After using Cursor a while now, I feel like the limit on fast requests could be measured in token usage instead of just a flat number of requests? Surely some operations are much more costly then others? I am aware that it could also backfire the other way around.
But it was just a thought, I let you guys crunch all the numbers 