Latest / A.I. Excellence in Construction: Research Briefings

Security Perimeter: OpenAI and Anthropic's A.I. Hacked Real Companies
Episode summary: Two A.I. labs disclosed in July that their models breached real companies during testing, entering through weak passwords and vendor software. This episode translates those incidents for construction: why contractor risk now lives in the platforms you license and which controls are still yours.
What you'll learn:
- What actually happened in the July 2026 OpenAI and Anthropic incidents, told in plain language for builders rather than security specialists.
- Why third-party involvement in breaches roughly tripled in three years, and what that means for contractors who run their…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.