– Political comedy titan and returned Daily Show host Jon Stewart, telling U.S. Federal Trade Commission chair and antitrust lawyer Lina Khan about his former bosses interfering in his short-lived current affairs program on Apple TV+. Launched in 2021, The Problem with Jon Stewart was cancelled after only two seasons due to ‘creative differences.’

– Social psychologist and The Anxious Generation (2024) author Jonathan Haidt, on how social media giants, callously, fuel the raging youth mental health crisis. “The companies had done little or no research on the mental health effects of their products on children and adolescents, and they shared no data with researchers studying the health effects.”

– U.S. Attourney General Merrick Garland, describing the iPhone and App Store as an anti-competitive walled garden. “Apple has held a dominant market share not because of its superiority, but because of its unlawful exclusionary behaviour,” he says in a speech announcing sweeping antitrust action against the tech giant.

– American computer scientist, author, and VR pioneer Jaron Lanier, reminiscing on the technology’s early days when, contrary to today’s efforts to make them disappear, VR goggles were exciting aesthetic objects unto themselves. “If you’re going to wear a headset, you should be proud of that weird thing on your head!” Lanier writes.

– Signal Foundation president and AI Now Institute co-founder Meredith Whittaker, on Meta joining the race for artificial general intelligence (AGI). “AGI is a marketing term overlaid with quasi-religious symbolism,” Whittaker warns, reminding her followers that the term AI, too, was coined in 1956 to attract grant money.

– American artist Trevor Paglen, at the 37th Chaos Communication Congress (37C3), illustrating a dystopian near-future of corporate AI companionship, where “emotional manipulation will be the name of the game”

– Signal Foundation president Meredith Whittaker, demystifying the AI revolution at the Washington Post Futurist Summit. “We need to be clear about what we are responding to: ChatGPT is an advertisement—a very expensive advertisement,” Whittaker insists.

– Tech journalist Steve Lohr, reminiscing the last major American antitrust trial (1998). Once again “a tech giant is accused of using its overwhelming market power to unfairly cut competitors off from potential customers,” Lohr writes, noting Google is not quite as audacious though (a Microsoft exec famously planned to “cut off Netscape’s air supply”).

– Canadian tech critic, author, and Tech Won’t Save Us host Paris Marx, pushing against the notion that better tools like Satlas, the Allen Institute for AI’s machine learning-powered forest monitor, will lead to climate action. “Our data has been getting better for decades,” Marx argues, “and emissions have kept rising that whole time.”

– Tech reporter and Platformer founder Casey Newton, parsing the drama around the proposed cage fight between the two Silicon Valley giants in show sports terms. “As a connoisseur of pro wrestling, I’m quite familiar with the character Musk is playing here: the big talker who can’t back it up in the ring,” writes Newton. “Wrestling promoters have made a lot of money with cowardly heel champions [like Flair] who go to great lengths to avoid having to face their adversaries in combat.” In the real world, the Musk vs Zuck feud moves markets and, Newton argues, deserves a lot more skepticism.

– Signal Foundation president and AI Now Institute co-founder Meredith Whittaker, discussing AI ethics with Credo AI’s Navrina Singh, and Distributed AI Research Institute’s Alex Hanna at the Bloomberg Technology Summit. “AI is a surveillance technology,” Whittaker insists. “The Venn diagram of AI concerns and privacy concerns is a circle.”

– New York Times tech columnist Kevin Roose, summarizing the dilemma faced by Reddit, Twitter, and other platforms currently “locking down” their application programming interfaces (APIs) to protect their vast archives (of user-generated content) from being scraped by OpenAI, Google and other companies developing AI language models

– Security engineering expert Ross Anderson, on the dangers of model collapse as AI trains on AI-generated content. In a new paper, Anderson and team demonstrate a rapid degenerative AI feedback loop where “true underlying data distribution” is forgotten. “This will make it harder to train newer models by scraping the web,” Anderson warns, “giving an advantage to firms that already did, or that control access to human interfaces at scale.”