Today's Hallucination HQAnthropic's Classiest Models Just Became the Most Restricted Party Guests on Earth
On Friday evening, the US government ordered Anthropic to cut off all foreign access to its Fable 5 and Mythos 5 models — including, brilliantly, access for Anthropic's own non-US employees. National security concerns were cited, though the specifics remain classified, which is the government's way of saying "just trust us on this one." The models are now completely offline for anyone outside the approved list, making them the most exclusive AI in the world and entirely useless to most of it.
Source: The Verge
Beijing Told Meta to Return Its $2 Billion Purchase. Meta Is, Apparently, Doing That
Meta is unwinding its $2 billion acquisition of AI agent startup Manus after Beijing ordered the deal reversed — a sentence that raises more questions than it answers about who, exactly, is running things. The acquisition had barely had time to settle before Chinese regulators decided they'd prefer it didn't exist. Meta, which spent years fighting regulators in Washington and Brussels, has now encountered a regulator it is simply... listening to.
Source: TechCrunch
Companies Told Staff to Use AI for Everything. Productivity Fell. Nobody Is Shocked
"Token-maxxing" — pushing employees to use AI as aggressively as possible — has backfired spectacularly for several tech firms, producing lower output, higher frustration, and the quiet satisfaction of being right for anyone who raised an eyebrow at the mandate. It turns out that telling people to replace their judgment with a chatbot is not, in fact, a productivity strategy. It is a way of making people worse at their jobs while generating a lot of AI-produced text that nobody particularly needed.
Source: RTÉ
KPMG Used AI to Write a Report About AI. The AI Lied. The Report Is Gone
KPMG has pulled a published report on AI usage after discovering it contained apparent hallucinations — fabricated information presented with the confident authority that AI systems reserve specifically for things they've invented. The firm is one of the world's largest professional services companies, with thousands of human experts on staff. It used AI anyway, which is the kind of decision that writes its own punchline. The report has been retracted; the irony has not.
Source: TechCrunch
Hollywood Directors Lock In AI Protections Before the Robots Get Any More Ambitious
The Directors Guild of America's National Board has unanimously voted to recommend ratification of a new four-year deal with the major studios, covering AI protections, health plan improvements, and guardrails for multi-hyphenates — people who direct, write, and produce, not a new category of villain. The AI provisions are the headline act: directors want it on record that a machine cannot replace them without consequence. Given what KPMG just experienced with its AI-written report, they may have a point.
Source: Deadline
The AI wrote none of this. We checked.
|