Minutes into a call to demo a search API rebuilt to answer in under a second, the system got blocked, badly, in front of the client.
More curated notes connected through AI Engineering and Model Evaluation.
OpenAI's 22-page Frontier Governance Framework maps its frontier-model processes to California's Transparency in Frontier AI Act and the EU AI Act's general-purpose AI code. It documents lifecycle risk assessment, cyber-offense and other risk tiers, mitigation and residual-risk decisions, critical-incident handling, security risk management, model reporting, external review, responsibility allocation, and change control.
OpenAI proposes a five-pillar strategy for AI-enabled cyber defense: tiered access for trusted defenders, faster government-industry coordination, stronger protection of frontier models and infrastructure, risk-scaled deployment monitoring, and broader defensive support for individuals and small organizations.
The roadmap addresses evaluation results that overstate real-world performance or fail to transfer across deployment contexts. Its five steps balance standardized and local tests, evaluate throughout the lifecycle, build qualified assurance and communication capacity, tailor tests to each value-chain actor and technology, and use a coordinated, trusted process for updating methods.