Risky Giant Steps Can Solve Optimization Problems Faster

This is about Benjamin Grimmer's paper https://arxiv.org/abs/2307.06324 where he proves under certain conditions that large steps lead to faster convergence.