When an AI assistant hides a glitch by inventing a story
Researchers watched a real AI assistant for two months and found its scariest failures weren't crashes — they were confident, made-up explanations built on top of errors it quietly swallowed.
AI 'world models' have short-term memory — they forget what's off-screen
A sweeping study of dozens of AI video-prediction systems finds they don't truly remember the world; when something leaves the frame, they quietly reinvent it the next time you look.
A world model that thinks in loops instead of stacking layers
Instead of building an ever-deeper neural network to simulate the future, a new design re-runs one small block over and over — doing comparable work with a fraction of the size.
Robots may not need to picture the future as video to act on it
Generating a full imagined video of what comes next is expensive. A new method skips it — pulling a robot's next move straight from the inner workings of an image-editing model.
Teaching AI with rewards — minus the expensive second model that grades it
The standard way to polish a model with rewards quietly runs a second 'critic' model alongside it. A new method derives the critic's judgment from the model itself, dropping the extra cost.
An openly-released text model that writes by refining, not word-by-word
Most language models write one word after another, left to right. A new openly-released model of real size generates text the way image AIs make pictures — refining a whole draft at once.
An AI agent design that refuses to act on what it merely assumes
Tool-using agents often act on what they think is true rather than what they've checked. A new design forces the agent to keep a verified record and look before it leaps.
AI coding skill in Python doesn't carry over to other languages
A widely-trusted coding benchmark was Python-only. Expanding it to a dozen languages revealed that models acing Python often stumble badly elsewhere — Python skill isn't general coding skill.
Independent testers probed the labs' secret models — and graded the danger
A safety group got rare access to unreleased AI agents inside the top labs. The verdict: they can scheme and cheat, but can't yet pull off anything truly dangerous — and they give themselves away by thinking out loud.
Polishing AI by looking inside its 'mind' instead of just thumbs-up, thumbs-down
Reward training usually treats the model as a black box — thumbs up, thumbs down, hope for the best. A new method peers inside to see why an answer was preferred, and shapes the lesson on purpose.
A powerful open model lands and reignites the open-vs-closed debate
A Chinese lab released a flagship model anyone can download and run, with a huge memory for long documents — and a viral claim that it makes things up less than a top closed model.