Cass Sunstein explains how 'enclaves and niches' develop on the internet, locking people into self-confirming communities of belief in which the information and the range of arguments they have to confront are limited and their viewpoints protected from serious challenge. It leads, Sunstein says, to a danger of 'enclave extremism'. One can acknowledge the existence of the tendencies he is identifying and still think that - as I've said before in debate with Oliver Kamm - it's rather too quick to see this as a serious problem if you haven't first marshalled some comparative evidence. In particular, we would need to know whether the enclave-producing tendencies that Sunstein writes about outweigh such counter-tendencies as are also part of the internet. I will mention two: (1) the ease of access to a much wider range of sources - newspapers, other media, blogs, etc - than anyone could possibly have, or have had, without the internet; (2) the culture of vigorous public argument and polemic that exists there. Both of these, I would argue, are likely to encourage greater exposure to views in some variety, including views opposed to one's own. Without a way of quantifying the relative weight of these tendencies and counter-tendencies, I don't think any firm conclusion can be drawn about the dangers of the internet, as compared with its beneficial consequences, for public discussion.