The paper studies a class of mathematical objects called automorphic forms, which are highly symmetric functions that encode deep connections between number theory, geometry, and representation theory. The specific setting involves 3-by-3 matrices with integer entries and determinant 1, forming a group called SL(3, Z). The authors focus on certain subgroups of this group, called Iwahori congruence subgroups, defined by requiring that the matrices look upper triangular when you reduce their entries modulo a prime number p. The central question is whether these subgroups carry nontrivial "cuspidal cohomology," which is a way of detecting special, genuinely new symmetric functions that are not built from simpler pieces.
To carry out their computations, the authors extend earlier technical machinery developed in the 1980s by Ash, Grayson, and Green, which was originally designed for slightly different subgroups. They then systematically check all primes p up to 227, computing the relevant cohomology groups and looking for nonzero cuspidal classes. Out of all those primes, they find exactly three where nontrivial cuspidal classes exist. For two of those primes they go further and explicitly compute how certain algebraic operators called Hecke operators act on these classes, which gives detailed arithmetic information about the underlying automorphic forms.
The significance of these examples is that they are the first known instances of a particular type of automorphic representation, called non-essentially-self-dual, appearing at Iwahori level with trivial cohomological weight and a property known as being Steinberg at p. Being non-essentially-self-dual means the representation cannot be matched with its own dual in the usual way, making these objects rarer and harder to find than their self-dual counterparts. Discovering concrete examples of them through explicit computation provides new data points that researchers can use to test and refine broader conjectures in the Langlands program, an ambitious web of conjectures connecting number theory and representation theory.