The disconnect here is that while the "job" of a grad student is to do things like research and teaching, it's not the tuition part that is paying for that "job

The disconnect here is that while the "job" of a grad student is to do things like research and teaching, it's not the tuition part that is paying for that "job", it's the stipend. The stipend is already taxed. Put another way, one could reasonably say that an entry level trainee in science is worth about $30K a year pre-tax (roughly the grad stipend at Stanford while I was there). Maybe they're even worth $50K, which is roughly what can be gotten in industry as an undergrad fresh out of college in the biosciences. They are definitely not worth $70-$80K a year (stipend + tuition at a place like Stanford), and even so we would be taxing them on a salary they're not actually seeing. It's much better to see these waivers as scholarships, which generally aren't taxable. I don't know how the tax plan treats scholarships or financial aid, but I hope they haven't decided that those are taxable too. If they have decided scholarships still aren't taxable then I suppose universities could call everything a scholarship and use that as a loophole. In which case this part of the tax bill is pointless, incurring a massive administrative change for zero benefit of any sort. Also some universities, particularly public universities may have regulatory issues around limits to number of scholarships they can issue or how they can decide to issue scholarships.