Paper proving black-box PIR with preprocessing still hits lower bounds, so the free lunch is, naturally, imaginary.

Paper proves the Gaussian binary tree mechanism is optimal for approximate-DP continual counting, after the lower bound finally shows up.

Sparse structure still doesn't save private PCA here, polynomial-in-d samples can survive the sparsity assumptions.