3 ms·

However, are "your people" actually transitively closed? My impression is this isn't really the case for most people (or at least not if you like to talk about

by schoen 6y ago

However, are "your people" actually transitively closed? My impression is this isn't really the case for most people (or at least not if you like to talk about more than one thing). I'm reminded of this article I saw several years ago https://plausiblydeniable.com/five-geek-social-fallacies/ https://plausiblydeniable.com/five-geek-social-fallacies/ which suggests that we shouldn't assume friendship is transitive. I know I have friends whom I would never want to introduce to one another or have in the same conversation or social gathering! I agree with your suggestion that being able to ignore someone easily could make communications environments better; it seems like there was probably a lot of direct and indirect benefit in USENET killfiles (although I'm not such a fan of how people used them performatively, publicly announcing that they had killfiled other people -- usually by writing "PLONK!" at the end of a long rejoinder).

renewiltord 6y ago
No, you're absolutely right. The closure of anyone over friendship is likely all people. But I think there's a scoring mechanism that's possible using a graph model that can usually reveal whether you want to see a person's comments or not. For the moment, I just settle for the opposite bias: there are people who I want to read what I say, but I don't let them. Instead I just settle for a group that most of the time also feels that way.
- chaboud 6y ago
  "The closure of anyone over friendship is likely all people." Except Ted. Nobody likes Ted, and he has no friends. Yes, you, Ted. You know who you are. Joking aside, an aggregate measure of proximal association and affiliation is useful, but it is also likely fundamentally incomplete and incoherent over even a small subset of modes of affiliation (e.g. coders, windsurfers, rock climbers, pianists, frisbee players, race drivers, book collectors, DJs, etc). That said, a heuristic that generally improves the real probability of alignment, enjoyment and engagement is definitely possible. The most entertaining part of this is that, with Aggregability being an NP-Hard problem, there will probably always be room for improvement in this problem category. So game on.

by schoen 6y ago

renewiltord 6y ago

No, you're absolutely right. The closure of anyone over friendship is likely all people. But I think there's a scoring mechanism that's possible using a graph model that can usually reveal whether you want to see a person's comments or not. For the moment, I just settle for the opposite bias: there are people who I want to read what I say, but I don't let them. Instead I just settle for a group that most of the time also feels that way.

chaboud 6y ago
"The closure of anyone over friendship is likely all people." Except Ted. Nobody likes Ted, and he has no friends. Yes, you, Ted. You know who you are. Joking aside, an aggregate measure of proximal association and affiliation is useful, but it is also likely fundamentally incomplete and incoherent over even a small subset of modes of affiliation (e.g. coders, windsurfers, rock climbers, pianists, frisbee players, race drivers, book collectors, DJs, etc). That said, a heuristic that generally improves the real probability of alignment, enjoyment and engagement is definitely possible. The most entertaining part of this is that, with Aggregability being an NP-Hard problem, there will probably always be room for improvement in this problem category. So game on.