Yesterday, despite the manifest, glaring problems with its porn filter, Tumblr turned on mandatory porn-blocking for all its users' content, so that anything that its bots identified a pornographic would be invisible.
Tumblr faced a backlash over its announcement that it would ban "adult content," and not just because Tumblr had traditionally nurtured communities built around queer and marginalized sexuality -- but also because the company defined "adult content" in part as being "female presenting nipples."
In response to concerns about blocking breastfeeding images, breast-cancer awareness posts, and other "legitimate" uses of "female-presenting nipples," Tumblr put up a post illustrating and explaining the kinds of "female-presenting nipples" the service would allow: "exposed female-presenting nipples in connection with breastfeeding, birth or after-birth moments, and health-related situations, such as post-mastectomy or gender confirmation surgery.
The images that accompanied the post? Blocked by Tumblr's porn filter.
When Gizmodo posted the gif to Tumblr ourselves, it was immediately flagged as a potential violation and hidden by the platform’s filter. When the images shown in the gif were uploaded individually, two of the four examples—which appeared to show a breast ultrasound and a pro-choice protest, respectively—were similarly flagged and hidden.
Tumblr's Porn Filter Flags Its Own Examples of 'Permitted' Nudity [Hudson Hongo/Gizmodo]
(via Beyond the Beyond)
Creative technologist Nicole He modified OpenAI’s GPT-2 language model to generate questions for happy mutant pop star Billie Eilish and also write Eilish-esque lyrics. Vogue Magazine published Eilish’s answers to the AI’s wonderfully odd questions like: “Who consumed so much of your power in one go?” and “Have you ever seen the ending?”
From the excellent “Two Minute Papers” YouTube channel, a discussion of a paper titled “DeepRemaster: Temporal Source-Reference Attention Networks for Comprehensive Video Enhancement,” that demonstrates the results of a neural network that fixes and colorizes aged, blurry, scratchy films.
Face-synthesizing algorithms often struggle with facial details like eyes and teeth. These features sometimes get pinned to a fixed spot as a head turns, resulting in an uncanny valley dweller. A new algorithm, StyleGAN2, fixes this problem and produces “eye-poppingly detailed and correct images.” It can also generate never-before-seen cars, churches, and animals. Image: YouTube
You hear the stories all the time. You know the ones about how a new vehicle instantly loses hundreds, even thousands of dollars in value the second a new owner drives it off the lot. Depreciation is a killer, especially when the item itself works just as well — or maybe even better than all […]
Spring cleaning time is right around the corner, and, if we’re being honest, the whole thing is kind of a drag. But keeping your home clean in any season is necessary for both health and happiness, so why not make it a little less daunting? The correct tools help make any work easier, and that’s […]
Quality knives matter — and in more ways than you might expect. At the most basic level, having a functioning set of proper kitchen knives means you can actually get in a kitchen, start putting together a meal and have the right tools to handle all tasks. From the blunt force of the clever to […]