the last cecilia

cecilia, elderly (33), they/them

[staff]

Hi, Tumblr. It’s Tumblr. We’re working on some things that we want to share with you. 

AI companies are acquiring content across the internet for a variety of purposes in all sorts of ways. There are currently very few regulations giving individuals control over how their content is used by AI platforms. Proposed regulations around the world, like the European Union’s AI Act, would give individuals more control over whether and how their content is utilized by this emerging technology. We support this right regardless of geographic location, so we’re releasing a toggle to opt out of sharing content from your public blogs with third parties, including AI platforms that use this content for model training. We’re also working with partners to ensure you have as much control as possible regarding what content is used.

Here are the important details:

We already discourage AI crawlers from gathering content from Tumblr and will continue to do so, save for those with which we partner. 

We want to represent all of you on Tumblr and ensure that protections are in place for how your content is used. We are committed to making sure our partners respect those decisions.

To opt out of sharing your public blogs’ content with third parties, visit each of your public blogs’ blog settings via the web interface and toggle on the “Prevent third-party sharing” option. 

For instructions on how to opt out using the latest version of the app, please visit this Help Center doc. 

Please note: If you’ve already chosen to discourage search crawling of your blog in your settings, we’ve automatically enabled the “Prevent third-party sharing” option.

If you have concerns, please read through the Help Center doc linked above and contact us via Support if you still have questions.

[slowedmountains]

by the way also, since staff's instructions aren't especially clear it's here in your blog settings (on browser select your blog and then blog settings on the right)

The screen looks like this:

then scroll aaaall the way to the bottom:

This does not apply universally across your account, you need to do this for each of your Tumblr Blogs individually.

[fromthebaronies]

Hobbit Day is a lovely day to celebrate a year on testosterone

[fromthebaronies]

What if I just went back to pretending I don't exist?

[milkovichy]

#happy dickfuck day to those who celebrate it

[nudityandnerdery]

Good news for you, this August 23rd.

[nudityandnerdery]

It’s that time of year again.

[funkittoys]

Sliding Scale NoFrillDo | funkittoys

One of Funkit Toys' main guiding principles is minimizing waste. Another is making safe sex toys accessible to as many people as possible. W

Pay-what-you-can NoFrillDos are temporarily back in stock. These only restock a couple times a year, as they're made of leftover silicone from other pours. Limit one per customer.

I make these specifically because I know even NoFrillDos are out of reach for some. They forward my ideals of minimizing material waste and maximizing financial accessibility, a perfect distillation of Funkit company values. Which are my values, since it's just me here.

[nacho-joe-yo-deactivated2022012]

Monster fucker this, monster fucker that. What if I want a monster RELATIONSHIP huh?! Monster HAND HOLDING, monster INTIMATE CONVERSATIONS, monster COMFORTABLE SILENCE??

[assassinationtipsforladies]

[fromthebaronies]

Almost 10 months on t

[chocmartini-deactivated20210419]

might flop but….. rb with ur sign & what kind of drunk you are

[girlhorrror]

WHY are you pitting vampires and werewolves against each other they should be having lesbian sex

[dispatchesfromtheclasswar]

[fromthebaronies]

Spotify, what the hell

[lgballt]

Are you sad that June is over and you don't have a pride month anymore? Fear not, friends! There's a different pride month just beginning! lgballt

[fromthebaronies]

Just got off the phone with the pharmacy again. Still no idea why my testosterone prescription refill has been denied or what happens next. The appeal could take 30 days, I guess, except it's been almost that long since I didn't get a refill in the first place. I think I have 20 doses left.

I haven't had a period since last fall, which means it's been about that long since I last really thought about killing myself. Not to be dramatic, but this is the longest stretch of, like, feeling normal? Since, oh, 1999?

[borntoruns]

Favorite Albums: Born in the U.S.A. Bruce Springsteen • 1984

Born down in a dead man's town The first kick I took was when I hit the ground You end up like a dog that's been beat too much Till you spend half your life just covering up...

[trentcrimms]

James Lance as Trent Crimm TED LASSO | 3.06

[griseldajane]

Glaze is out!

Tired of having your artwork used for AI training but find watermarks dismaying and ineffective?

Well check this out! Software that makes your Art look messed up to training AIs and unusable in a data set but nearly unchanged to human eyes.

I just learned about this. It's in Beta. Please read all the information before using.

[digitaldiscipline]

Art thieves already hate it:

[fierceawakening]

Dude, if you're stealing, you deserve to have the data poisoned. Because you could have asked and you didn't.

[spooky-octagon-deactivated20240]

The link is only in the original post inside an image, not as text, so here it is as plain text: https://glaze.cs.uchicago.edu/ and the paper about how it works: https://arxiv.org/abs/2302.04222

[esoanem]

As links (because some of us are on mobile and can't easily copy and paste to our browser), those are:

https://glaze.cs.uchicago.edu

&

GLAZE: Protecting Artists from Style Mimicry by Text-to-Image Models

Recent text-to-image diffusion models such as MidJourney and Stable Diffusion threaten to displace many in the professional artist community

[sliceosunshine]

A bit of a TLDR for some questions I saw in the notes:

The team that created Glaze is from the University of Chicago. Their names are each listed in full on the Glaze download website. (This group of students/professors did this for their SPRING BREAK 😱 so go give them some love lol)

It is free to download. No, they won’t ask for or raise money from/for this project.(stated by one of the lead professors of the project).

Glaze is designed to protect artists’ STYLE--which a bunch of ai people have been deliberately fine-tuning their models to mimic (and specifically of current living artists--small or big).

It currently does not protect against composition/trace-like theft (as seen when run through img-to-img) but that would be protected by copyright anyway while STYLE is not.

The University Team has stated that they are dedicated to continuing to improve the tool, like fixing bugs (like overheating older computers by taking up lots of energy when Glazing--it currently runs on CPU so they’re trying to change that to GPU, I believe) and expanding the type of protection given to artists (like working against img-to-img theft).

It currently only works directly on your computer (phones not advised due to current overheating issue, no tablets, or iPads, and no website runthrough since that would be insecure to breaches/scraping/hacks)

It currently works best on painterly artwork, but can still be used on other forms (team is working on improving this)

IT WORKS BY calculating the changes each image needs for the best protection against style theft by AI, and adds tiny changes throughout the piece, so that your style will, for example, confuse the ai into seeing van gogh. But the ai thieves will see a regular image in your style, feeding it into their model labeled as your work (thus starting the “data poisoning”).

Do not post the original unGlazed piece of your artwork after posting your Glazed version (obviously)

The Team worked directly with over 1,000 artists that were being impacted by the ai theft. Because the team listened to those artists, Glaze accounts for regular art thieves too (i.e. Glaze can’t be removed/cropped etc. like signatures or watermarks when reposted. It’s just part of the image, so even if it ends up on another site and scraped, the Glazing is still in effect)

When you run your artwork through Glaze, no information is sent back to the Team. (Aka, no scraping on their part. The app receives information from the Team (like updates) but no information from you is given to them through the app. Basically Team servers ---> You and NOT Team servers <--->You) One-way data street.

Brief misunderstanding happened over an open-source license for the front-end part of the app. (Used open-source coding for front-end, not knowing that code’s use-license states it is only for other open-source uses, not closed-source (the back-end code of the app is private to prevent counter-counter measure developments)). The Team took down the app until they replaced the front-end code with code written from scratch by the team. They are now not in violation of that open-source license since they are no longer using it. (you have 30 days to remedy a license breach once informed; they did so in 2)

The Team is currently in touch with Japanese artists to better expand the tool for use to protect their art styles

From what I understand of it, Glaze is an AI tool designed to be anti-AI (Think Arnold Schwarzenegger in Terminator 2: one Terminator robot vs. all the other Terminators 😂)

You can download it from their website and also contact them through email there with any questions, problems, or bugs. The website: https://glaze.cs.uchicago.edu/