Abstract: From the constitutional Founding until Erie Railroad Co. v. Tompkins, 304 U.S. 64 (1938), federal courts routinely applied a body of “general law” to resolve a wide range of legal issues, including issues relating to commercial law, torts, international law, conflict of laws, equity, and procedure. This general law regime became politically fraught and difficult to administer, and the Supreme Court finally repudiated it in Erie, declaring that “there is no federal general common law” and requiring the federal courts to ground all rules in either federal or state law. Erie and its progeny created what this Article terms the “Erie algorithm,” which requires federal courts to convert pre-Erie rules premised on one understanding of law and federal judicial power into rules premised on different understandings of both, almost always with a different meaning and different legal status than prevailed at the Founding and in the nineteenth century. This algorithm undergirds nearly every corner of contemporary federal courts doctrine. Despite its willingness to rethink other major structural constitutional law precedents, the Supreme Court has accepted—and, indeed, embraced—the Erie algorithm and incorporated it throughout modern public law. By contrast, a number of scholars in recent years—primarily but not exclusively originalists—have sought to revive the pre-Erie general law. “General law revivalism,” this Article argues, overlooks how dysfunctional the general law regime had become before Erie and fails to appreciate the incompatibility of that regime with the post-Erie constitutional order. In addition, those who suggest jettisoning Erie have not made the case for absorbing the massive system costs that such an effort would generate. The Article concludes that many strands of originalism are left with “the problem of 1938”: the necessity of reconciling originalist commitments with a legal system fundamentally shaped by Erie and its transformational algorithm.