Skip to content
InfoNews

Security for the agentic era, powered by Microsoft Perception 

Published
Record updated
View JSON

Summary

Microsoft's security blog describes updates to Microsoft Perception and a new Microsoft Security cyber stack for agentic security. The post argues that frontier AI models now find and exploit software vulnerabilities far faster than before, citing PHP's two critical remote code execution flaws, where about 29 million internet-facing services remained vulnerable nearly two weeks after disclosure. A working exploit for those flaws appeared within roughly a day of disclosure.