alt Hacker NewsInteresting to see Anthropic now downplaying the new vulnerabilities that Mythos discovered:
> We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly-available models are able to discover them as well without requiring a bypass
Replies
> Interesting to see Anthropic now downplaying the new vulnerabilities that Mythos discovered:
That is absolutely NOT what is being said there.
They are referring to a very specific thing that you must have clearly seen and chosen to ignore—a jailbreak for LLMs that is used on other models and to some effect with Fable 5.
Their hubris just became lethal for their business. Whoops, I guess.
AFAICT this is not talking about Glasswing stuff. They are saying that they were sent a demonstration of Fable 5 being used/abused in some specific way that led to the "discovery" of some minor, already-known vuln, and that other models can find it too. IOW, they're claiming that the USG's complaint is baseless and dumb.
I think what they're saying is that this prompt/jailbreak only lets Mythos discover some really easy vulnerabilities that it probably fixes from a simple "Find and fix bugs in this code" and that this can be easily done by other models like GPT-5.5. Which is very different from targeted security research.