Having G and R separated into two people can be quite nice though[...]
This part seems dubious to me since the tasks done by G and R often share a lot of conceptual work. e.g. finding a counterexample to a given claim might naturally suggest a related claim which is plausibly true. Mathematical collaboration can still be useful but I think usually both parties do a bit of G and a bit of R.
I agree. I was mostly making the claim that on small time-scales (e.g., one hour at a whiteboard) it can make sense to have one G and one R (although you shouldn’t be too rigid about it imo).
This part seems dubious to me since the tasks done by G and R often share a lot of conceptual work. e.g. finding a counterexample to a given claim might naturally suggest a related claim which is plausibly true. Mathematical collaboration can still be useful but I think usually both parties do a bit of G and a bit of R.
I agree. I was mostly making the claim that on small time-scales (e.g., one hour at a whiteboard) it can make sense to have one G and one R (although you shouldn’t be too rigid about it imo).