What does S4 get? S4 paper doesn't have model 98. JD pulls out results of internal communication. The results for S4 are also biased.
For r=0; still need to optimize patch. Perhaps work with JD?
Delensing: Sebastian needs clean maps from MR. They need simulations to estimate the mean field, and they need the E-field. Until now MR hasn't generated the E-mode. He needs to start generating the E-mode.