Four Months, Not Four Years: Capability, Authority, and a Clock Recognition Was Never Built to Match

In June 2026, a change in United States export licensing suspended public access to two frontier artificial intelligence systems within three days of their release, restored partial access to one of them, for a narrow set of vetted institutions, within a further two weeks, and restored the other to general availability a week after that, all without any court, legislature, or philosophical body reaching, or attempting to reach, a conclusion about what kind of thing had been released. This essay takes that episode as a concrete instance of a more general pattern: that whatever governs access to, and authority over, an advanced artificial system currently operates on a timescale entirely disconnected from the timescale on which recognition of that system's status would need to be decided. Volume I of this series showed that the bench deciding questions of personhood has never had to justify its own composition; this volume shows that the bench, whatever its composition, is not the fastest-moving actor in the room. Drawing on a published measure of how quickly the length of tasks an artificial system can complete autonomously has been doubling, on the company most directly responsible for advancing that measure and its own account of what the trend implies, on a framework for identifying authority exercised without formal authorization, and on an existing, entirely legal mechanism by which a sufficiently capable system could already exercise practical control over property without any legislature's involvement, the essay argues that capability and consequence are accumulating on a clock measured in months, while recognition, on the evidence of Volume I alone, has historically run on a clock measured in years, decades, or centuries. It engages directly with a recent philosophical argument that this mismatch may make the entire vocabulary of controlling such a system, not merely recognizing it, unstable, and with the methodological objection that the pace of the capability clock is itself less certain than it is usually presented as being. It closes by examining the position of the company that produced the central measurement this essay relies on, an organization simultaneously demonstrating the trend and warning that the institutions meant to keep pace with it cannot.