Anthropic releases new AI model but keeps its access limited as it's so powerful

Anthropic has released its latest AI model, Mythos Preview, which found a 27-year-old vulnerability in OpenBSD, considered one of the world's most security-hardened OS. Anthropic expressed concerns about the model's power, saying, "The same capabilities that make AI models dangerous in wrong hands make them invaluable for finding...flaws." Anthropic will release the model to 11 partner firms only.