What does Anthropic's Mythos, with its ability to autonomously find and weaponize software vulnerabilities, mean for the future of cybersecurity? Affiliate Bruce Schneier and Barath Raghavan suggest that Mythos is the first in (what's sure to be) a series of incremental steps toward long-term changes in the relationship between AI and security. The authors note that not all technologies will be patched at the sane rate, so we may see "a few years of constant hacks" while software companies catch up. Read more from IEEE Spectrum.