For decades, scientists have debated whether supercooled liquids stop flowing below a glass transition temperature T$_{g0}$ or whether motion continues to slow gradually down to zero temperature. Answering this question is challenging because human time scales set a limit on the largest measurable viscosity, and available data are equally well fit to models with opposite conclusions. Here, we use short simulations to determine the nonequilibrium shear response of a typical glass-former, squalane. Fits of the data to an Eyring model allow us to extrapolate predictions for the equilibrium Newtonian viscosity ηN over a range of pressures and temperatures that change ηN by 25 orders of magnitude. The results agree with the unusually large set of equilibrium and nonequilibrium experiments on squalane and extend them to higher ηN . Studies at different pressures and temperatures are inconsistent with a diverging viscosity at finite temperature. At all pressures, the predicted viscosity becomes Arrhenius with a single temperature-independent activation barrier at low temperatures and high viscosities (ηN>10$^3$ Pa⋅ s). Possible experimental tests of our results are outlined.