As so many other activities nowadays, modern science revolves around the competition for attention.Unlike in so many other attention games, in science those who seek attention are more or less the same people who are giving it. An important characteristic is the skewness of the distribution of scienti®c attention. We discuss the e ect these characteristics have on scienti®c institutions. An important thesis of ours is that scientists converge in clusters of likeminded scientists. Given the character of scienti®c organization and communication we expect that the digitalization of scienti®c communication will not a ect the basic scienti®c institutions as the principles upon which the Internet functions coincide more or less with the way science functions.However, violation of these principles can in principle disrupt science and fundamentally change its character.Diversity, the key element of scienti®c conversation,may be destroyed.