Hsieh and Moretti (2019) find that relaxing land use regulations in three productive US cities would increase GDP by 3.7 percent. In this comment, I revisit their findings. I first attempt to replicate their result and find that their counterfactual would lower output. I document errors in their code that explain this discrepancy. I next show that the results of their model depend on the arbitrary choice of population unit. I propose a modification to their model that eliminates unit dependence. Their experiment raises output in the modified model, but the effect is two orders of magnitude smaller than what they report. (JEL E23, J24, J31, R23, R31)