"Anthropic has unveiled Claude Mythos, a new AI model capable of discovering critical vulnerabilities at scale," writes Slashdot reader wiredmikey. "It's already powering Project Glasswing, a joint effort with major tech firms to secure critical software. But the same capabilities could also accelerate offensive cyber operations." SecurityWeek reports: Mythos is not an incremental improvement but a step change in performance over Anthropic's current range of frontier models: Haiku (smallest), Sonnet (middle ground), and Opus (most powerful). Mythos sits in a fourth tier named Copybara, and Anthropic describes it as superior to any other existing AI frontier model. It incorporates the current trend in the use of AI: the modern use of agentic AI. "The powerful cyber capabilities of Claude Mythos Preview are a result of its strong agentic coding and reasoning skills... the model has the highest scores of any model yet developed on a variety of software coding tasks," notes Anthropic in a blog titled Project Glasswing -- Securing critical software for the AI era. In the last few weeks, Mythos Preview has identified thousands of zero-day vulnerabilities with many classified as critical. Several are ten or 20 years old -- the oldest found so far is a 27-years old bug in OpenBSD. Elsewhere, a 16-years old vulnerability found in video software has survived five million hits from other automated testing tools without ever being discovered. And it autonomously found and chained together several in the Linux kernel allowing an attacker to escalate from ordinary user access to complete control of the machine. [...] Anthropic is concerned that Mythos' capabilities could unleash cyberattacks too fast and too sophisticated for defenders to block. It hopes that Mythos can be used to improve cybersecurity generally before malicious actors can get access to it. To this end, the firm has announced the next stage of this preparation as Project Glasswing, powered by Mythos Preview. Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely. "Project Glasswing is a starting point. No one organization can solve these cybersecurity problems alone: frontier AI developers, other software companies, security researchers, open-source maintainers, and governments across the world all have essential roles to play." Claude Mythos Preview is described as a general-purpose, unreleased frontier model from Anthropic that has nevertheless completed its training phase. The firm does not plan to make Mythos Preview generally available. The implication is that 'Preview' is a term used solely to describe the current state of Mythos and the market's readiness to receive it, and will be dropped when the firm gets closer to general release.

Read more of this story at Slashdot.

A New York Times analysis found Google's AI Overviews now answer questions correctly about 90% of the time, which might sound impressive until you realize that roughly 1 in 10 answers is wrong. "[F]or Google, that means hundreds of thousands of lies going out every minute of the day," reports Ars Technica. From the report: The Times conducted this analysis with the help of a startup called Oumi, which itself is deeply involved in developing AI models. The company used AI tools to probe AI Overviews with the SimpleQA evaluation, a common test to rank the factuality of generative models like Gemini. Released by OpenAI in 2024, SimpleQA is essentially a list of more than 4,000 questions with verifiable answers that can be fed into an AI. Oumi began running its test last year when Gemini 2.5 was still the company's best model. At the time, the benchmark showed an 85 percent accuracy rate. When the test was rerun following the Gemini 3 update, AI Overviews answered 91 percent of the questions correctly. If you extrapolate this miss rate out to all Google searches, AI Overviews is generating tens of millions of incorrect answers per day. The report includes several examples of where AI Overviews went wrong. When asked for the date on which Bob Marley's former home became a museum, AI Overviews cited three pages, two of which didn't discuss the date at all. The final one, Wikipedia, listed two contradictory years, and AI Overviews confidently chose the wrong one. The benchmark also prompts models to produce the date on which Yo Yo Ma was inducted into the classical music hall of fame. While AI Overviews cited the organization's website that listed Ma's induction, it claimed there's no such thing as the Classical Music Hall of Fame. "This study has serious holes," said Google spokesperson Ned Adriance. "It doesn't reflect what people are actually searching on Google." The search giant likes to use a test called SimpleQA Verified, which uses a smaller set of questions that have been more thoroughly vetted.

Read more of this story at Slashdot.