Recombinant gene delivery vectors derived from naturally occurring or genetically engineered adeno-associated viruses (AAV) have taken center stage in human gene therapy, fueled by rapidly accumulating and highly encouraging clinical data. Nonetheless, it has also become evident that the current generation of AAV vectors will require improvements in transduction potency, antibody evasion, and cell specificity in order to realize their full potential and to widen applicability in larger patient cohorts. Fortunately, in the recent past, the field has seen a flurry of exciting new developments that enhance our understanding of AAV vector biology, including virus–host interactions, and/or that expand our arsenal of technologies for AAV capsid design and evolution. This review highlights a collection of latest advances in these areas, which, in the authors' opinion, hold particular promise to propel the AAV vector field forward in the near future, especially when applied in combination. These include fundamental novel insights into the AAV life cycle, from an unexpected role of autophagy and interactions with other viruses to the (re-)discovery of a universal AAV receptor and the function of AAV-AAP for capsid assembly. Concurrently, recent successes in the rational design of next-generation synthetic AAV capsids are pointed out, exemplified by the structure-guided derivation of AAV mutants displaying robust in vivo immune evasion. Finally, a variety of new and innovative strategies for high-throughput generation and screening of AAV capsid libraries are briefly reviewed, including Cre recombinase-based selection, ancestral AAV capsid reconstruction, and DNA barcoding of AAV genomes. All of these examples showcase the present momentum in the AAV field and, together with work by many other academic or industrial entities, raise substantial optimism that the remaining hurdles for human gene therapy with AAV vectors will (soon) be overcome.