Hacker News new | past | comments | ask | show | jobs | submit login

For that, though, you ultimately want groups.

The problem with G+ Circles (or Diaspora* Aspects, which were an earlier version of the same thing) is that the people and profiles in your circles have no idea how you've classified them. One persistent argument in the early years of G+ was some Circler yelling at other a Circlee whom the Circler had classified invisibly to the Circlee that "they (the Circlee) were holding it wrong" --- not posting to G+ in the way in which the Circler had anticipated.

It turns out that what you actually really want are groups, not Circles.

(Preferably some kind of light-weight group with an easy join/quit dynamic and little overhead, but also robust moderation tools for larger cases, another aspect G+ never delivered on.)

I ranted about G+ failings for a long time, but ultimately reached a rather frustrated equanimity about it. At one point I commented to the effect that "It's a simple tool, designed for simple problems." That is, it lacked many features I and others would have liked to see.

One of the people +1'ing that particular post was Google+'s chief architect.

The best use I've found for Circles is to group profiles very roughly by interest level, usually into 2--4 tiers, from greatest to least interest. This permits following a fairly large group (though I prefer keeping even that limited) but without being overwhelmed by content. I've used that model on G+, Diaspora*, and Mastodon, pretty effectively.




I was more commenting on the general concept than a specific implementation. To be honest, while I tried to be optimistic about Google+ I never used it much and didn't dive into things like Circles hardly at all.

Certainly there can be a need to partition things. Today, I pretty much do it by using different social media for different purposes and, to the degree I blend some things, I keep it mostly uncontroversial and not overwhelming for those who may be only interested in some aspect.


Circles turned out to be immensely painful to manage, for numerous reasons. The easiest approach was often to simply nuke them all and start over.

Surprisingly, this was remarkably non-lossy. One of the more interesting episodes I had on G+ was when my primary profile was blocked (authentication issues) and I created a new one. In a fit of pique I named that "The Real Slim Shady", and started connecting up with a few key contacts. Within a few days I'd certainly managed to restore my core list of about 100 or so connections. Several people commented on how clear it was to them that this was in fact the same individual.

(I did confirm the association through several other sites, which helps.)

It also revealed the value of quashing even only a small number of sources of major noise.

(See my recent HN submission of Yonatan Zunger's "Privacy, Triage, and Risk" talk: https://news.ycombinator.com/item?id=29249630 https://yewtu.be/watch?v=9HQBCEW9odU https://www.usenix.org/conference/pepr19/presentation/zunger for the relationship between harm-reduction / risk mitigation vs. value-creation. The former is vastly more effective.)

This episode gave me a few insights on the nature of identity and trust. It also reminds me of how rapidly even massively-damaged cities and countries (war, natural disaster, etc.) tend to recover, especially if essential culture remains intact. Perhaps not entirely per their previous trajectory, but often with remarkably little long-term impact. This contrasts with trying to raise a specific region up out of poverty, a lack of institutions and infrastructure, and often a low-trust / high-corruption culture.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: