The paper re-examines Ricardo's argument on machinery, criticizes and re-uses Samuelson's recent interpretation, and produces a new reconstruction of the argument. It separates long-run effects of inventions from short-run effects of capital conversion, concluding that both viewpoints are correct. Technological unemployment and changes in technical coefficients are not the main focus of Ricardo's argument. The conclusions are based on classical distinctions between value and wealth, and between free capital and invested capital.