Internal documents obtained by 404 Media show that Tumblr staff compiled users' data as part of a deal with Midjourney and OpenAI.
According to this article, they're discussing an opt-out model of harvesting data, which is definitively worse than an opt-in model (because a lot of people are not going to go through the logistical hurdles to avoid their stuff getting stolen) and very much a bandaid on a open wound.
Needless to say, everyone should be screaming about this to @staff because the more friction the better, and backing everything up if they need to shut shit down rather than have their work be "legally" stolen by a bunch of Silicon Valley vultures who are all going to lose their shirts in a couple years.
Not exactly sure what Plan B I'm going to go with yet, but needless to say my interest in getting an independently hosted wordpress.org blog (because that's distinct from wordpress.com) and/or checking out Squarespace or the like has just gone way up.