Open Questions in High Dimensional Regression: - Inference for heterogeneous data that is different from the classical mixture models - Aggregating information across a large number of simple and not perfect machines - Data adaptive and model free inference - New views on the Efficiency considerations