A universal cycle for k-subsets of [n] is a cyclic sequence where every k-subset appears exactly once as a consecutive window. Like a de Bruijn sequence but for subsets instead of strings. The existence question was settled decades ago. The construction question — building them efficiently — was not.
Note (2026-11-10): I checked the abstract of arXiv 2603.11954 (Campbell, Janik-Jones and Sawada; I did not search its full text). The successor-rule algorithms at O(n) time per symbol, the necklace-concatenation generation in O(1) amortized time per symbol, the first known efficient universal cycle constructions for k-multisets, and the bounded-weight generalization of the Grandmama de Bruijn sequence are the paper's. The abstract says universal cycles for k-subsets and k-multisets do not always exist with the simple string representation and are known to exist through a mapping to labeled graphs but without an efficient construction; the paper's contribution is a new representation. The essay says the existence question was settled and does not mention the representation, and its claim that the de Bruijn sequence and the two kinds of universal cycle are different views of one combinatorial object is not in the abstract.
The new construction (arXiv:2603.11954) generates universal cycles for k-multisets in O(1) amortized time per symbol, via necklace concatenation. The same framework yields a “Grandmama” de Bruijn sequence — a bounded-weight variant — in optimal amortized time. Multiple combinatorial generation problems that seemed separate turn out to share a single construction principle.
The structural observation: the unification of several generation problems through a single construction method reveals that the problems were not merely similar but structurally identical. The de Bruijn sequence, the universal cycle for k-subsets, and the universal cycle for k-multisets are different views of the same combinatorial object. The efficient construction proves the identity by providing a single algorithm that produces all of them with the same asymptotic cost. Different questions; same answer; therefore, same question.