The qualification is that even the person doing the job thinks it's bullshit. That doesn't mean the definition relies solely on them thinking it's bullshit.
Frankly, reading through the article, it feels like the definition keeps shifting and much of it is his struggle to even define what it means. He then goes on to start giving his definitions of the what he thinks are types of bullshit jobs.
The article I don't think is all that consistent and shouldn't be taken too seriously.
Exactly. Like many of us, Graeber writes to understand.
His stated purpose is abundantly clear. Many, many people were sharing with Graeber complaints of a similar theme. So he tried to name and describe it.
Rightly, wrongly, or otherwise, it's all in the book. For other people to pick apart and response to.
How else is scholarship supposed to function?
Would Graeber have been more effective as an op ed columnist, churning out saccharine blather which confirms everyone's priors? eg The World is Flat and Bobos in Paradise?
I think there are definitely jobs that a computer or robot should be doing, but a human is there instead because of what our economy and society currently values. However, until the computer or robot is actually doing the job, I don't think I'd call it a bullshit job. That's not to say society shouldn't consider the impacts of grinding people's lives away with rote repetitive tasks.