Anthropic's super-scary bug hunting model Mythos is shaping up to be a nothingburger
Theregister
Anthropic's Mythos AI model, despite being hyped as a dangerous vulnerability-finding tool, appears to be an over-marketed, human-level productivity aid.Summary
Anthropic's Mythos model, developed to identify software vulnerabilities, has faced criticism regarding both its security and its actual capabilities. While Anthropic initially restricted access due to safety concerns, unauthorized users recently accessed the model via a third-party vendor. Security experts and researchers now argue that the model is significantly over-hyped; it functions as an efficient tool for human researchers rather than an autonomous 'zero-day machine.' Analyses suggest that the vulnerabilities it identifies can also be found by humans or existing models, and that Anthropic's claims regarding the scale of its impact may be exaggerated.
(Source：Theregister)