@jim It sort of works for now because of the relatively small scale and lack of legal attacks, but as soon as some defamation lawyer or prosecutor or regulator gets interested those centrally controlled chokepoints will come back to bite (for example, the central appview and relay server can certainly be forced to stop distributing certain posts). Engagement is highly concentrated due to default recommendations, which increases risks. https://bnewbold.net/2024/atproto_progress/ https://anderegg.ca/2024/11/15/maybe-bluesky-has-won
I've been digging a bit on #bluesky moderation and the "firehose" approach to almagamating user posts, searchability and tagging.
On the one hand, compared to current closed models, it looks like an improvement as it opens up decision making and moderation. On the other, it looks to me at least like it is vulnerable to the same problems as all other centralised systems