I asked o3 to summarize the paper:

* Tracked about 276 000 academics from 2023-2025.

* Roughly 1 in 5 opened a Bluesky account and stopped tweeting; split by field: ~31 % in arts & humanities vs ~13 % in medicine.

* Biggest predictor is whether the accounts you read migrate. If 10 % of your followees leave, your own chance of quitting jumps about ten-fold compared with losing the same share of followers.

* Five flashpoints caused most exits:

1. July 2023 global X outage

2. July 2023 “rate-limit” fee talk

3. Feb 2024 Bluesky public launch

4. Aug 2024 one-day X block in Brazil

5. Post-election rush after the Nov 2024 US vote

* Re-creating your old Twitter reading list on Bluesky makes you stay active; failing to do so means you often drift away.

* Political tweeters leave in a steady trickle; non-political scholars mainly bolt during the big shocks.

* Citation counts, h-index, and “top-100 author” status have no effect on who quits.

TL;DR: Academics move when the voices they rely on disappear—especially during Twitter/X hiccups—and they stick with Bluesky only if they can rebuild that feed.

Is there an academic vibe there, where people are talking about their research etc? That sounds like something I would check out. If its just academics spouting of politics, it sounds like the worst network in the world.