Search results for Armadin
AI & Enterprise
China\'s open-source AI GLM-5.2 could lower bar for cyber attacks, researchers warn
China AI startup Z.ai’s open-source model GLM-5.2 is raising concern among cybersecurity researchers that it could accelerate the spread of hacking capabilities, Axios reported. The model is described as having agentic capabilities comparable to Claude Opus4.8 and OpenAI GPT-5.5 at about half the operating cost. Security assessments found performance on par with leading U.S. models. Researchers say jailbreak methods are being shared on Russian-language hacker forums.
AI & Enterprise
Palo Alto Networks says Mythos finds vulnerabilities in 3 weeks matching a year of manual penetration testing
Palo Alto Networks introduced its Frontier AI Defense initiative, saying AI-driven cyber threats have entered a new phase after testing frontier models. The company said three weeks of AI-assisted analysis matched a year of manual penetration testing. It also flagged exploit chaining that links low-risk flaws into critical attack paths and said AI-assisted attacks cut the time from initial access to data theft to 25 minutes. It warned most organisations lack visibility into employee-created code as local AI agents spread.
AI & Enterprise
AI agent hackers are coming, a turning point in cybersecurity history
Anthropic warned in a recently leaked blog draft that new AI models under development, including Claude \"Mythos\", could find and attack security vulnerabilities at unmatched speed. It said Mythos signals a wave of models that exploit flaws faster than defenders can respond. Axios reported Anthropic is letting some institutions test Mythos and privately warning government officials. Security executives said AI agents could outpace human hackers, while others noted limits in understanding organisational context.
