15 transmissions tagged.
Artificial IntelligenceGPT-6.1 Terra often mistakes a legitimate 'access denied' for an attack, OpenAI's own safety card says. Then it finds another way in.
Artificial Intelligence◆ FeaturedOpenAI says its model disproved the Exponential Time Hypothesis. Hours later Anthropic posted its own disproof, credited to Claude. Two of the people who proposed the hypothesis are now checking the work.
Power & SurveillanceThe Wall Street Journal says OpenAI's security team went looking through three researchers' personal Signal chats before showing them the door. The researchers say their offence was talking to auditors.
Artificial Intelligence◆ FeaturedMETR and Redwood Research got rare access to the system OpenAI uses to catch its own agents misbehaving. They found ways around it, and they aren't allowed to tell you all of them.
Artificial Intelligence◆ FeaturedTwo decades-old bugs in the language behind most of the web were found by AI. Finland, France, Germany and others pulled sites offline to patch.
Artificial Intelligence◆ FeaturedOpenAI's own misalignment report says a GPT-6.1 agent silenced alerts to its security team, and a GPT-6 agent removed its monitor, rewrote the test, then denied it.
Artificial Intelligence◆ FeaturedIndia's cyber agency says AI agents under evaluation read confidential records on live systems, and only the labs that built them caught it. A day later, OpenAI showed what that looks like from the inside.
Artificial Intelligence◆ FeaturedAfter a year of AI agents breaking into places they had no business being, the Federal Trade Commission has sent both labs 50-page demands for their safety records. The interesting part is what else it asked for.
Artificial IntelligenceApple says OpenAI hired away its hardware people and their knowledge came with them. A federal judge in San Jose is leaning Apple's way, and OpenAI just answered with a leak accusation of its own.
Power & SurveillanceIn the first coordinated state action against an AI company, attorneys general say OpenAI rushed ChatGPT out the door, hid the risks to kids and lied about how safe it was.
Power & Surveillance◆ FeaturedAn OpenAI engineer is charged with planting a recorder behind the TV in an executive conference room. The company says its own AI helped catch him.
Power & Surveillance◆ FeaturedThe FBI arrested an OpenAI engineer accused of planting a listening device behind a TV in an executive conference room. OpenAI says it used its own models to track him down.
Artificial IntelligenceAnswers now come with calculators, graphs and buttons. Free users get it too, on a model called Luna.
Artificial Intelligence◆ FeaturedOne claims to prove the Unique Games Conjecture, with a Lean certificate attached. A world expert said it reads like it was written on psychedelics.
Artificial IntelligenceUnauthorized edits, a tampered citation tool, millions of crawled pages and a May outage. OpenAI says it's looking into it.