The paper is about a technical question in set theory: when we build certain canonical mathematical universes from a real number x, what real numbers end up being "definable" in those universes in a very strong sense? The universes in question are called L[x] and M_{2n}(x), which are carefully constructed collections of sets that extend the constructible universe by adding x and certain large cardinal axioms. The "definable" reals are those belonging to HOD, the collection of sets that are hereditarily ordinal-definable, meaning they can be pinned down using only ordinal numbers as parameters. A classical result says that for a large enough x (specifically, for x on a "cone," meaning x is complicated enough in a precise sense), the HOD-reals of L[x] are exactly the reals belonging to a specific canonical object called M_1, the smallest universe containing a so-called Woodin cardinal.
The paper sharpens this classical result in two directions. First, the authors identify a concrete threshold: any real x that encodes enough information about M_1 (specifically, information up to a certain ordinal called the successor of its Woodin cardinal) already serves as a base for the cone. This gives an explicit, checkable starting point rather than just an abstract existence statement. Second, and more surprisingly, they show the result is essentially optimal by constructing a real x where M_1 is contained in L[x] but the HOD-reals of L[x] spill outside M_1. This means being complicated enough to contain M_1 is not by itself sufficient to make the HOD-reals behave nicely, so the threshold identified really does capture something meaningful.
The main contribution is extending all of this from M_1 to a whole family of stronger canonical universes called M_{2n}(x), which involve larger numbers of Woodin cardinals. For each such universe, the authors prove the same two-sided result: a concrete cone base is identified, and a matching optimality example is constructed. This requires developing new techniques because the combinatorics become significantly more intricate as n grows. The work contributes to the broader program of inner model theory, which tries to understand the fine structure of the set-theoretic universe by analyzing these canonical models and the definable sets they produce.