Anthropic research covered by The Decoder suggests AI systems can build exploits from security patches within hours rather than weeks. That compresses the window defenders have after vulnerability details become public.
The finding matters because patch releases can act like roadmaps. If AI can quickly infer exploit paths from the changes, organizations may face faster attacker adaptation after every disclosure.
For security teams, the practical takeaway is urgency: patch latency, monitoring, and exploit prediction become more important as AI lowers the cost of turning public fixes into working attacks.