So these are your friends, it’s just on public policy you haven’t deliberated over dinner, and now you know. And then the next thing is that it shows it’s possible to converge. At the beginning of each interaction, often you see people clustering in four corners with very little overlap between them. But because the mechanism rewards the opinions that converge people’s views, it ranks to the top. So whereas people vote for the things with 70% consensus, then they would get competitive and think, "OK! I can prove something that wins 80% consensus," and once somebody has 80% someone would want to think "but I can do it better, I can do a 90% consensus point" and so because we reward the consensus, people spend their time thinking of how to convince people who host different positions, so we then get much more nuanced views out of this system.