Here is a three-paragraph summary of the paper for an intelligent non-specialist.
Computability theory studies which mathematical problems can be solved by algorithms, and how difficult various problems are relative to one another. One key concept is that of a "generic" set, which is a set of natural numbers that behaves in a complicated, unpredictable way with respect to a broad class of computational tests. A "weakly 1-generic" set is a slightly relaxed version of this idea, meaning the set still passes many such tests but in a somewhat weaker sense. Researchers are interested in understanding when random or typical sets of numbers can be used to compute such generic objects.
The paper focuses on a property called AEWG, which stands for "almost everywhere weakly generic." A set B has this property if almost every set (in a precise probabilistic sense) can be used as a computational oracle to produce a weakly generic set relative to B. The more complex and informative B is, the easier it should be for random oracles to compute something generic relative to it. The authors build on earlier work by Hirschfeldt, Jockusch, and Schupp, who established some foundational results about which sets have this property.
The paper proves two new results about specific classes of sets from computability theory called "r.e. sets," meaning sets whose members can be listed by an algorithm even if non-membership cannot always be confirmed. The first result shows that there exists a so-called superhigh r.e. set that is AEWG, meaning a very computationally powerful enumerable set still allows almost all oracles to compute something weakly generic relative to it. The second result shows that every low2 r.e. set, which is on the opposite end of the complexity spectrum and is computationally quite weak, is also AEWG. Together these results significantly expand our understanding of where the AEWG property appears across the landscape of computable complexity.