AI companies make powerful tech
Anthropic’s Claude Mythos AI is said to be frighteningly capable, but we shouldn’t get carried away by the hype
TL;DR
- Anthropic released Claude Mythos, an AI model focused on cybersecurity, which they claim is adept at exposing software weaknesses.
- Mythos reportedly found thousands of vulnerabilities in commonly used applications, prompting Anthropic to form Project Glasswing with cybersecurity specialists to manage its distribution.
- Cybersecurity experts are questioning the significance of Anthropic's claims, suggesting the company is brilliant at marketing.
- Concerns about AI capabilities and their potential misuse are reminiscent of past debates, such as OpenAI's GPT-2 release.
- The article also touches on a Guardian investigation that exposed child sex trafficking on Facebook and Instagram, leading to a legal case against Meta.