Will AI make cybersecurity obsolete or is Silicon Valley confabulating again?

ZDNet reviews new AI-driven code security tools from Anthropic (Claude Code Security), OpenAI (Aardvark), and Google DeepMind (CodeMender), noting their ability to find and patch long-standing vulnerabilities and the potential to disrupt traditional AppSec and observability vendors. The article argues these capabilities are valuable for reducing avoidable software flaws but cannot substitute for the broader cybersecurity ecosystem—firewalls, SIEM, endpoint security, human-run incident response—or address deeper risks from agentic AI design and complex software artifacts.