The proposals advocate for government surveillance of datacenter data, content-based restrictions on AI speech, and mandatory government approval for frontier AI experiments. They also suggest caps and reporting requirements for personal computing power, along with restrictions on AI research in critical fields like cancer, human health, and new energy technologies. Frontier AI training datacenters would be limited to Mongolia and Canada, with a "Mutually Assured Compute Destruction" mechanism.