In late 2022, while working on scaling Readwise’s infrastructure for their Readwise Reader app, Simon Hørup Eskildsen realized the high costs of implementing article recommendations and semantic search using traditional relational databases. This led him to develop turbopuffer, a search engine built on object storage and smart caching that significantly reduces costs while maintaining high performance. The innovative approach offers cost efficiency without compromising speed, making it appealing to companies like Cursor, who saw a 10x reduction in costs and improved latency. With a focus on scalability and reliability, turbopuffer aims to revolutionize search engines for the future.
https://turbopuffer.com/blog/turbopuffer