Elasticsearch 8.0 introduces approximate nearest neighbor (ANN) search, enhancing vector search capabilities alongside traditional methods like term-based scoring. The surge in vector search interest arises from machine learning models that convert content into high-dimensional vectors, capturing similarities beyond surface characteristics. Traditional k-nearest neighbor (kNN) searches for vectors most similar to a query but can be inefficient for high-dimensional data. ANN algorithms, such as Hierarchical Navigable Small World (HNSW) graphs, offer a more scalable approach by trading off some accuracy for speed and efficiency, making them suitable for large datasets. Elasticsearch's ANN search, built on the Apache Lucene library, allows for efficient retrieval by integrating deeply with existing functionalities and ensuring compatibility across versions. The new _knn_search endpoint, introduced as a technical preview, efficiently retrieves similar vectors and is designed to iterate quickly, addressing open questions and integrating with Elasticsearch's core search functionalities in future updates.