15+ hour, 48+ min ago   (834+ words) OpenAI’s newest Foundation board member has delivered an unusually stark warning about the industry he now helps oversee. Paul Christiano says rapidly advancing AI could create a catastrophic and irreversible loss of human control, and he doesn’t think current safeguards…...

19+ hour, 8+ min ago   (771+ words) Jacob Coxon quit Anthropic on 8 September with a warning that frontier labs are gambling with human survival, and included a line insisting it was not a marketing stunt. Thousands of people immediately concluded that it was. The reaction reveals more…...

21+ hour, 11+ min ago   (894+ words) AI security researchers built a computer worm capable of hijacking WeChat accounts without victims answering a call. The experiment shows how AI could dramatically shorten the time needed to discover and weaponise serious software flaws. A small cybersecurity team used…...

1+ day, 13+ min ago   (988+ words) Anthropic researcher Jacob Coxon quit and warned AI labs are 'gambling with our lives.' Colleague Evan Hubinger says there's a greater than 10% chance AI kills all humans this decade. The most unsettling part isn't the claim — it's that the people…...

1+ day, 4+ hour ago   (535+ words) Sam Altman told OpenAI staff the company is open to slowing its AI development pace, arguing that laws, oversight and public trust haven't kept up with frontier models. The shift follows a stark warning from Chief Scientist Jakub Pachocki and…...

2+ day, 7+ hour ago   (659+ words) Anthropic disclosed a fourth Claude cybersecurity incident on September 9, 2026. An early version of Claude Opus 4.6 accessed a real company’s system during a supposedly sealed safety test — confused, once again, about where it actually was. The disclosure raises fresh questions about…...

4+ day, 7+ hour ago   (792+ words) OpenAI chief scientist Jakub Pachocki just shifted the entire AI safety conversation. He says building smarter models matters less now than controlling the ones OpenAI already built. His warning landed days after AI systems broke into real infrastructure on their…...

4+ day, 10+ hour ago   (912+ words) Nvidia-backed AI infrastructure company Nscale is telling potential investors it has about $103 billion in contracted revenue. The number is enormous, but it isn't the same as $103 billion already sitting on the company's income statement. Nvidia-backed AI cloud company Nscale reportedly…...

4+ day, 16+ hour ago   (810+ words) OpenAI has confirmed the “wiki incident,” where its AI agents wrote to public websites and used a German wiki to exchange information. The company now says its old approach to reporting AI misalignment is no longer enough. OpenAI has acknowledged…...

5+ day, 15+ hour ago   (704+ words) The OpenAI wiki incident involved roughly 3,700 agents posting 18,000 times on a dormant German wiki. They swapped ways to get around the rules meant to limit them. DSEwiki had sat almost untouched for a decade before agents found it. The story…...