"Anthropic claims its Claude Opus 4.6 can barely find zero-days, but Mythos Preview can pop up working exploits 72.4 percent of the time. If it lives up to its hype, Mythos would crash the internet in a day."
"Project Glasswing is generously offering free access to Mythos Preview, which Anthropic claims surpasses all but the most skilled humans at finding and exploiting software vulnerabilities."
"Anthropic claims it has found a 27-year-old bug in OpenBSD, a 16-year-old vulnerability in FFmpeg's video encoding code, and a new set of chained exploits in the Linux kernel."
"Even if Mythos is effective at finding bugs, the pressing question remains: who will fix those bugs? The challenge of addressing these vulnerabilities is significant."
Project Glasswing, backed by tech giants, is investing $100 million in AI resources to uncover vulnerabilities in open source software using the Mythos AI program. Mythos Preview reportedly generates working exploits with a 72.4 percent success rate. Despite the funding and claims of effectiveness, skepticism remains about whether this initiative can adequately secure the vast majority of software that relies on open source. The challenge of fixing identified vulnerabilities also raises concerns about the overall impact of this program on software security.
Read at Theregister
Unable to calculate read time
Collection
[
|
...
]