To illuminate how generative AI models like Midjourney and Stable Diffusion derive their worldview from its 5.8 billion image and text pairs, German data journalist Christo Buschek and Canadian software artist Jer Thorp deconstruct the (only) open-source foundation dataset LAION-5B in an incisive, visual essay. Digging deep into its troubled contents, algorithmic—not human—curation, and entanglements with other systems, the two warn about stacking “models on top of models, and trainings sets on top of training sets.”

Anicka Yi’s solo exhibition “A Shimmer Through The Quantum Foam” opens at Esther Schipper, Berlin, evolving the Korean-American artist’s notion of the “biologized machine” with new works. Visitors enter a hybrid ecosystem of fleshy landscapes created with machine learning models and suspended luminescent pods resembling Radiolaria. As the soft glow of an aqueous ooze—indicative of life’s marine origins—sprawls across the gallery floor, a custom-made scent by perfumer Barnabé Fillion fills the air.

American software artist Casey Reas returns to Berlin’s DAM Projects with “Conjured Terrain,” a solo exhibition of new Untitled Film Stills and Compressed Cinema digital video works set to (and driven by) the electroacoustic soundscapes of German artist and composer Jan St. Werner. Building on a body of images ‘conjured’ from feature films fed to generative adversarial networks (GANs) in 2018, Reas revisits—and celebrates—the raw visual grammar of early machine learning experiments from that era.

MouthPad^, a smart oral splint that enables hands-free human-computer interaction using tongue, jaw, head, and breathing gestures, wins the grand prize in Innovation: Early Stage Technology at Cannes Lions International Festival of Creativity. Developed by MIT Media Lab spinoff Augmental, the forthcoming machine learning-powered ‘tongue trackpad’ gives users (not just those lacking mobility) full reign over their gadgets—to send emails, edit photos, or play videogames—while being invisible to others.

Full of playful examples—statistically modelling dropping cannonballs from different heights, a neural net theory of cat recognition—Stephen Wolfram breaks down how ChatGPT works. Working from the simple claim “it’s just adding one word at a time,” the computer scientist describes how neural nets are trained to model ‘human-like’ tasks in 3D space, how they tokenize language, and concludes with a rumination on semantic grammar that recognizes the language model’s successes (and limits).

The procedurally generated Seinfeld spoof Nothing, Forever is temporarily banned on Twitch after lead character Larry Feinberg made transphobic remarks. The show’s developers blame switching from OpenAI’s GPT-3 Davinci model to its predecessor, Curie, after the former caused outages. “We leverage OpenAI’s content moderation tools, and will not be using Curie as a fallback in the future,” they state on Discord. Launched in December 2022, the show became a viral hit for its nonsensical humour, nondescript style, and audience activity.