliver your favorite kind of pizza or beer. It could even be something that lets you chat with other people. We'll go with dog walking as the example here.
Now think about the entirety of humanity. At the moment, there seem to be about 7.8 billion people running around on this planet (plus a handful in orbit). For the sake of this thought experiment, assume all of them have Internet access and actually have some use for dog walking services.
Consider this: just how many really good people do you suppose it would take to saturate the market and provide service to the entirety of humanity for that dog walking dispatch?
Me? I think it's about 100 people, tops. Granted, I'm talking about the top 100 people in the population for solving this specific problem: running apps that dispatch dog walkers to dogs... for all ~8 billion of us.
They need not work at the same company. For the sake of some realism, imagine them split up somehow. It could be 20 companies with 5 people each, 5 companies with 20 people each, or 10 companies with 10 people each. Whatever.
Now let's say you looked at the actual marketplace and determined there were closer to 100,000 people actually working on these dog-walking apps. What do you suppose that means? What could possibly be going on there?
If humanity can solve the problem with 100 people and yet has thrown 100,000 people at it, there's a very good chance that 99,900 of those people are, well, superfluous. Some of them are probably doing work that doesn't need to be done. Others are probably doing negative work, that is, they are screwing things up, and still other people need to work to undo the disasters the first group created.
If you have two people doing negative work and two more undoing that negative work, you probably now need a fifth person as a line manager just to supervise the whole mess. Repeat this pattern a couple of times and now you need a second-level manager to manage the managers! Keep on going and now you have full-on middle management at the director level.
Meanwhile, the actual level of dog-walking app service hasn't improved. If anything, it probably degrades as the random inputs of the thousands of unnecessary workers distort things for their own needs. "Resume-driven development" kicks in. Odd things start being shipped right before the end of quarter or the end of the half so that people can "make their OKRs". The problem space is never "solved" because it's not in anyone's best interest to make it work and call it done.
The problem is, what else can we do? Those 99,900 other people still need jobs. They need food, shelter, healthcare, protection for their families, and all of that good stuff. You can't just kick them out just because they are unneeded in some hypothetical ideal situation.
Imagine, though, if those people were working at doing literally anything else, and preferably something they truly enjoyed at that. Imagine if it paid just as well, and wasn't obviously a "bullshit job".
It'd be great, right?