The same team wrote another interesting paper arguing that there's no "double descent" in linear regression, trees, and boosting, despite what many argued befor

The same team wrote another interesting paper arguing that there's no "double descent" in linear regression, trees, and boosting, despite what many argued before (in this paper they don't tackle deep learning double descent, but remark that the case may be similar regarding the existence of different complexity measures being conflated).