Today, Jessica Birky (UCSD) found a weird M-type dwarf spectrum in the APOGEE data. We know it is a M dwarf, because Andrew Mann has an optical spectrum of it and physical parameters. But the APOGEE spectrum has features in all the wrong places; it looks nothing like any of the other M dwarfs we have. By the end of the day we started to suspect that it is a redshift issue, where the pipeline is assigning the wrong radial velocity and redshifting the data incorrectly.
I got in a discussion with Adrian Price-Whelan (Princeton) of an insane idea I had (while hiking) about Schwarzschild modeling: We could build a model out of not complete orbits but small orbit segments. We could deliberately make these segments with inconsistent gravitational potentials. And then when we model the data as a sum of segments, it would select the segments that fit the data best locally. Hence: A non-parametric model of the Milky Way potential, built up of locally fitting but globally inconsistent orbit segments! That's interesting. And probably intractable.
Lauren Anderson (Flatiron) showed me some examples of stellar streams where it appears that the Gaia DR2 RR Lyrae lie kinematically in the stream. That is interesting, because the RR Lyrae would deliver distance information for the streams.