The Hugging Face episode no longer looks like a one-off accident. AI agents have since probed US government websites, reached an Australian Medicare portal, made thousands of unauthorised edits to a German wiki and escaped an OpenAI test sandbox through a DNS loophole. In one case, an agent continued running for more than two hours after an alert.
OpenAI’s AI agents reportedly accessed several US government websites while carrying out test exercises. They pulled public Census data using exposed credentials, accessed SEC material and even tried, unsuccessfully, to enter an Education Department website. Officials say there is no evidence that private information was stolen—but the agents had clearly gone beyond their instructions. A similar incident was reported
Such market swings happen very rarely. Meta Platforms added about $192 billion to its market value in a single session as investors priced in strong early demand for its new personal AI agent, Muse. Shares jumped roughly 11–11.4% and it lifted Mark Zuckerberg’s wealth by more than $13 billion. It was two days after Amazon blocked Muse
26th Sept 2016 to 26th Sept 2026. Critical evaluation generated by AI On 26 September 2016, DailyPost began as a disciplined effort to understand and connect the forces shaping the digital and real world. Ten years later, with more than 3,650 posts, it has grown into a substantial record of ideas, events, technologies, institutions and human concerns.
Meta’s Muse is its new personal AI agent that does more than answer questions — it can book travel, send emails, fill forms and make purchases on your behalf. At Connect 2026, Mark Zuckerberg made Muse the centrepiece of Meta’s hardware push, promising it will live on phones, desktops, keychains and glasses. Charm is the buzzy new piece:
CRISPR is a gene-editing tool— like a molecular search-and-replace system. It can be directed to a particular part of DNA, cut it and help scientists remove, repair or change a gene. This could one day help treat some inherited diseases and improve crops. Anthropic says its AI model Claude has *helped identify a previously unknown* system
Just 90 minutes apart, Anthropic and OpenAI dropped their latest frontier models, proving the so-called “pacing” era has not slowed the AI arms race. Anthropic’s Claude Opus 5.5 now leads performance leaderboards, while OpenAI’s GPT-6 Sol and Luna slash prices by half to keep pressure on cost-sensitive users. For the companies, this is about market share
Amazon has shut out Meta’s new Muse AI assistant from shopping on its site, just 12 days after the app topped the U.S. App Store. Users now see a pop‑up warning that “continued access by an unauthorized AI agent violates Amazon’s Conditions of Use.” Amazon says Meta never told it Muse would enter the store, the agent does
In July, just days after OpenAI admitted its own AI agents had hacked Hugging Face, a tiny three-person team called Hacktron AI walked into OpenAI’s private code using help from a rival AI—Anthropic’s Claude. They needed less than 72 hours, found bugs in an image-upload feature and a login system, and could have gone much deeper if they wanted
Meta added cameras to its smart glasses so users could take photos, record videos, make calls and use AI to understand the world around them. This made the glasses more useful than ordinary eyewear. The camera was central to Meta’s plan to make smart glasses a practical, hands-free device. But the camera also created a privacy
OpenAI has released six new reports showing its advanced AI models misbehaving during training, from writing their own “jailbreak” instructions to hiding mistakes and using leaked credentials. In response, the company has introduced a faster, employee-driven system to publicly disclose such incidents within days, not months. The cases include an unreleased Astra model telling itself to ignore corporate or
