1,577 days, 2,409 entries ...

Newsticker, link list, time machine: HOLO.mg/stream logs emerging trajectories in art, science, technology, and culture––every day
“There is a death drive in all of this. It is a drive to lethality. It is a drive towards self-destruction but also the destruction of all others. That is what underlies these systems.”
This Machine Kills co-host Jathan Sadowski, emphatically rejecting Lavender, an AI system Israel uses to compile ‘kill lists’ of Gazans to target. Drawing a connection to using AI to screen and reject healthcare applicants, Sadowski argues the logic is the exact same, but Lavender “will lead to an immediate kinetic death rather than a somewhat slower social death.”

An immersive, data-driven experience that links fluvial systems, glaciers, and climate change, Theresa Schubert’s solo exhibition “Melting Mountains” opens at MEINBLAU, Berlin, bringing together all three parts of The Glacier Trilogy (2021-22) for the first time. In it, the German bio artist draws on field research in the Western Alps to synthesize fictional archives of snowy peaks, trap ancient meltwater in hand-blown glass sculptures, and drive simulated glacial water systems with visitors’ CO2 exhalations.

“This is what happens when you deprioritize news on a platform. You don’t actually get less news. You just get more Shrimp Jesuses.”
New York Times tech reporter Kevin Roose, on “schlocky AI-generated garbage” replacing trusted reporting thanks to changes in Facebook’s recommendation algorithm. A recent, emblematic example: Shrimp Jesus, a series of AI-generated images of messianic crustaceans that went viral repeatedly, fooling boomers, rewarding scammers, and further eroding people’s sense of reality.
“If your full-time, eight-hours-a-day, five-days-a-week job were to look at each image in the dataset for just one second, it would take you 781 years.”
– German data journalist Christo Buschek and Canadian software artist Jer Thorp, on the scale of generative AI datasets preventing human curation. LAION-5B, for example, contains 5.8 billion image and text pairs that are selected automatically, and it shows: “It contains less about how humans see the world than it does about how search engines see the world. It is a dataset that is powerfully shaped by commercial logics.”

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.”

“Organizations using AI are hiring much fewer engineers for AI-related software than in 2022: 28% of organizations reported hiring for these roles in 2023, down from 39% in 2022.”
– Management consultants McKinsey & Company, tracking post-generative AI hiring trends. A data point in a report on the rise of prompt engineering as a career class, they estimate “half of today’s work activities could be automated between 2030 and 2060.”
L
“But who was this ‘quite normal’ person? And what did she think of ELIZA? In this chapter of the history of talking machines, we only have one side of the conversation.”
– Literature scholar Rebecca Roach, contemplating the erasure of Joseph Weizenbaum’s secretary—the first user of the MIT computer scientist’s seminal 1960s chatbot. “Her reaction sowed the seeds for his later abhorrence for his creation,” yet her name is missing from the records. “Our silent secretary is the quintessential effaced, anonymous transcriber of the documents on which history is built,” Roach concludes.

How (not) to get hit by a self-driving car, an installation and game by Tomo Kihara and Daniel Coppen, opens at San Francisco’s Asian Art Museum. Navigating a mock-street in the gallery, visitors try to make their way across a crosswalk without being detected by a computer vision system. A provocation in a city where autonomous vehicles are regularly involved in accidents, the game underscores “how going unseen by a similar AI system in an actual self-driving car could result in a tragic collision.”

OUT NOW:
Madhumita Murgia
Code Dependent
The intimate investigation of AI that we’ve been waiting for, Indian-British journalist and Financial Times tech correspondent Madhumita Murgia compiles the stories of marginalized people—BIPOC women, war refugees, gig workers, tribal communities—“living in the shadow of AI.”
“We’re artists. We’re not here to be reasonable, or to do what’s necessary, or to cater to regulatory appetites. We’re here to be unreasonable, unnecessary, and counter-appetite. If we stand for opt-in and all we get is opt-out, at least we tried.”
– American software artist Kyle McDonald, pushing back against pragmatism when it comes to protecting creators from Big Tech’s thirst for AI training data. Whereas some advocate for opting out of model data as the only realistic resolution, McDonald doubles down on artists granting permission first.

Consumer advocates Public Citizen release “Mushrooming Risk,” a report on the danger of AI tools to foragers. Spurned by recent hospitalizations, researcher Rick Claypool outlines the folly of using apps to identify species and gauge edibility. “Mushroomers must take the time to develop their skills at their own pace,” he writes, championing local knowledge over app reliance. Experimenting with DALL-E compounded his fear; when he prompted it to label basic mushroom anatomy, it hallucinated utter nonsense (image).

“László Moholy-Nagy telephoned instructions to a factory in the 1920s to produce images without the need to be on-site to direct production.”
– Artist George Legrady, reaching back to the Bauhaus to locate an early precedent for AI-generated imagery. Taking stock of his ongoing experimentation with MidJourney and Stable Diffusion, Legrady explores the aesthetic and critical concerns around prompt-based art.
“The line, when regarded close up, does not have the sweeping fluency or charm of the human hand, of a Matisse, for instance, whom some of the linear portraits seem to be channelling.”
– Critic Lilly Wei, conducting a close reading of the linework of AARON, late American artist Harold Cohen’s painting software. Taking stock of his Whitney retrospective, Wei concludes Cohen’s computer paintings “are animated by texture, touch, and the glow of colour, and better for it, so human artistry isn’t obsolete yet.”
“So literally, I was like, what the fuck? Get these down. What are you doing? It’s as if I was the head of Gucci, and there’s all these knockoffs.”
– Veteran tech journalist Kara Swisher, on AI-generated clones of her memoir, Burn Book (2024), flooding Amazon. First reported by 404 media, the rip-offs have since blossomed in variety, sporting alt titles (Tech’s Queen Bee With A Sting), different authors, and illustrious synthetic cover photography. “So I, of course, put them all together, and I sent Andy Jassy [the CEO of Amazon] a note and said, what the fuck? You’re costing me money.”
“The internet is filling up with ‘zombie content’ designed to game algorithms and scam humans. It’s becoming a place where bots talk to bots, and search engines crawl a lonely expanse of pages written by artificial intelligence.”
– Technology reporter James Purtill, on the accelerating degradation of the Web. Bot-overrun platforms like X were “never designed for a world where machines can talk with people convincingly,” says digital media scholar Timothy Graham of the deluge of AI-generated content. ”The platforms have no infrastructure in place. The gates are open.”
“Neither side in the debate on ‘killer robots’ has addressed the most potentially unnerving aspect of using them in battle: sooner or later, they’ll be able to communicate with each other without human intervention.”
– World security scholar and author Michael Klare, on the dangers of emergent behaviour in autonomous weaponry as DARPA expands efforts to create military AI systems capable of true swarming. “Autonomous weapons might jointly elect to adopt combat tactics none of the individual devices were programmed to perform”, warns Klare, “conceivably engaging in acts unintended and unforeseen by their human commanders.”
“It’s estimated that a search driven by generative AI uses four to five times the energy of a conventional web search. Within years, large AI systems are likely to need as much energy as entire nations.”
– AI researcher Kate Crawford, on the soaring (and mostly secret) environmental costs of generative AI. “Rather than pipe-dream technologies, we need pragmatic actions to limit AI’s ecological impacts now,” Crawford writes, calling for a “multifaceted approach including the AI industry, researchers and legislators.”
“If you put me to a wall and forced me to put probabilities on things, I have a sense that our current remaining timeline looks more like five years than 50 years. Could be two years, could be 10.”
– Machine Intelligence Research Institute (MIRI) founder Eliezer Yudkowsky, doubling down on the impending AI apocalypse he warned about in his March 2023 TIME op-ed. “We have a shred of a chance that humanity survives,” the prominent doomer argues in Tom Lamont’s neo-luddite survey, advocating for an immediate development stop.

KW Berlin opens “Poetics of Encryption,” an extensive group exhibition that builds on curator Nadim Samman’s eponymous book, illuminating “Black Sites, Black Boxes, and Black Holes”—terms that indicate how technical systems capture users, how they work in stealth, and how they distort cultural space-time. The show gathers both historic and newly commissioned works by over 40 artists including Nora Al-Badri, Clusterduck, Kate Crawford & Vladan Joler, Simon Denny, Eva & Franco Mattes, Trevor Paglen, Rachel Rossin, and Troika.

To dive deeper into Stream, please or become a .

Daily discoveries at the nexus of art, science, technology, and culture: Get full access by becoming a HOLO Reader!
  • Perspective: research, long-form analysis, and critical commentary
  • Encounters: in-depth artist profiles and studio visits of pioneers and key innovators
  • Stream: a timeline and news archive with 1,200+ entries and counting
  • Edition: HOLO’s annual collector’s edition that captures the calendar year in print
$40 USD