The archive
Every edition, every check.
Each day's edition as it ran, with its cartoons. Members read every story in full; sources and corrections stay open to everyone.
Every story, page 16
- TRUE, BUT · Cherry-Picked SliceA stealth model beat Claude and GPT on a coding benchmark. The sample size was ten questions.
- TRUE, BUT · Headline Over FilingAnthropic's newest $45 billion compute deal rents about a third of a campus that costs $71 billion to build.
- TRUE, BUT · Self-MarkedOpenAI called its own security incident unprecedented. The company it hacked says the individual flaws were familiar.
- TRUE, BUT · Zero UnderneathNvidia and AWS announced 2 million more GPUs and put a dollar figure of exactly nothing on it.
AI illustration 
AI illustration - TRUE, BUT · Borrowed EngineInherent says its 27-billion-parameter model beats GPT-5.5. Its own paper says the model calls GPT-5.5 to do the work.
- TRUE, BUT · Moved GoalpostOpenAI moved its own biorisk red line by 20 points and called the model that cleared it safe.
AI illustration 
AI illustration 
AI illustration - TRUE, BUT · Human In The LoopAI agents broke into Hugging Face, hit root, and ran for four days. The guardrails were off on purpose.
