This paper uses smartphone GPS data to study the impact of highways on intracity non-work travel within the Seattle MSA. Using a discrete choice framework, I show that destinations that are accessed by traveling on or across a highway are visited 17% less than other destinations, ceteris paribus. This effect is non-linear in travel distance and is especially significant for short trips by travelers in urbanized zones. I then quantify the welfare effects of two counterfactual scenarios with alternative highway systems using a quantitative urban model that incorporates non-work travel. I find that welfare increases by 10.2% if the urban highways are replaced with an underground system and by 9.0% if they are replaced with primary surface roads. For both exercises, residents in the urban core benefit from amenity improvements, albeit at the cost of reduced amenities in suburban zones.