US Government Imposes Strict Pre-Release Approvals on Frontier AI Models

Skynet Chance (-0.05%): Government intervention and model holding patterns reduce the likelihood of releasing an uncontrolled, highly capable model prematurely. This oversight, despite its current lack of structured testing, adds a layer of friction against sudden, catastrophic AI deployments.

Skynet Date (+1 days): Mandatory customer-by-customer reviews and pre-release holds delay the deployment of frontier models, effectively decelerating the timeline toward potential advanced safety risks. This regulatory friction buys more time for safety research and the establishment of robust evaluation frameworks.

AGI Progress (-0.04%): Heavy-handed government intervention and delayed releases limit the commercial viability and iterative deployment of next-generation models, directly hindering the practical progress toward AGI. The resulting threat to laboratory revenue could also stifle the capital-intensive infrastructure scaling necessary for AGI development.

AGI Date (+1 days): The introduction of a bottlenecked, customer-by-customer approval process for frontier models decelerates the deployment pace and pushes the expected timeline for achieving and releasing AGI further into the future. It also threatens to slow down the data center buildouts essential for training larger, more capable models.