DiskANN

Microsoft Azure Data | Senior Software Engineer (Vector DB) | US Redmond/Remote We build the engine (https://github.com/Microsoft/DiskANN) that brings vector search capabilities to Azure Databases and most other at-scale search/RAG systems at Microsoft. See aka.ms/AboutDiskANN Looking for Rust developers to join our team of researchers and engineers. Experience with performance engineering, OSS development, Database internals, or numerical methods a plus. https://apply.careers.microsoft.com/careers?query=200019455&...

Microsoft Azure Data | Senior Software Engineer (Vector DB) | US Redmond/Remote We build the engine (https://github.com/Microsoft/DiskANN) that brings vector search capabilities to Azure Databases and most other at-scale search/RAG systems at Microsoft. Looking for Rust developers to join our team of researchers and engineers. Experience with performance engineering, OSS developments, Database internals, or numerical methods a plus. https://apply.careers.microsoft.com/careers?query=200019455&...

Answering my own question (with the help of replies on Twitter: https://twitter.com/simonw/status/1798368111038779610 ) Best guess is that there's some level of semantic analysis going on beyond just OCR, which tips the balance (at least for me) to being an "AI" feature based on loose current usage of that term. The clue is that in this demo https://www.youtube.com/watch?v=aZbHd4suAnQ&t=1062s they show searching for "blue pantsuit with sequin lace" finds something that was described as "peacock" in the text. It looks like this is based on embedding search against image embeddings. And... one of the 3 SQLite databases created by Recall is called "SemanticImageStore" - which suggests to me there may be a CLIP-style image embeddings model being run on-device here. Those databases contain "diskann" columns which are likely a reference to Micr

DiskANN seems to be a recent & effective approach: https://github.com/microsoft/DiskANN

My team at Microsoft works on DiskANN, which was designed for this kind of use case. https://github.com/microsoft/DiskANN

Found code for it https://github.com/microsoft/DiskANN

Spot on. There is zero moat and the self-hosted alternatives are rapidly improving (if not better) than Pinecone. There are good open-source contributions coming from bigcorp beyond Meta too, e.g., DiskANN (https://github.com/microsoft/DiskANN).