AI
MosaicLeaks: Can your research agent keep a secret?
MosaicLeaks is an experiment that tests the confidentiality of AI-powered research agents. These agents are designed to assist researchers by gatherin...
AI
AI
New usage analytics and updated spend controls for enterprises
OpenAI has released updated features for its ChatGPT Enterprise platform. The changes include new usage analytics and spend controls designed to help ...
AI
Improving health intelligence in ChatGPT
ChatGPT's health and wellness responses have been improved through the introduction of GPT-5.5 Instant. This update enhances the model's ability to re...
AI
Using AI to help physicians diagnose rare genetic diseases affecting children
Researchers have been using a reasoning model from OpenAI to aid physicians in diagnosing rare genetic diseases affecting children. By applying this A...
AI
Beyond LoRA: Can you beat the most popular fine-tuning technique?
Hugging Face researchers propose a new method called PEFT, which they claim outperforms LoRA in some tasks. PEFT uses a different approach to fine-tun...
AI
Is it agentic enough? Benchmarking open models on your own tooling
Researchers are questioning whether current AI models, such as those developed by Hugging Face, truly have human-like agency. To address this concern,...
AI
From the Hugging Face Hub to robot hardware with Strands Agents and LeRobot
Amazon's AI research team has developed a platform that enables users to deploy their models from the Hugging Face Hub directly onto robot hardware. T...
AI
A near-autonomous AI chemist improves a challenging reaction in medicinal chemistry
Researchers at OpenAI and Molecule.one have developed a near-autonomous AI chemist that uses the GPT-5.4 model to improve a challenging reaction in me...
AI
GLM-5.2: Built for Long-Horizon Tasks
Hugging Face has released version 5.2 of its General Language Model (GLM) architecture, designed for tasks that require long-term planning and underst...
AI
Introducing LifeSciBench
LifeSciBench is a new benchmark designed to evaluate the performance of AI systems in handling real-world tasks and decisions related to life science ...
AI
Agentic Resource Discovery: Let agents search
Researchers have developed a new approach called Agentic Resource Discovery, which allows AI agents to autonomously search for and discover new resour...
AI
Unlocking UK house-building with AI-accelerated planning
The UK government is teaming up with Google's DeepMind to develop an artificial intelligence (AI) system that can accelerate the planning process for ...