Kris gives an update on his work on low frequency foregrounds
unpublished evidence from Quijote for significant variation of synchrotron beta
Planck data for \ell>100 is dominated by noise, so we don't know the small scale structure
Kris has a prescription for filling in small scale structure in a non-gaussian way; reminder: PySM is gaussian fill-in, so the new procedure is interesting, but Kris is not yet saying how the filling is done.
Brandon mentions that people (e.g. Clark) suggest using HI data. Kris says he is not using HI data.
Discussion of \ell relevance: agreement that this is all relevant for lensing reconstruction at high \ell, not for low \ell data. MR points out that there is an indirect effect: averaging of small scales to larger patches will impact the large angular scale data as well.
Clem understands the reservations HK has regarding the signal. Simulations will need to be done with higher resolution. He is more puzzled about the noise issues.