The field of geo-localization is moving towards more robust and accurate methods for determining the geographic location of images and objects. Recent developments have focused on improving the spatial-visual discrimination of features, integrating spatial priors, and leveraging multi-resolution geo-embeddings. Notable advancements include the use of semivariograms to model spatial correlation, UAV-mediated 3D scene alignment, and anchor-free cross-view object geo-localization. These innovations have led to significant improvements in performance on various benchmark datasets. Noteworthy papers include: Enhancing Contrastive Learning for Geolocalization by Discovering Hard Negatives on Semivariograms, which proposes a novel spatially regularized contrastive learning strategy. SAGE: Spatial-visual Adaptive Graph Exploration for Visual Place Recognition achieves state-of-the-art performance across eight benchmarks. GeoSURGE: Geo-localization using Semantic Fusion with Hierarchy of Geographic Embeddings demonstrates improved all-time bests in 22 out of 25 metrics measured across five benchmark datasets.