The fact that the person doing the job thinks so doesn't necessarily make it so. There's plenty of people who don't see any real purpose in safety belts in cars, too.
I agree with your second statement, but your first statement doesn't necessarily follow, it depends on the proportion of jobs that are bullshit....
If a human is 1% likely to think a not-bullshit-job is bullshit, and 10% likely to think a bullshit-job is bullshit, but 99% of jobs are not-bullshit, then for every bullshit job that the worker thinks is bullshit, there are 9.9 not-bullshit jobs that the worker thinks are bullshit.
Your hypothetical would lead to roughly 98% thinking that their jobs are not bullshit when 99% actually were. At that level we wouldn't be having this discussion at all.
The actual numbers are more like 60% non bullshit / 40% bullshit, though.
I bet everyone in tech who has been around for a while has been in enough of _those_ meetings to be able to at least relate to what a bullshit job it.
Most probably have also seen the type of manager that is less concerned with the work people underneath them do (because they don't understand or care) and more with the prestige, making sure people stay there, even if they contribute nothing - either because they can't or wouldn't. Worse, eventually the former turn into the later with enough time to resign to the situation.
There's also an insidious pattern where entire layers of management exist strictly to report status to higher layers of management.
It's challenging in such organizations to have meaningful cross-functional communication between the people who do the actual work and who need to understand each other. Why? Because these management layers also [try to] mediate all cross-functional interaction in formalized meetings which they run like court proceedings-- they're called by various names: "hand-offs", "gates", etc. The point is they're one-direction process-flow "ceremonies" that don't have real discussion between the parties involved.
What has to happen to workaround this is that the individuals who need to really communicate must seek each other out informally (or "offline" as the PM-derps would say), sometimes this works, sometimes it's a setup for one party getting snubbed by the other because they're beset with too many meetings.
So yeah, there's often of fat layer of middle management that do bullshit jobs in large-enough orgs.
For years, I've made slides to report status on things (as a line manager, not a project manager) and I've tried to figure out ways to automate it. No amount of tags or epics or whatever in Jira seem to do it. It always has to be done by hand because what's actually happening is a lot of editorial curation, contextualization with other things going on at the company or in the industry, etc. (Please, if someone has a better idea, let me know!)
At larger companies, I can imagine this semi-journalism taking lots of people, because you can't just aim a firehouse of information at upper management. In theory, this could all be replaced with some clever dashboards, but the reality always seems messier.
> There's also an insidious pattern where entire layers of management exist strictly to report status to higher layers of management
I think everyone agrees.
> So yeah, there's often of fat layer of middle management that do bullshit jobs in large-enough orgs.
Here we disagree. You are positing that it's possible for there to be a large org without this fat layer, and I think this is unlikely. It's like complaining about entropy or the n^2 communication cost. Something may be terrible yet impossible to avoid. Making big bureaucracies as efficient as small bureaucracies is a really, really, hard problem. It may well be an unsolvable problem that comes with scale, sort of the flipside of how many things are more efficiently produced in large quantities, and it could be that the "optimal" size of a business is where these two trade-offs balance out -- e.g. where the scaling costs of the bureaucracy meets the efficiency gains of scale production.
