We’re introducing, a new evaluation that measures model performance on economically valuable, real-world tasks across 44 occupations.

While many working people are reasonably worried about AI taking their jobs and leaving them on the street, another consequence of the AI revolution is filling seats in mental health facilities.
The mass adoption of large language model (LLM) chatbots is resulting in large numbers of mental health crises centered around AI use, in which people share delusional or paranoid thoughts with a product like ChatGPT — and the bot, instead of recommending that the user get help, affirms the unbalanced thoughts, often spiraling into marathon chat sessions that can end in tragedy or even death.
New reporting by Wired, drawing on more than a dozen psychiatrists and researchers, calls it a “new trend” growing in our AI-powered world. Keith Sakata, a psychiatrist at UCSF, told the publication he’s counted a dozen cases of hospitalization in which AI “played a significant role” in “psychotic episodes” this year alone.
Measuring the object’s non-gravitational acceleration, the team believes they found something “anomalous”
Ternary computing uses-1, 0, and 1 instead of just 0 and 1, and for a brief moment in the 1950s, it looked like it could redefine how we build computers. A Soviet team even built a working ternary machine called Setun. So why did the world choose binary? And could ternary still make a comeback?
Sources, transcript and more available on codeolences.com.
🟣 Subscribe for more deep dives into strange tech, forgotten history, and curious ideas in computing.
▶️ This video is part of a series called \.
Synthetic biology offers a toolkit to engineer microbes capable of surviving in outer space and for biomanufacturing materials to support astronauts on long missions.
Stressful factors like chronic pain, low income, less education and other social risks were associated with older-looking brains. Those links seemed to make less of an impression over time. What stood out more clearly were protective elements: things like getting restorative sleep, maintaining a healthy weight, managing stress, avoiding tobacco and having supportive relationships.
Study participants who reported the most protective factors had brains eight years younger than their chronological age when the study started, and their brains went on to age more slowly over the next two years.
We release Code World Model (CWM), a 32-billion-parameter open-weights LLM, to advance research on code generation with world models. To improve code understanding beyond what can be learned from training on static code alone, we mid-train CWM on a large amount of observation-action trajectories from Python interpreter and agentic Docker environments, and perform extensive multi-task reasoning RL in verifiable coding, math, and multi-turn software engineering environments. With CWM, we provide a strong testbed for researchers to explore the opportunities world modeling affords for improving code generation with reasoning and planning in computational environments. We present first steps of how world models can benefit agentic coding, enable step-by-step simulation of Python code execution, and show early results of how reasoning can benefit from the latter.