21ST CENTURY PROBLEMS: How Dangerous is Mythos? Why Did Anthropic Decide Not to Release the New Model?

Anthropic’s Mythos has been able to independently complete a whole set of attack processes. From discovering zero-day vulnerabilities, writing exploit code, to linking multiple steps to access core systems, tasks that originally required top hackers to collaborate for long periods have been compressed to hours or even minutes.

This is also why, at the moment the model was disclosed, Scott Bessent and Jerome Powell would convene Wall Street firms to ask them to conduct ‘self-checks.’ When the ability to discover vulnerabilities is released on a large scale, the financial system faces not sporadic attacks, but continuous scanning.

The deeper change lies in the supply structure. In the past, vulnerability discovery relied on the experience accumulation of a few security teams and hackers, which was slow and non-replicable. Now, this capability is beginning to be output in bulk by models, lowering the threshold for both attacks and defenses. A metaphor from an insider is very straightforward: giving the model to an ordinary hacker is akin to equipping them with special operations capabilities.

Developing…