Search Talks

In this talk, I describe some very simple but significant phenomena predicted by quantum theory. These are described simply in terms of what is observed in the laboratory, without making any presumptions about what sort of microscopic picture of reality might account for these observations. With these phenomena in hand, I demonstrate two simple applications of quantum theory to cryptography: how to build counterfeit-proof money and how to detect eavesdroppers on a channel (and thereby distribute a secret key which can be used to encode messages in a way that cannot be deciphered by one who does not have the key). Moving from quantum information theory to quantum foundations, I show how the qualitative features of these phenomena can be reproduced in a toy theory wherein systems have well-defined properties but observers can only come to know a limited amount about them. This shows the merits of the notion of \"hidden variables\" underlying quantum theory. Finally, I describe a phenomenon that is predicted by quantum theory but that *cannot* be reproduced by a natural class of hidden variable models, namely, those that are *local* in the sense that changes in one region cannot instantaneously affect the state of affairs in another. The phenomenon in question is the existence of certain strange correlations between the measurement outcomes on distant systems. It is illustrated in terms of a two-party game that is played out by a few lucky members of the audience.