Abstract: The goal of compressed sensing is to learn a structured signal $x$ from a limited number of noisy linear measurements $y \approx Ax$. In traditional compressed sensing, ``structure'' is represented by sparsity in some known basis. Inspired by the success of deep learning in modeling images, recent work starting with Bora et.al has instead considered structure to come from a generative model $G: \R^k \to \R^n$. In this paper, we prove results that (i)establish the difficulty of this task and show that existing bounds are tight and (ii) demonstrate that the latter task is a generalization of the former. First, we provide a lower bound matching the upper bound of Bora et.al. for compressed sensing from $L$-Lipschitz generative models $G$. In particular, there exists such a function that requires roughly $\Omega(k \log L)$ linear measurements for sparse recovery to be possible. This holds even for the more relaxed goal of \emph{nonuniform} recovery. Second, we show that generative models generalize sparsity as a representation of structure. In particular, we construct a ReLU-based neural network $G: \R^{k} \to \R^n$ with $O(1)$ layers and $O(n)$ activations per layer, such that the range of $G$ contains all $k$-sparse vectors.