Dropbox

Universal Search is maybe the holy grail of enterprise info management, and Dropbox has made a major entry into this field. I was an alpha user of universal search app Command-E (founded by Ben Standefer and Tom Uebel), which was acquired by Dropbox and formed a key part of the strategy behind Dash. Before Product Hunt, I worked on new products with Drew at Dropbox and was struck by his ambition and how hands-on he was with new product teams (#foundermode before it was in the popular consciousness!). I'm really excited to see what happens with Dash.

What kind of search technology is being used here? Is this more along the lines of Elasticsearch/BM25, or are we dealing with vector search and cosine similarity? Are we searching within documents, and if so, how is document chunking handled? This seems like a complex task to make universally available, so I’m a bit skeptical. I’d be happy to discuss the technology behind it—I'm working with a similar system and have experienced the limitations of different methods firsthand. I'd also be interested in discussing enterprise implementation if there's something worth exploring here