Fabian Jahr's Brink grant covers four strands: investigating innovations enabled by Schnorr signatures, dedicated in-depth code review of significant Bitcoin Core initiatives, work on ASMap, and alternative code hosting solutions.
ASMap is the most concrete of those to explain. A node's defence against an eavesdropper or a partition attack depends on its peers being spread across the network, and using raw IP address ranges to judge that is easy to game, because one operator can hold many ranges. Mapping peers to the autonomous systems that actually route them gives a better measure of whether a node's connections are genuinely independent.
The code hosting strand is a governance question wearing engineering clothes: a project with no legal entity, whose review history is its institutional memory, has a real problem if that history lives on a single company's platform. Jahr has said a full-time grant is what allows him to review intricate proposals thoroughly enough to understand their implications for the network.