For the almost three decades that I have worked in the domain, cybersecurity operated on a reassuring asymmetry: Attackers needed extraordinary skill, time, and patience to find and exploit software vulnerabilities, while defenders, though perpetually behind, could at least count on that human bottleneck to slow the threat. Claude Mythos Preview, Anthropic’s frontier AI model announced on April 7, has shattered that assumption.
n controlled evaluations, Mythos Preview could execute multi-stage attacks on susceptible networks and discover and exploit vulnerabilities autonomously, tasks that would take human professionals days of work. Mythos has a fundamentally different architecture from its predecessors, enabling it to chain multiple small vulnerabilities into a single devastating attack, reconstruct source code from deployed software to find exploitable weaknesses, and, once inside a network, automatically map systems, move laterally, and build custom tools to extract data — all within hours.

Link: https://indianexpress.com/article/opinion/columns/mythos-shows-that-in-age-of-ai-no-system-is-safe-10657202/lite/