Information is inevitably tied to a physical representation, and therefore to all the possibilities and restrictions allowed by our real physical universe. The theory of computational limits is reviewed in a historical fashion. After some widespread initial errors, it was eventually understood that statistical mechanics and elementary quantum mechanics do not provide any limits. The energy requirements of the communications channel are particularly emphasized; it is an area where lower bounds, accepted for decades, are circumventable. The utility of the time-modulated potential going from monostability to bistability and back, is emphasized. Despite its use by von Neumann, Feynman, and many others, it has not received broad attention, i.e.by those not actually invoking it for their own purposes. I revisit my long-standing contention that our real universe does not permit the unlimited chain of infallible operations, envisioned in continuum mathematics, and that this has an influence on the ultimate nature of physical law. Finally, in the spirit of a volume dedicated to Richard Feynman’s impact, I deplore the strong effects of fashions in science.