March 13, 2026
I Joked About This on Stage — Then McKinsey Got Hacked
When McKinsey's AI platform "Lilli" was breached in March 2026, 46.5 million chat messages and 728,000 confidential files were exposed — not through sophisticated AI manipulation, but through unauthenticated API endpoints and basic SQL injection. In this post, Keno Dreßel argues that the real lesson isn't about AI security. It's about old engineering failures meeting a relentless new class of attacker.