Presented this talk at a school at Petropolis, Brazil. Lorentz covariance requires that gravity be velocity dependent. But old physicists can’t get over the 19th c. idea, reiterated by Eddington, that introducing velocity dependence in gravity makes the two-body problem unstable! They never studied functional differential equations, and no physics texts yet mentions my first solution of functional differential equations in a serious physical context. Old-time physicists confuse the mathematical theory of ordinary differential equations with physics. But young people do better.