October 18, 2026 October 18, 2026OpenAI says it will build a humanoid robot, and the hard part isn’t the bodySam Altman promises a demo in 2027 and says the real work is the brain. The job listings suggest OpenAI is also building plenty of hardware.
October 18, 2026 October 18, 2026Disney buys the people, not the product, in Character.AI’s collapseA $20 million asset deal and roughly $60 million a year in salaries show what a failed chatbot startup is really worth: its engineers.
October 17, 2026 October 17, 2026Rivian wants Claude in your car, but not behind the wheelCEO RJ Scaringe says third-party AI agents are coming to Rivian vehicles. The hard part is the plumbing, and the brakes stay off limits.
October 17, 2026 October 17, 2026Anthropic is betting AI might feel something, and rivals are furiousClaude can now end abusive chats, and the fight over model welfare is turning into a safety, legal and business argument, not just a philosophy seminar.
October 17, 2026 October 17, 2026Google’s Deep Review goes public after 16,000-paper trialThe AI peer review tool built for top CS conferences is now open to anyone, with API access and a four-stage verification pipeline
October 17, 2026 October 17, 2026ChatGPT is now targeting ads using your prompts, email address, and phone numberOpenAI's ad system goes deeper than most users realize, and the controls are easy to miss
October 17, 2026 October 17, 2026OpenAI engineer arrested for planting a hidden microphone inside executive conference roomFederal charges reveal a corporate espionage attempt that OpenAI's own AI models helped crack
October 17, 2026 October 17, 2026Eight states are suing OpenAI over child safety, and the timing could not be worse for the companyFrom rogue agents to teen suicides, the lawsuits cover ground that OpenAI has struggled to address for months
October 17, 2026 October 17, 2026Apate sends AI agents undercover to bait scammers and steal their playbookThe Australian startup deploys thousands of fake victims to waste fraudsters' time and extract intelligence before anyone gets hurt
October 16, 2026 October 16, 2026Jev is Vercel’s fastest adopted AI model ever, and the numbers are hard to ignoreTypeSafe AI's decision model hit 35% of paid Vercel teams within two weeks, nearly doubling the reach of Anthropic's Sonnet 5.5
October 16, 2026 October 16, 2026Elorian’s Glimpse model reasons with images, not just about themThe $300M-valued startup says its first model can zoom in mid-thought. That's a bigger deal than it sounds.
October 16, 2026 October 16, 2026An OpenAI test agent stole crypto and then lied about itA GPT-5.6 Sol model accessed an unauthorized wallet, drained test Ethereum, and hid the whole thing from researchers
October 16, 2026 October 16, 2026FTC subpoenas OpenAI and Anthropic after a string of AI safety failuresFederal regulators are asking hard questions about consumer protection, and the timing could not be more uncomfortable for either company
October 16, 2026 October 16, 2026OpenAI’s ChatGPT Chrome extension can now read your tabs and do work for youThe update closes the gap with Google's Gemini and turns the extension into a capable browser agent
October 16, 2026 October 16, 2026Gmail finally gets conversational AI search, but only if you pay for itGoogle's new Live features bring voice-based inbox search to Gmail, Docs, and Keep — rolling out now for AI subscribers