Artificial intelligence
Every day we read hundreds of stories. Only the most important ones are here — short, in our own words, with a link to the original source.
- Cybersecurity · Artificial intelligence
OpenAI teams up with Hacktron on vulnerabilities
OpenAI is working with Hacktron to find vulnerabilities. The partnership also tests connections with third-party services and aims to strengthen defenses with frontier AI.
Full story at openai.com ↗ - Networking · Open source · Artificial intelligence
NVIDIA open source tools support 6G research
NVIDIA open sourced its Aerial wireless software a year ago, joining Sionna and other tools. Researchers use them to link simulation and AI training with experiments on real over-the-air systems and to build 6G-ready networks.
Full story at NVIDIA ↗ - Artificial intelligence
Predicting molecule behavior from limited data
Finding new molecules for drugs or materials is still largely trial and error. Recent progress in AI and machine learning aims to supplement or speed up this laborious work.
Full story at MIT News ↗ - Cybersecurity · Artificial intelligence
OpenAI teams up with Hacktron on security testing
OpenAI is partnering with Hacktron to find vulnerabilities in its systems. The collaboration also tests third-party service connections and strengthens defenses using frontier AI.
Full story at openai.com ↗ - Artificial intelligence
Australian universities expand ChatGPT access
Australian universities are giving students broader access to ChatGPT and integrating AI into teaching. The aim is to prepare graduates for an AI-enabled workforce.
Full story at openai.com ↗ - Artificial intelligence · New AI models
OpenAI unveils benchmark for lab-science AI agents
OpenAI introduced BioBasicsBench, a 100-task benchmark testing how AI agents handle everyday wet-lab science. It measures how well these agents interpret raw experimental data.
Full story at openai.com ↗ - Artificial intelligence · New AI models
AI model swears hundreds of times in benchmarks
Thinking Machines' Inkling 2 Small cursed 316 times across 2,655 benchmark runs, according to an analysis by Vals AI. The profanity mostly appears in long runs that fail, but it is not a reliable warning signal since most failed runs had no cursing.
Full story at The Decoder ↗ - Chips & hardware · Smartphones & apps · Artificial intelligence
Smart ring adds a screen for AI control
The Aina Ring is a wearable device featuring a tiny display and a predictive operating system. It is now available for pre-order.
Full story at PhoneArena ↗ - Artificial intelligence · Open source
ToyBlocks offers controlled setting for robot learning
Hugging Face released ToyBlocks, a controlled environment for training and evaluating robot learning. It aims to help researchers test robot learning algorithms more consistently.
Full story at huggingface.co ↗ - Artificial intelligence · Chips & hardware
NVIDIA speeds up weight updates in reinforcement learning
NVIDIA introduced the Reshard Collective, a method to refresh model weights faster during RL post-training. It improves efficiency when inference and training run on separate GPU resources.
Full story at NVIDIA Developer ↗
An AI writes the short texts from the source's stories.
Free by e-mail
Choose your topics. You only get the most important news on them.