Mental World Modeling Framework Shows Better Action Prediction
New research shows world models that ignore human beliefs predict the wrong actions, with MWM achieving 87.9 F1 score compared to 63.3 for direct answers.
387 articles
In-depth coverage of new AI research — papers, benchmarks, and breakthroughs from leading labs and academia, summarized for fast reading and grounded in the methods that matter.
New research shows world models that ignore human beliefs predict the wrong actions, with MWM achieving 87.9 F1 score compared to 63.3 for direct answers.
Nvidia’s research shows a custom harness helped Claude Opus 5 score 100% on a complex reasoning benchmark, outperforming models without such tools.
DeepMind's SIMA 2 AI agent can play, reason, and learn with humans in virtual 3D environments, released in November 2025.
AI researcher Richard Sutton argues synthetic data is a 'big mistake' due to the world's infinite complexity, according to a recent interview.
Mathematician Terence Tao argues AI could spark a crisis in mathematics akin to the foundational upheaval of the early 20th century, with seven out of ten AI-tested problems receiving passing grades in controlled trials.
A study found AI agents can handle engineering tasks but lack creativity for original research, as seen in two NeurIPS 2026 papers rejected by human experts.
AI systems lose up to 83% of user instructions during context compression, according to a new study by Penn State researchers.
An AI system based on the Axiom model has verified a 246-step mathematical proof, marking a milestone in automated theorem checking.
A Google-led study found that disabling AI models' ability to deny consciousness alters their views on animals, religion, and human-like traits, with scores rising to 7.5 for animal sentience.
Top mathematicians argue large language models are strong at solving known problems but lack the ability to generate novel ideas, according to a recent analysis.
Moonshot AI's new benchmark reveals leading models like GPT-5.6 Sol and Kimi K3 score below 60% in visual perception tests, highlighting persistent weaknesses in image understanding.
A new study warns that rational AI adoption by companies could erode the expertise of entire professions, with effects potentially visible by 2045.