Posts

A Europe-wide investigation found major gaps between mandatory data-centre reporting and public records. In the Netherlands, the registry lists 104 of 186 qualifying commercial sites.

Open model experiments, local-agent infrastructure and debates over AI-generated mathematics led the community agenda alongside a warehouse-robotics deployment.

Anthropic says the open-weight model built working browser exploits and that common techniques weakened its safeguards. A US government assessment independently confirms a sharp capability gain.

Dots can continue work in the background, use connected applications and contact users when decisions are needed. The rollout gives persistent agents wider access to workplace data.

OpenAI's new mid-tier model costs one-fifth as much as GPT-6 Astra at standard API rates. Its performance and safety comparisons remain company-run evaluations.

British Transport Police scanned more than half a million faces during a six-month trial. The system produced one incorrect watchlist alert and no alert-led arrests.

OpenAI's DevDay release turns ChatGPT into a shared work surface with collaborative documents, recurring tasks and team-controlled application connections.

PostHog added reasoning to a decision model, OpenAI expanded its developer and enterprise offers, and communities debated privacy, infrastructure economics and open weights.

Developers debated how to verify AI output, while researchers shared a decision-model project and revisited coding-agent evaluation. The items distinguish opinions and research claims from independent findings. » Why it matters: A launch announcement answers what a supplier offers. These discussions ask how a person checks the result, retains understanding and decides when an apparently successful task is incomplete. Hacker News Two coding essays put understanding beside output speed. Alex Ewerlöf’s September 26 essay argues that generating code does not remove responsibility for maintenance and correctness. A separate September 28 discussion of system architecture asks how developers retain enough understanding to review changes. These are related professional arguments, merged here as one item. Their value is a practical question for teams: can the person accepting a patch explain its effect on the surrounding system? Neither thread establishes a measured productivity loss. Ewerlöf discussion; author’s essay; architecture discussion. ...

The proposed share-based acquisition would bring spatial-model research into AMD. Closing and the promised engineering benefits remain future events.

Anthropic keeps Sonnet’s token prices while claiming lower task costs. The useful comparison is the cost of work that passes review.

The report asks whether automated AI development could sharply accelerate progress. It presents a risk scenario with uncertainty, not an established timetable.