It seems to me that the best way to capture Kant's intent and avoid these sort of computer programmer problems is to interpret the first categorical imperative as saying something like "always choose in such a way that your choice can universally rationally be regarded as valuable." This seems to make it possible to endorse some people being computer programmers while others farm, without so far as I can tell endorsing having some people tell the truth while others lie. It also, incidentally, helps slightly narrow the gap between the first and second categorical imperatives, which is an advantage as a matter of Kant exegesis since Kant insists the first and second (and third) CI are in fact the same. I'm not actually certain what this interpretation would say about the choice to be childless, but then one of the reasons I'm not a Kantian myself is that I think it's much harder than Kant pretended, indeed sometimes impossible, to figure out what the CI actually recommends. Only a few cases work out as neatly as Kant's preferred example of lying (and even that case is probably more difficult than Kant admits).