Varies a lot by country. In the U.S., most PhD students do TA/RA duties, because that's the main funding mechanism. It's rare for universities to give direct funding for an entire PhD, though some do fund students' first year directly. A smallish number of students also get multi-year fellowships, either from the university or from an external source like the NSF Graduate Research Fellowship Program. The common case, though, is that each semester they're hired as either a TA or RA. In the best case, the student's advisor has a grant precisely aligned with their thesis topic, and the student is hired as an RA on that grant, in which case they're being paid more or less just to do their PhD. If they're less lucky, they have to work on a grant-funded project that's more tangential to their thesis, and then essentially have a side job.
On the professor's side this system is also a pretty big source of stress, because every semester there's a question of where your students are going to be funded from the next semester. Have to bring in enough grants to hire most as RAs (ideally on projects that lead to progress on their theses), rotate some into TA positions if there aren't enough grants to pay everyone as an RA this semester, etc.
In a different form RAships are also becoming common here in Denmark, though they weren't traditionally. PhD students are all hired into a 3-year salaried contract, but increasingly those contracts are funded by European Union projects, and a student hired into a PhD position funded from that source will typically have project-related responsibilities in addition to their thesis.
The 3 papers makes a thesis also ignores the fact that it needs to be a coherent whole and pass the thesis defence, where the committee includes both a person external to the department and one external to the university. While I'll grant 3 papers is probably a thesis, it entirely depends on the papers.
Everybody in STEM courses at my university did both an RA and a TA unless they had a large scholarship and didn't need any extra money. In the CS department, it was actually expected (I was co-supervised by a CS prof for my Master's so I know what the other people in my lab did).
I had a girlfriend who did her PhD in Mechanical Engineering and she was actually required to do contract work in her lab. She couldn't even get paid for it and was still required to do it. She was in the Canadian military and they were paying for her to do the doctorate so she couldn't get paid from anyone else (but she had a scholarship which was allowed). The contract work delayed her thesis research, but the professor required her to do the work. So, you're extrapolating your knowledge and assuming it's the same everywhere which isn't true.
>They already get a low but liveable salary without having to pay tuition. I don't think many PhDs do TA/RA duties - they're too busy researching.
In most places I've seen, PhD students are the primary source of TA/RA labor. If we're lucky, we get a base stipend and TA/RA adds to it.
>PhD students don't have grades or qualifying exams
I've never seen a PhD program with no coursework, no qualifying exams, and no grade requirements for either. Could you show me where such an arcadian academic track exists?
>and three publishable research papers is what most PhD students achieve anyway
That was my point, yes.
To begin my PhD I made contact from cold with a professor who had similar research interests to mine at the University I wanted to go to. We did an informal interview and he offered me the studentship. I didn't go through any kind of central application process, do any entrance exams or write any essays or anything.
When I started my PhD I immediately began working on my own research and writing my own papers. I never did any coursework or exams. My only assessment is a yearly review presentation and then the thesis and viva at the end. I'm not involved in any kind of group project so I'm not working for anyone else as 'cheap labour'.
I get paid by a grant from the government of around $21k a year. That's tax-free so I guess it's maybe the equivalent of $25k. I have done about 50 paid hours of TAing (demonstrating) during three years, but it was optional and I did it just to meet some new masters students really.
PhD students don't have grades or qualifying exams, and three publishable research papers is what most PhD students achieve anyway. The thesis is just these papers written up into one continuous piece and submitted to the people who will become their peers.