Anthropic improved the generalization of alignment training with Model Spec Midtraining.
A method for optimizing Qwen3.6 inference performance by grafting an MTP layer onto an Unsloth quantized model is introduced.
Friday Studio, an AI agent runtime that converts natural language into repeatable YAML workflows and executes them, has been released.
Anthropic has released the Model Context Protocol (MCP), an open standard for connecting AI assistants with data sources.
In llm-gemini 0.31, Gemini 3.1 Flash-Lite has transitioned to a stable version.
The US Department of Energy and NVIDIA emphasized the power needed for AI scaling and their collaboration on the Genesis Mission.
The 'mode collapse' phenomenon, where ChatGPT repeats certain phrases, has become controversial among Chinese-language users.
In Anthropic's Gmail connector, the existing message-reading tool has been removed and replaced with a tool that only provides metadata.
OpenAI is introducing a Trusted Contact feature for adult ChatGPT users that sends alerts in cases of serious safety concerns.
OpenAI unveiled three real-time voice models for the Realtime API that support reasoning, translation, and transcription.
Don't just rely on CI passing for agent PRs—you need to directly review for duplication and security.
NASA and IBM's Prithvi geospatial foundation model has completed its first orbital demonstration aboard the ISS and a satellite.
Simon Willison built Big Words, a tool that creates simple slides from a URL.
A Hugging Face model impersonating a legitimate one has reportedly been found downloading malicious scripts.
xAI is providing Colossus 1 to Anthropic to increase Claude capacity.
The UK is introducing AI camera technology into cars to detect drunk and drugged drivers.
Canadian privacy regulators found that OpenAI's data collection and training practices violated privacy law.
Canadian privacy regulators determined that OpenAI's collection of personal information violated the law.
Court evidence has revealed that Elon Musk attempted to recruit OpenAI's founders to an AI lab within Tesla in 2018.
At the Code with Claude event, Anthropic unveiled Managed Agents and Claude Code updates.
A legal NER model trained on 33,000 Indian Supreme Court judgments has been released, showing improved performance over existing models.
Mozilla used Claude Mythos Preview to find and fix hundreds of Firefox security vulnerabilities.
The Trump administration is reviewing plans to invite the CEOs of Nvidia and Apple to join a delegation visiting China.
OpenAI has added GPT-Realtime-2 and two translation/transcription models to the Realtime API.