Anthropic's powerful AI model leaked via data cache

Anthropic's most advanced AI model, codenamed Capybara, was inadvertently exposed through an unsecured data cache. This new model is reportedly more capable than any previous Anthropic creation. The company has highlighted significant cybersecurity risks associated with this leak, emphasizing the unprecedented nature of the potential threats. The exposure raises concerns about the security of highly advanced AI technologies and their potential misuse.