If there are potentially hundreds of millions of nodes the user has access to, then that uid list will be stored in memory, correct? So that would cause an issue if multiple search requests are being called at once? Would cause a huge burden on compute resources, right?