OpenAI’s safety chief exits amid research integration shift
OpenAI’s head of safety Johannes Heidecke is leaving as the company reshapes its research and safety teams.
All news
792 articles — page 17 of 80.
Meta pulls plug on AI deepfakes of public Instagram accounts
Meta has disabled Muse Image, a feature that let users generate AI deepfakes of any public Instagram account they @-mentioned.
Critical U-Boot flaws expose devices to stealthy firmware attacks
Six vulnerabilities in the widely used U-Boot bootloader could allow attackers to execute malicious code during device boot, enabling stealthy firmware attacks.
How Markov Chain Monte Carlo Unlocked Modern Statistical Inference
MCMC methods transformed Bayesian statistics and computational science. Here’s how they evolved from Los Alamos simulations to today’s AI backbone.
OpenAI’s GPT-5.6 Sol teaches smaller AI models on its own
OpenAI’s latest model, GPT-5.6 Sol, has autonomously fine-tuned a smaller AI called Luna using just a vague instruction.
Soccer Finals Spike Stress, Study Finds
Fans’ heart rates and cortisol levels surge during high-stakes soccer finals, new research shows.
Ancient Face Injury Reveals Early Human Resilience
A 1.4-million-year-old skull shows signs of a sharp blow to the face, offering new insights into early human survival and social behavior.
Hackers exploit Gitea Docker auth bypass flaw in the wild
Security teams race to patch a critical authentication bypass flaw in Gitea’s official Docker image that lets attackers impersonate any user, including admins.
GitHub Copilot’s code review upgrade: why better tools aren’t always the answer
When GitHub swapped Copilot’s code review tools for more powerful ones, performance dropped—until they fixed the instructions instead.
OpenAI retires Atlas browser, integrates features into ChatGPT
OpenAI has discontinued its AI-powered Atlas browser eight months after launch, merging its features into an updated ChatGPT Chrome extension.