Rockset announces the expansion of its platform's vector search capabilities, following the launch of its high-speed database of the same name. The new feature aims to support faster search of cloud data to facilitate artificial intelligence applications. Vector search has become an essential factor for databases supporting generative AI and other types of AI applications. It allows storing unstructured data such as images and handwriting as vector embeddings, which are mathematical structures that enable indexing and searching of the information. As a result, vector search can create more powerful and accurate AI models. By adding vector search capabilities earlier this year, Rockset became a viable database choice for generative AI models. With today's update, the Rockset database now supports "approximate nearest neighbor computation" (ANN) search, meaning it can achieve "billion-scale similarity search" in the cloud. However, as the company explains, vector search alone is not enough because the most powerful large-scale language models generate vector embeddings with thousands of dimensions, making precise nearest neighbor computation (ENN) search an extremely complex and computationally intensive task. By adding support for ANN, Rockset enables the creation of vector embeddings and fast similarity search for any AI model in the cloud. The new feature is used in conjunction with Rockset's LlamaIndex and Langchain integrations to help developers iterate faster and create more relevant AI experiences. According to the company, developers can now store and index billions of vectors and hundreds of terabytes of critical data, such as text, JSON, geospatial, and time series data. They will be able to build real-time updates for their AI applications by inserting, updating, and deleting vectors and metadata stored in the Rockset database. New data is reflected in database searches within milliseconds, meaning it is immediately available for use by AI models. Venkat Venkataramani, Co-founder and CEO of Rockset, stated that combining real-time signals and updates with vector search is not an easy task. However, without them, AI applications would not be able to realize their enormous practical value. He added, "We spent years designing Rockset to achieve real-time updates, and we are excited that enterprises can now build scalable AI applications on real-time streaming data." Rockset's high-speed database is designed to support various real-time applications, not just AI. For example, it is widely used in IoT devices and sensors that need to record and analyze incoming data to ensure efficient and uninterrupted machine operation. If data collected by a sensor indicates a potential machine failure, Rockset can record and provide immediate troubleshooting feedback to prevent the failure. It also has applications in areas such as network security and e-commerce, where real-time data is highly valuable. Rockset's high-speed database can receive and analyze data within seconds. Additionally, the database is claimed to be more reliable, thanks to its innovative compute-compute separation feature, which provides two independent compute pools for ingestion and analysis records, unlike other databases that use only one compute resource pool. According to Rockset, this is particularly useful for demanding AI applications. The company's claims are not unfounded, as it recently raised $44 million in funding. Investors have expressed particular interest in its practicality for AI applications. Rockset also disclosed that its revenue and customer base have more than doubled in the past two years. Some of Rockset's customers have already deployed AI applications at scale using its advanced vector search capabilities, such as low-cost airline JetBlue Airways. The Senior Manager of Data Science and Analytics at JetBlue Airways stated in a recent case study that the speed of iteration and AI is the most critical factor for a database. "We have seen the tremendous power of real-time analytics and AI in enhancing and automating JetBlue Airways' real-time decision-making, as cobbling together 3-4 database solutions slows down application development," he said. "Through Rockset, we found a database that can keep up with JetBlue Airways' rapid innovation pace." Venkataramani appeared on theCUBE's mobile live studio, participating in the coverage of AWS re:Invent 2022 Global Startup Program, where he discussed in detail Rockset's real-time analytics capabilities and the hundreds of applications it can support.