Mistral Large 4, an open-weight model trained in Europe, owned the day and framed the bigger question: who controls the AI other people run. The same thread runs through OpenAI publishing machine-made mathematics, a mega media merger clearing its last legal hurdle, and a privacy fiasco at Meta.
The day’s biggest story puts an ordinary chatbot user in court: a Florida woman faces a felony charge after Claude flagged an entry she wrote like a diary. Privacy, platform permissions, and stalled regulation run through the rest of the digest, with science, business, and culture each holding their own.
An obituary topped the day: Bob Cringely, who told the personal-computer story to a general audience, has died. The rest of the issue splits AI’s two faces — running large models on a home PC and reining in runaway bills — alongside resurrected old hardware, a war-time energy case, and arguments about trades work.
A court stopped Utah’s anti-VPN law on the grounds that it asks for a technical impossibility, opening a day of debate about how far public rules can reach into everyday tools; meanwhile a government-run dating service, desktop AI helpers, and archives searched by models keep the conversation wide. The science desk is lighter: a 12-year planetary time-lapse, a new theory of aging as lost identity, and a dodo sighting pulled from a Dutch ledger.
Cheaper near-frontier models and always-on agents promise more capability, while trials of live facial recognition, leaks from AI chat apps, and stories from Delhi, California tables, and American living rooms ask what guardrails and everyday resilience we need to keep up.
A weirdly empathetic search box, unsealed filings on training data, a cheaper way to think, and a fluid that clicks — today’s stories trace how accuracy, accountability, and cost are renegotiated across search, AI, law, and everyday repair.
As models compete on price and endurance, platforms make refusal harder and tracking quieter. Today’s issue weighs efficiency against consent, traceability, and the human work of writing and history.
As tools get lighter — GET-only exfiltration, transparent image layers, and peer-to-peer model archives — the harder questions move to defaults for privacy, memory, and health, with hardware and math keeping score in the background.
From poster fatigue to 33-millisecond decisions, from a brain that is really two systems to a new cat species, today’s stories ask not how powerful the tools are, but how humans keep judgment, variety, and verifiability.
From copyright on trial to passkeys that lock you out, from a coding app that quietly copied whole repos to fines and AI intel failures, the question is not just what technology can do, but whether we can verify, refuse, and own it.