The recent developments in the research area of matching markets and algorithms have shown significant advancements in several key areas. There is a notable shift towards addressing more complex and realistic scenarios, such as many-to-many matching with preferences and costs, and the integration of learning mechanisms in decentralized markets. Innovations in approximation algorithms for weighted Nash Social Welfare with submodular valuations have been particularly impactful, providing the first constant approximation algorithm for this challenging problem. Additionally, the field has seen progress in handling ties in stable matching problems, with new approaches to guarantee logarithmic optimal stable share ratios. Online Budgeted Matching with general bids has also been tackled, removing assumptions that limited previous algorithms and introducing novel meta algorithms with provable competitive ratios. Lastly, the incorporation of spacing objectives in budgeted auctions, focusing on the distribution of wins over time, has led to efficient learning algorithms with low regret, addressing practical concerns in various domains. These developments collectively push the boundaries of what is possible in matching theory and its applications, emphasizing the importance of considering real-world constraints and leveraging learning to improve outcomes.