A lively research field has recently emerged that uses experimental methods\nto probe the linguistic behavior of modern deep networks. While work in this\ntradition often reports intriguing results about the grammatical skills of deep\nnets, it is not clear what their implications for linguistic theorizing should\nbe. As a consequence, linguistically-oriented deep net analysis has had very\nlittle impact on linguistics at large. In this chapter, I suggest that deep\nnetworks should be treated as theories making explicit predictions about the\nacceptability of linguistic utterances. I argue that, if we overcome some\nobstacles standing in the way of seriously pursuing this idea, we will gain a\npowerful new theoretical tool, complementary to mainstream algebraic\napproaches.\n