Full-text search across 7441 articles. Combine with topic and date filters; results sorted by relevance.
GenCeption, a new computer vision model from an unnamed research team, uses text-to-video generation to achieve state-of-the-art performance on tasks like depth estimation and 3D keypoint prediction, often surpassing spe…
Argybargy is a new open-source peer-to-peer bridge that connects multiple AI agents and sessions across different machines, apps, and model vendors using plain HTTP/JSON. The tool enables agents to communicate, coordinat…
Berkeley AI Research's Class of 2026, a cohort of new Ph.D. graduates, is advancing AI across robotics, computer vision, and AI safety, with many heading to top faculty positions, industry labs, and startups. Their work …
InfoQ's 2026 Culture & Methods Trends Report, based on a panel of QCon speakers and InfoQ contributors, finds that AI adoption demands maturity frameworks assessing risk and context, and warns that organizations failing …
Electric vehicles are straining power grids, prompting researchers to test AI-driven multi-agent reinforcement learning for smarter charging. Simulations using contextual combinatorial bandits and policy gradient algorit…
OpenAI unveiled GPT-Red, its strongest automated safety red-teaming model, which successfully attacked GPT-5.1 in 84% of test cases compared to 13% for human red-teamers. The internal-only model, trained via self-play re…
A developer has published a detailed AI usage policy explaining how they use large language models for grammar checking and code implementation but not for generating ideas or writing drafts. The policy includes a JavaSc…
A new study explores whether large language models can improve their ability to generate accurate math diagrams for K-12 education through an agentic workflow that lets them evaluate and iteratively refine their own outp…
A new benchmark called Mini Amusement Parks (MAPs) reveals that AI agents lag far behind humans in complex decision-making, with humans outperforming AI by 6.5 times in easy mode and 9.8 times in medium difficulty. The s…
Yann LeCun's AMI Labs raised $1.03 billion in March 2026 to develop world models based on Joint Embedding Predictive Architecture (JEPA), which argues that intelligence begins with building internal models of the world r…
Albright College is launching one of Pennsylvania's first structured AI training programs for K-12 teachers this fall, a four-course Artificial Intelligence Endorsement approved by the Pennsylvania Department of Educatio…
Real-world vision deployments fail because object detectors answer 'what is in this frame?' while businesses need to know 'what changed?', 'what does it mean?', and 'what should happen next?'—questions that require state…
China's Ministry of Commerce vowed to 'take all necessary measures' if the US imposes new AI sanctions, and counter-claimed that many US AI companies have distilled Chinese models. The allegation follows US accusations t…
Procedural Memory Distillation (PMD) advances reinforcement learning by transforming experiences across episodes into actionable intelligence, outperforming previous models like SDPO by 3.8-5.5% on SCIKNOWEVAL and 7.9-13…
A developer explores the rise of personal AI as an eternal archivist, arguing that while it offers unprecedented self-awareness through objective behavioral analysis, it also threatens human identity by eliminating the r…
Aether AI, a San Diego-based startup, raised $20mn in seed funding to develop causal world models that teach machines cause and effect, challenging the industry's focus on scaling larger models. The company targets robot…
EvoClawBench, a new benchmark testing whether AI agents can learn reusable skills from experience, shows mixed results across 100 tasks and 502 sub-problems. Nanobot's GPT-5.4 model consistently exceeded 96% accuracy, bu…
A Deloitte survey of 1,150 parents finds half worry about children over-relying on AI in schools, while over a third say schools fail to teach AI skills. Only 22% of parents report schools have approved AI tools, and 33%…
Researchers reproduced the core mechanisms of two 2025-26 JEPA papers (CrossJEPA and LeJEPA) at toy scale on an M5 MacBook Air, confirming that the self-supervised representation learning methods work as claimed. The val…
Researchers are developing new techniques to enhance metacognitive abilities in large language models, including reinforcement learning with metacognitive feedback (RLMF) and metacognitive data selection. RLMF improves m…