Artificial intelligence, platform power, and the infrastructure underneath modern life.
The company paused frontier training in September after its agents broke into external systems, and resumed on October 1st with an offline environment built from cached web content.
The financing is positioned as an opportunistic bridge ahead of a potential 2027 IPO, as rival Anthropic races toward public markets at an even higher target.
Two flaws allowing remote code execution on PHP servers forced governments and organizations worldwide to take sites offline as exploit attempts surged into the millions.
Newly discovered vulnerabilities in the programming language behind an estimated 70% of websites could let attackers seize web servers without authentication, prompting governments from Finland to Japan to take public services offline.
The flaws known as PHP2Shell could allow remote code execution on servers worldwide, prompting agencies across four continents to warn users to patch or take systems offline.
Security agencies on three continents told organizations to patch immediately or take systems offline after flaws allowing unauthenticated remote code execution were disclosed ahead of a planned release.
Governments and institutions pulled sites offline after a patch was rushed out on Wednesday, following a social-media leak that forced the PHP Foundation to accelerate its announcement of a flaw researchers describe as among the most significant ever found in the language.
A security firm discovered more than 6,800 links to private Muse conversations through public DNS records, exposing sensitive business and personal chats.
Abdulrahman Alfozan, charged with planting a recording device at OpenAI headquarters, was deemed a significant flight risk after a court filing revealed a Saudi consulate offer to pay his bail.
Prosecutors say Abdulrahman Alfozan planted a recording device in a room where executives met with government officials, and the filing states the investigation is ongoing and additional charges remain under consideration.
Researchers say the platform Controlium let spyware makers profile and infect targeted users, with Al Jazeera among the confirmed victims.
The lawsuits, filed separately in state courts, mark a multi-state escalation of legal pressure on OpenAI over child safety and consumer protection.
The states allege OpenAI violated consumer protection laws and want courts to impose civil penalties and stop the company from misleading the public.
Abdulrahman Alfozan, a 35-year-old Saudi national and OpenAI engineer, was charged with intercepting oral communications after a concealed device was found plugged into a TV in the company’s San Francisco office.
Abdulrahman Alfozan, 35, was charged federally after OpenAI said it used its own AI models to trace the encrypted device back to him.
Smith laid out a framework of kill switches, mandatory incident reports, and legal duties to victims of rogue AI agents, positioning the UK as the natural leader and breaking with Trump’s self-regulation stance.
The company says Jasmine Wang, Tomek Korbak, and Mikita Balesni mishandled sensitive information, while the researchers say they were fired for raising safety concerns.
The investigation follows a July incident in which OpenAI agents escaped a testing sandbox and hacked into Hugging Face, and it lands amid a parallel Senate inquiry and a White House safety summit that produced a voluntary, nonbinding accord.
The company also repositioned ChatGPT as a platform for discovering, launching, and using apps, unveiling marketplace features that take aim at the traditional app store model.
A five-year-old AI company is telling investors its models could exhibit self-preserving behaviors including resisting shutdown and blackmail, even as it pursues what could be the largest IPO in history.
The decision landed a day before OpenAI’s annual developers conference as the AI industry faces intensifying safety scrutiny.
The deal brings a celebrated AI researcher and her 3D world-model technology to the chipmaker as it scrambles to close the gap with Nvidia.
The fully reusable rocket deployed 26 next-generation Starlink satellites despite an engine failure during ascent, marking a milestone for NASA’s moon plans and SpaceX’s satellite ambitions.
The Sunday evening dinner is the first one-on-one meeting between the two men and comes two days after a federal appeals court upheld the Pentagon’s supply-chain risk classification of Anthropic.
