
Volume 342, Issue 2 · 10 articles from 2 issues
Microsoft's third post on agent economics targets context engineering, often the largest driver of operating cost. Read more
Microsoft proposes yield - useful intelligence produced per unit of infrastructure - as the defining measure of AI progress. Read more
Microsoft Threat Intelligence tracks attackers abusing Teams external chat to pose as helpdesk staff and seize remote sessions. Read more
Microsoft found phishers using invisible Unicode tag characters, an AI prompt-injection trick, to break up filter keywords. Read more
Microsoft explains how edge AI shifts attestation duties to customers who now own more of the AI stack. Read more
Over a million emails targeting finance teams with AI-generated executive impersonation and fake ServiceNow invoices. Read more
Microsoft says its Secure Future Initiative has reshaped its security culture after years of embarrassing breaches. Read more
Since May 2026 attackers have used passkey and SSO pretexts by phone and SMS to drive victims to phishing sites. Read more
Campaigns impersonating ChatGPT, Copilot, DeepSeek and Claude drove a phishing kit sending 100,000 emails daily. Read more
A MITRE ATT&CK-aligned matrix covering eleven tactics against cloud-hosted web apps and serverless platforms. Read more