The internal documentation details a messy and controversial process within Tumblr itself. One internal post made by Cyle Gage, a product manager at Tumblr, states that a query made to prepare data for OpenAl and Midjourney compiled a huge number of user posts that it wasn't supposed to. It is not clear from Gage's post whether this data has already been sent to OpenAl and Midjourney, or whether Gage was detailing a process for scrubbing the data before it was to be sent. Gage wrote: "the way the data was queried for the initial data dump to Midjourney/OpenAl means we compiled a list of all tumblr's public post content between 2014 and 2023, but also unfortunately it included, and should not have included:
private posts on public blogs
posts on deleted or suspended blogs
unanswered asks (normally these are not public until they're answered) private answers (these only show up to the receiver and are not public)
posts that are marked 'explicit' / NSFW / 'mature' by our more modern standards (this may not be a big deal, I don't know)
content from premium partner blogs (special brand blogs like Apple's former music blog, for example, who spent money with us on an ad campaign) that may have creative that doesn't belong to us, and we don't have the rights to share with this-parties; this one is kinda unknown to me, what deals are in place historically and what they should prevent us from doing."
Gage's post makes clear that engineers are working on compiling a list of post IDs that should not have been included, and that password-protected posts, DMs, and media flagged as CSAM and other community guidelines violations were not included.
Second image: the Visibility section of the tumblr blog settings. Each item has an on/off switch. Items:
Hide evilwizard from people without an account
Discourage external searching of evilwizard
Exclude evilwizard from Tumblr search
Prevent third-party sharing for evilwizard
Number 4 is circled. End ID.]
germansingergirl
turnnoffyourmind
vaporwave
stormborns
theartofmadeline
