The foreseeable impacts of climate change on humans depend critically on the ability of societies to adapt to new climatic signals. Recent literature based on the US experience suggest little to no adaptation to climate trends in crop agriculture. We revisit this question with a novel panel econometrics framework that allows yearly outcomes to jointly respond to contemporaneous weather and climate signals, delivering estimates of both short-run and long-run climate change impacts within a single regression framework. In our most conservative model, which strictly relies on the existence of trends in climate to identify adaptation potential, we find evidence of adaptation to beneficial temperatures (as captured by growing degree days) and precipitation, for both corn and soybean yields. Evidence of adaptation to extreme heat is more nuanced, with suggestive evidence of adaptation for corn and no evidence for soybeans. The presence of adaptation matters for the calculation of warming impacts: the net counterfactual impacts of warming are negative in the short run but become positive in the long run.