Research

Spotlight Memory addresses a small region of expandable storage at each step. Percepta reports strong retrieval beyond the models’ training length.

A visual-reasoning study targets the information held between seeing an image and answering a question. Its training method makes those hidden states more sensitive to relevant image changes.

A controlled study separates who generates training answers from how a student model learns. Changing the objective and learning rate explains much of the difference.

Research, local tools, technical essays and discussions: 57 concise items with direct sources.

A 1,050-leader survey finds widespread AI gains but only 13% of implementers scale projects as originally planned.

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.

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.

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