This paper characterizes the optimal labor income taxes in an environment where individual labor supply choices are subject to adjustment frictions. Agents incur a fixed cost of adjusting their hours of work in response to changes in their idiosyncratic wages or their tax rates. This fixed cost can be thought of as the cost of searching for a new job in an economy where hours are constrained within the firm. I derive a formula that characterizes the optimal long-run progressive tax schedule in this economy. Adjustment frictions generate endogenously an extensive margin of labor supply conditional on participation. In addition to the standard intensive margin disincentive effects of taxes, the optimal schedule takes into account their effects on the option value of adjusting hours of work, and therefore depends on several new elasticities and marginal social welfare weights. I then evaluate the quantitative magnitude of these novel theoretical effects and show that for a given intensive margin labor supply elasticity, the optimal long-run tax schedule is less progressive than a frictionless model would predict, because an increase in progressivity raises the dispersion of individual incomes around their desired values. The welfare miscalculations by wrongly assuming a frictionless economy can be large, and are decreasing in the size of the intensive margin labor income elasticity. The insights of this paper apply more broadly to models where fixed costs interact with non-linear policy instruments to yield long-run aggregate effects.