Yann LeCun’s AMI Labs debuts to build AI world models
Yann LeCun’s startup AMI Labs has launched to develop “world models” for industrial and healthcare applications, with reports suggesting it might raise funding at a $3.5 billion valuation.
Sections
AI & Models
398 storiesMercor benchmark shows AI models struggle with professional tasks
Mercor’s new APEX-Agents benchmark reveals that most AI models struggle to perform complex, multi-domain white-collar tasks, with the highest-performing model achieving 24% accuracy.
Google DeepMind CEO questions OpenAI's early move into AI ads
Google DeepMind CEO Demis Hassabis expressed surprise at OpenAI’s early move to test ads, stating Google has no current plans to introduce ads in its AI chatbot.
Google launches free SAT practice exams powered by Gemini
Google has launched free, Gemini-powered SAT practice exams, a move that aims to democratize test preparation while potentially disrupting the traditional private tutoring industry.
Google’s AI Mode now accesses Gmail and Photos for users
Google is integrating "Personal Intelligence" into its AI Mode, allowing subscribers in the U.S. to receive tailored responses based on their Gmail and Google Photos data.
Anthropic redesigns hiring tests as AI models match human performance
Anthropic is redesigning its technical hiring tests because its latest AI models, specifically Claude Opus 4.5, have become capable of matching the performance of top human candidates.
Anthropic updates Claude’s constitution with new safety guidelines
Anthropic has released a revised version of its 'Constitution,' an 80-page document detailing the ethical principles and safety guidelines that theoretically govern its Claude chatbot.
GPTZero finds hallucinated citations in NeurIPS research papers
GPTZero identified 100 hallucinated citations across 51 NeurIPS papers, a finding considered not statistically significant, though organizers note the research is not necessarily invalidated.
OpenAI adds age prediction to ChatGPT to protect minors
OpenAI has launched an age prediction feature for ChatGPT that uses behavioral signals to identify minors and automatically apply content filters to their accounts.
Signal co-founder launches privacy-focused AI service Confer
Signal co-founder Moxie Marlinspike has launched Confer, a privacy-focused AI service that uses Trusted Execution Environments to prevent data collection and model training on user conversations.
Chai Discovery raises $130M to scale AI-driven drug development
Chai Discovery raised $130 million at a $1.3 billion valuation and partnered with Eli Lilly to accelerate AI-driven drug discovery using its proprietary Chai-2 algorithm.
Anthropic hires former Microsoft India MD to lead local expansion
Anthropic has appointed former Microsoft India managing director Irina Ghose to lead its India business as the company prepares to open an office in Bengaluru.
AI talent shifts accelerate as OpenAI poaches from competitors
OpenAI has hired three executives from Thinking Machines Lab, with two more expected to follow, while losing safety staff to Anthropic and recruiting new engineering talent.
AI models are beginning to solve complex Erdős conjectures
OpenAI’s GPT 5.2 is anecdotally showing improved mathematical reasoning, with AI models now credited in 11 of 15 recently solved Erdős conjectures.
Google updates Trends Explore with Gemini AI capabilities
Google launched a revamped Trends Explore page featuring Gemini-powered capabilities, allowing users to automatically compare up to eight search terms on desktop starting today.
Skild AI raises $1.4 billion at $14 billion valuation
Robotics software startup Skild AI raised a $1.4 billion Series C round, pushing its valuation to more than $14 billion as investor interest in general-purpose robotic models grows.
Google adds 'Personal Intelligence' to Gemini beta
Google launched a Gemini beta feature called Personal Intelligence that allows the AI to reason across user data in Gmail and Photos to provide proactive, context-aware responses.
OpenAI and Anthropic launch new health-focused AI tools
OpenAI and Anthropic have launched new health-focused AI products, sparking debate over the balance between administrative efficiency and risks like data privacy and AI hallucinations.
Google updates Veo 3.1 for native vertical video creation
Google updated its Veo 3.1 AI video model to support native 9:16 vertical video creation, improved character consistency, and higher-resolution upscaling for professional and consumer users.
Anthropic launches Claude for Healthcare tools
Anthropic has launched Claude for Healthcare, a suite of tools designed to automate administrative tasks, aiming to differentiate itself from OpenAI’s consumer-focused health offerings.