OpenAI ships Daybreak security AI agent, Murati's startup targets interaction models

OpenAI launched Daybreak, an AI initiative that uses the Codex Security AI agent to detect and patch vulnerabilities before attackers find them. The system builds on the Codex Security AI agent that launched in March.

Security engineers should evaluate Daybreak against existing SAST tools in your pipeline. Engineering leads should audit current vulnerability detection workflows before this becomes generally available.

Thinking Machines, founded by former OpenAI CTO Mira Murati, announced it's building interaction models for AI systems. The company provided no technical specifications or release timeline.

AI engineering leads should track Thinking Machines for framework releases that could challenge existing agent tooling. Platform engineers should continue current agent framework evaluations without waiting for vaporware.

OpenAI launched a deployment company focused on legal technology applications. Single source — verify before acting.

CTOs in regulated industries should monitor this deployment model for compliance and data residency patterns. Engineering leaders should evaluate whether vertical-specific OpenAI offerings change build-vs-buy decisions for domain AI.