Dima - a related point is that both \Delta and \Nabla are extensions of the dualizing sheaf on S^-, as opposed to the constant sheaf, because the latter is only a pro-object
Related to the “tensor !”, I take it. (Yes, I realize that with ind-schemes the symmetry is already broken)
After my talk last week, I wished I had been clearer that defining the A/B-twists of T[G] as a 3d theory is okay. The hard part is seeing the G-symmetry, which is really an issue on the B-twist.