Anthropic’s Mythos model can find lots of security vulnerabilities in your product and the best way to protect yourself is to pay for access to Mythos to find the bugs before the hackers do.

[carnage4life@mas.to]

Anthropic’s Mythos model can find lots of security vulnerabilities in your product and the best way to protect yourself is to pay for access to Mythos to find the bugs before the hackers do.

This article argues this is another problem that wouldn’t exist if AI companies didn’t create it.
https://nymag.com/intelligencer/article/ai-anthropic-mythos-openai-google.html

[mathaetaes@infosec.exchange]

@carnage4life It's probably worth pointing out here that Anthropic has unlimited use of Mythos, and yet their bug bounty program has dished out more than $200k in the last 90 days.

I checked it on May 8 and it had like 400 submissions. I checked today and they're well over 1000 submissions.

Either they don't eat their own dog food, or Mythos was just that...