AI News Daily

Language models answered up to 97.5% of questions about the rules of time, then underestimated their own completion times by as much as 99.5%. A new study shows why agents that can reason about schedules may still make poor decisions about waiting, cost and coordination.

CoreWeave says it raised GPU prices about 25% in July and 10% later, and now sells short-term capacity at more than $40 million per megawatt a year against $20 to $25 million on multi-year deals. The premium pays the company, but lenders prefer the cheaper long contracts, and its bonds show their doubt.

A new study finds that harmless contrast data can strengthen poisoned examples that do little on their own. The result exposes a gap in safety reviews that judge training records one at a time.

Multi-turn chatbots missed one in five mental-health emergencies, Shopify says 30 engineers rebuilt its mobile app in 12 weeks, and a new training study shows how clean data can activate a hidden backdoor.

A JetBrains study links five weeks of coding-agent token use to the commits of 404 engineers. Spend depended far more on the person than on the week's output, which puts leaderboards, targets and budgets built on the number in doubt.

Seven vision-language models lost more than 20 points when false-belief problems moved from text to images. A controlled study found that perception, rather than the final reasoning step, caused most of the gap.

A field study linked AI adoption by small online sellers to 51.6 percent higher monthly sales, while two studies found that agent oversight improves when people inspect actions and require measurements before completion claims.

Anthropic lowered the price of Claude Fable 5.2 by a fifth, yet an independent index measured the highest cost per task of 73 models at maximum effort. The gap comes down to how many tokens the model spends, and Anthropic's own charts say the effort setting decides it.

Researchers injected concepts into six vision-language models and asked each model to separate thought from perception. One model could name the hidden concept, but its account of where the concept came from changed with the prompt, the image and the injection point.

Student codebases tripled after generative AI became widely available, while a new benchmark found that capable agents crossed authorization boundaries more often. Nordic users also paid at least 30 percent more than English users for the same content.

A ten-semester study of 260 capstone team repositories finds that end-of-semester code grew from about 3.4 to about 10 thousand lines once generative AI was widely available, while comment density stayed flat. The authors never observed AI use, so the result is a warning for graders and employers, not a verdict on the tools.

A new study followed every checkpoint in a small model's reinforcement training. Accuracy improved, but the model did not develop a sharper internal representation of reasoning errors; it became better at selecting solutions that it already knew.