Read a Twitter thread by Mx. Aria Stewart (Twitter)

It just crystallized for me what I think has been mistaken about thinking of unwanted interaction on social networks as a "privacy" problem. It's not.

A privacy problem is things becoming known more widely than they should, subject to surveillance and contextless scrutiny. 

The onslaught of sexual harassment on platforms like early Twitter (and later twitter for people of notability), @KeybaseIO, every naive social network is an attack on the right to exist in public. It is the inverse of a privacy problem. 

But the conceiving of this as a privacy problem brings the wrong solutions. It means we are offered tools to remove ourselves from public view, to restrict our public personas, to retreat from public life. It means women are again confined to private sphere, denied civic life. 

 It's so endemic, so entrenched, and so normal that women should have to retreat to protect ourselves that we think of this as part of femininity. A strong civic life is seen as unfeminine, forward. It poisons us politically, socially, and personally. 

It is, at its core, an attack on democracy as well. 

The only way to undo this is to reconceive of this, not as a privacy problem but as an attack on public life. There will be new problems with this but at least they will be new. 

There has been work done on this, but I've never seen it connected to civic life, and this connects with my thoughts and work on community. The unit that social networks must focus on cannot be the individual. We do not exist as individuals first but as members of our communities 

When a new user joins a social network, their connection must be to their peers, their existing social relationships. A new user can only be onboarded in the context of relationships already on the network. 

Early adopters form such a community, but extrapolating from the joining of those initial members to how to scale the network misses the critical transition: from no community to the first, not from the first users to the next. 

New communities can only be onboarded by connections from individuals that span communities. New communities must be onboarded collectively, or the network falls to the army of randos. 

The irony is that surveillance capitalism has the information to do this but not the will, because as objects of marketing, we are individuals, statistics and demographics, not communities. The reality lies in plain sight. 

There have been attempts at social networks, sadly none dense enough to succeed, but that treat people as part of a web, and that their peers can shield and protect them. The idea is solid. 

The other alternative is to stop trying to give people a solitary identity, a profile and onboarding to a flat network, but instead only provide them with community connections. Dreamwidth is this to a large degree, if too sparse for most people to connect. 

Our social networks must connect us, not to our "friends" but to our communities. The ones that succeed do this by intent or by accident.

Facebook has a narrow view of community, but for those it matches, it works. With major flaws, but it does. 

Twitter, its community of early adopters, its creepy onboarding by uploading your contacts and mining data to connect you works. If I were to join and follow a few people I know, it would rapidly suggest many more people in my queer and trans community. It works. 

And this is why Ello failed. This is why Diaspora failed. This is why Mastodon succeeded, if only by scraping by the bare minimum. This is why gnu social failed. This is why a random vbulletin forum can succeed. The ones that succeed connect a dense community. 

Note that gnu social and mastodon are the same protocol! But they are different social networks. The difference in their affordances and the community structures they encourage are vastly different, despite interoperating. 

I'd say I don't know how apparent this problem is to white men — the ones largely designing these networks — but I do know. I know because of the predictable failures we see.

Part of this, I think boils down to how invisible community is when you are the default user. 

At no time am I unaware that I am trans, that I am a woman, that the people I follow and who follow me are distinct from the background. I can spot my people in a crowd on the internet with precision, just like a KNN clustering can. 

Trans culture in particular is Extremely Online. We are exceptionally easy to onboard to a new platform. But the solution can scale if we focus on solving it. And by knowing who is in the community (likely) and who is not, we can understand what is and is not harassment. 

We don't need to even know what the communities are — Twitter does not — and yet it knows how we cluster, and that suffices.

If we stop thinking of this as a privacy problem — letting us hide from the connections that are our solution — we can enlarge public life. 

That exceptional article —

That Uplifting Tweet You Just Shared? A Russian Troll Sent ItMusic, Film, TV and Political News Coveragehttps://www.rollingstone.com/politics/politics-features/russia-troll-2020-election-interference-twitter-916482/

— about how bots sow division shows us another facet of this problem and way of thinking. Conceiving of this as a privacy problem fundamentally reacts with division when solidarity is needed. 

We can only fight this with a new, loose solidarity and an awareness of community boundaries. We can build technology that makes space for us to be safe online by being present with those that support us, and react together, rather than as individuals and separating us for safety 

This thread has meandered a bit, but I'm dancing around something important. We fundamentally need to stop organizing online activity the way we do. Follow and be followed is not where it's at.

It's join, manage attention, build connection. 

Stop sorting things topically and trying to find connections in content.

Start looking for clusters of relationships between people.

The question should not be "what is this about?" but "who is this for?"