The news desk
Recent developments in artificial intelligence, with the original sources one click away.
Researchers introduce benchmarks evaluating LLM capabilities in Wikipedia navigation, temporal video reasoning, and formal mathematical theorem proving.
Sources · arXiv AI (cs.AI)
Recent arXiv preprints explore knowledge distillation for interactive AI and fine-grained capabilities of vision-language models.
Sources · arXiv AI (cs.AI)
Researchers publish papers on applying neuro-symbolic AI to academic advising systems and grounding language models with formal mathematical knowledge.
Sources · arXiv AI (cs.AI)
Google executive identifies LLM wrappers and AI aggregators as facing pressure from shrinking margins and limited differentiation.
Sources · TechCrunch AI
OpenAI debated alerting authorities after Jesse Van Rootselaar's violent gun descriptions were flagged by ChatGPT monitoring tools.
Sources · TechCrunch AI
According to NYT, tech leaders worry that lack of public enthusiasm for AI could impact the industry's ambitious growth plans.
Sources · NYT Technology
Dueling AI super PACs are backing and opposing candidate Alex Bores over his proposed RAISE Act requiring AI safety disclosures.
Sources · TechCrunch AI
AWS details SageMaker AI enhancements including Flexible Training Plans, price performance improvements, and better observability features.
Sources · Amazon AWS AI
Anthropic's policies prohibiting AI use in autonomous weapons and surveillance could affect its ability to secure major military contracts.
Sources · WIRED AI
Tech workers show growing concern about AI replacing jobs, though some view it as another tool, according to New York Times report.
Sources · NYT Technology
WIRED examines whether AI data centers could be moved to orbit to address environmental concerns about terrestrial facilities.
Sources · WIRED AI
Google's new Gemini 3.1 Pro model achieves record benchmark scores and aims to handle more complex work tasks.
Sources · TechCrunch AI
OpenAI announces $7.5 million funding for The Alignment Project to support independent research on AGI safety and security.
Sources · OpenAI
Recent arXiv preprints explore frameworks for evaluating health LLMs, physical safety risks in robotic systems, and human mobility simulation.
Sources · arXiv AI (cs.AI)
Recent arXiv preprints tackle LLM-based recommendations, pedagogical risk assessment datasets, and privacy frameworks for children's AI applications.
Sources · arXiv AI (cs.AI)
Researchers propose novel techniques for pruning transformers, compressing LLMs, and accelerating language model generation.
Sources · arXiv AI (cs.AI)
Anthropic has updated its Consumer Terms of Service, restricting the use of OAuth tokens from consumer subscriptions with third-party AI agents like OpenClaw, urging users to switch to API keys.
Sources · Turn your Raspberry Pi into an AI agent with OpenClaw / OpenClaw - Wikipedia +9
Anthropic faces political pressure as debate intensifies over AI use in military applications and battlefield deployment.
Sources · NYT Technology
DeepMind researchers advocate for systematic evaluation of LLMs' ethical responses as their use in sensitive roles like therapy and medicine expands.
Sources · MIT Technology Review
Google's Gemini app now includes Lyria 3, enabling users to create 30-second music tracks from text or image prompts.
Sources · Google AI Blog / DeepMind