1: \begin{abstract}

2: We study the oracle complexity of nonsmooth nonconvex optimization, with the algorithm assumed to have access only to local function information.

3: It has been shown by Davis, Drusvyatskiy, and Jiang (2023) that for nonsmooth Lipschitz functions satisfying certain regularity and strictness conditions, perturbed gradient descent converges to local minimizers \emph{asymptotically}. 

4: Motivated by this result and by other recent algorithmic advances in nonconvex nonsmooth optimization concerning Goldstein stationarity, we consider the question of obtaining a non-asymptotic rate of convergence to local minima for this problem class. 

5: 

6: We provide the following negative answer to this question: Local algorithms acting on regular Lipschitz functions \emph{cannot}, in the worst case, provide meaningful local guarantees in terms of function value in sub-exponential time, even when all near-stationary points are global minima.

7: This sharply contrasts with the smooth setting, for which it is well-known that standard gradient

8: methods can do so in a dimension-independent rate.

9: Our result complements the rich body of work in the 

10: theoretical computer science

11: literature

12: that provide hardness results conditional on conjectures such as $\mathsf{P}\neq\mathsf{NP}$

13: or

14: cryptographic assumptions, in that ours holds unconditional of any such assumptions.

15: \end{abstract}

16: