Good pont. Still that's more or less assuming that the costs and returns of D and T are roughly similar. But talk is cheap, and it is not clear that the return

Good pont. Still that's more or less assuming that the costs and returns of D and T are roughly similar. But talk is cheap, and it is not clear that the returns earned are the same. if we are intent on maximizing u(d, t) = u(d) * u(t) we still need u(d) and u(t) to know what quantities of d and t will do the job (assuming that there is a maximum, or not more than one...)