Fixing Twitter

According to JestersCourt, Twitter can fix Twitter with just a few lines of code

I don’t have inside knowledge of Twitter’s architecture, and so can’t really comment on whether a “few lines of code” is accurate, but the essence is this: When someone blocks you for whatever reason, you can’t @mention them in a Tweet. Whether you get an error, or it’s just silently deleted, the Tweet will go nowhere and won’t be seen by your followers.

That fixes the biggest single aspect of Twitter’s harassment problem, when someone with a large bully pulpit sets their followers on someone who’s incurred their wrath.

Unless there’s a flaw in the solution I haven’t see, it’s hard to see why Twitter doesn’t just go ahead in implement something along these lines. You’re forced to suspect that Jack Dorsey is less interested in solving Twitter’s actual problems than he is currying favour with particular activist cliques.  In other words, virtual signalling trumps positive action.

The problem with Jester’s Court’s solution is it’s politically neutral. The same mechanism that would stop racist and sexist trolls would also damp down the witch hunts popular in social-justice circles. And because that would cramp the style of the people Jack Dorsey wants to curry favour with, it’s a non-starter; they have a weird “punching up/punching down” dynamic where it’s only defined as harassment if it’s a member of their outgroup targetting a member of their ingroup.

So instead Twitter seem to be going down the route of top-down content politicng and filtering by keywords known to be popular with outgroups. What could possibly go wrong?

This entry was posted in Social Media and tagged . Bookmark the permalink.

4 Responses to Fixing Twitter

  1. The other thing that annoys me about this, political implications aside, is that they’re using a technical solution to the problem which is a) far more complex, b) requires constant management (you need to tune the keyword list over time for it to work), and c) isn’t going to work anyway because filtering by keywords has proven to be uniformly worthless for managing Twitter content from any point of view!

  2. Tim Hall says:

    And filtering by keyword takes no account of context, and nobody had yet written an AI that’s smart enough understand whether somebody using a naughty word like “Cuck” is a member of the Alt-Right, or a Social Justice Warrior attacking the Alt-Right.

  3. Exactly. I had this problem when trying (and failing) to use keywords to filter my Twitter streams in the early days of my use of the service. I finally just gave up. Twitter isn’t really a stream-based medium; it’s more of a pool that’s added to by way of streams, and you dig into the pool and pull out stuff every so often. Following streams religiously is pointless.

  4. Tim Hall says:

    Yep, Twitter is something you dip in and out of. The way it threads conversations mean you can backtrack up to follow interesting conversations where you missed the start.

    Sometime it does feel as though Twitter is slowly dying at the hands of clueless management who don’t understand how and why people use their product.