TITaha Iqbal

Policy Analysis

Governance Research

Long-form essays on AI regulation, compute governance, geopolitical competition, institutional coordination, and catastrophic risk policy.

Topics Covered

AI RegulationCompute GovernanceGeopolitical CompetitionInstitutional CoordinationOpen-Source Model RiskAlignment PolicyCatastrophic Risk
GovernanceInstitutional RiskAgencyCompute Gov.

Beyond the Threshold: Designing Compute Governance That Actually Works

Why compute thresholds alone will fail — and what a complete three-layer governance architecture requires before the efficiency frontier makes them obsolete

The debate about compute governance is currently structured as a binary: either FLOP-based thresholds are the right tool for managing frontier AI risk, or the DeepSeek result has made them obsolete. Both positions are wrong. Compute governance is neither…

June 4, 2026 20 min read
GovernanceForecastingScaling LawsInstitutional RiskCompute Gov.

From Benchmark to Threat Model: Using Capability Trajectories to Assess Catastrophic AI Risk Windows

Why intelligence community methodology transforms benchmark data into strategic warning

The intelligence community has a concept called a warning failure — a situation where adequate signals existed but were not integrated into a coherent threat assessment before catastrophe arrived. We are currently in the early stages of an AI warning failure.…

June 4, 2026 14 min read
GovernanceForecastingCompute Gov.

Scaling Laws as Governance Clocks:

Identifying Intervention Windows Before Threshold Capabilities Arrive

Scaling laws are not just about how models improve. When used carefully, they can serve as a forecasting tool capable of predicting when governance-relevant capability thresholds are likely to emerge. My argument is that the empirical transition from Kaplan…

May 31, 2026 12 min read

Additional essays on alignment policy and institutional design forthcoming.