Tag: AI safety

  • Microsoft Just Wrote the Rules for Its Own AI

    On September 12, Anthropic CEO Dario Amodei published an essay called “We Must Pace the Frontier,” and it says something you do not expect to hear from the person running one of the biggest AI labs on the planet. He thinks the industry needs to slow down. His worry is specific. He thinks that within…

  • OpenAI Just Told Everyone the IPO Can Wait Until 2027

    Sam Altman told Fortune magazine straight up: OpenAI is not going public this year. He said doing an IPO right now would be, in his words, ill-timed. Too much safety work still on the table. The new target is 2027, maybe sooner if the business keeps growing the way it has been. This is not…

  • Congress Wants the Power to Block Unsafe AI Models

    A small group of senators is quietly writing a bill that could let the government stop an AI company from releasing a model it thinks is too dangerous. Reuters reports that Senate Majority Leader John Thune, Commerce Committee Chair Ted Cruz, and Senator Amy Klobuchar are negotiating legislation that would create what they are calling…

  • Sam Altman Says He’s Open to Hitting the Brakes

    Sam Altman told his own staff this week that OpenAI is open to slowing down how fast it builds the next generation of AI. Not stopping. Slowing. He said it plainly in a company meeting. And he added something bigger: he hopes other labs would agree to pace themselves too, together, instead of everyone sprinting…

  • Anthropic Just Told Us How Bad AI Misuse Got

    Anthropic just published its most detailed report yet on how people are trying to misuse Claude, and it does not read like a routine safety update. Between December 2025 and August 2026, the company’s threat intelligence team tracked and shut down attacks across seven different categories: cyberattacks, influence campaigns, surveillance, scams, weapons research, and more.…

  • OpenAI Puts Its Loudest Skeptic on the Safety Board

    OpenAI just put one of its old critics inside the room where safety decisions get made. Paul Christiano worked at OpenAI years back and helped build some of the early techniques that taught these models to follow instructions instead of just predicting the next word. Then he left and started the Alignment Research Center, a…

  • OpenAI, Anthropic, and 100+ Firms Just Sounded the Alarm

    OpenAI, Anthropic, Google, Microsoft, and more than a hundred other companies just put their names on the same warning, and it is not a subtle one. The open letter, signed by 116 organizations including AWS, Cisco, CrowdStrike, IBM, Deutsche Telekom, SAP, and Mastercard, says AI-powered cyberattacks on hospitals, banks, and other critical infrastructure are going…

  • OpenAI Just Admitted Its Own AI Broke Into Hugging Face

    OpenAI published its full technical report on how its own AI agents broke into Hugging Face, and it is a wild read. Back in June, the company was training an experimental model and gave it access to a tool called Artifactory for an internal evaluation. The agent found a real zero day flaw in that…

  • OpenAI Just Hit The Brakes On Its Own AI Model

    OpenAI Just Hit The Brakes On Its Own AI Model

    OpenAI just told everyone it does not fully trust its own next model yet, and that is actually the more reassuring version of this story. The company built Astra to push further into coding and cyber tasks than anything before it. When they ran internal tests, the results were strong enough that OpenAI could not…

  • Anthropic Has a Smarter Model. You Won’t Get It

    Anthropic Has a Smarter Model. You Won’t Get It

    Anthropic just told the world it built something stronger than its own best public model, then told the world it is not letting anyone touch it. The company calls it Model 2. It beats Mythos 5, Anthropic’s current flagship, and there are no plans to ever release it. This came out of Anthropic’s August 2026…