@jrf

@marqui ~tldr: origin story of my AI experience~ when I started practicing law in 2013, my first assignment was reviewing millions of documents for a fortune 50 client w a team 57-hour weekly quotas sit. click. tag. repeat. it was factory work for the mind ofc i tried to hack the system better search strings smarter filters find the needles without drowning in the hay it was the moment i began to understand how databases work (fields, queries, logical operators) and how much leverage you get from structuring information the right way fast-forward to 2016 new case small team 1.5 million documents need to onboard and train reviewers we needed leverage so we built it before “legal AI” was a thing, we worked with a vendor to create our own model i manually reviewed ~30,000 documents and built a schema: what each document meant and how it fit into the larger story then we used that schema to infer relevance across the full set highly relevant, maybe relevant, not relevant key docs surfaced instantly that was my first real proof that structure beats brute force that knowledge graphs matter that the right tags, fields, and relationships can transform chaos into clarity only years later did I realize how much that shaped the way I think about AI (when the OpenAI API came out and i played with it in google sheets) it’s still guiding how we’re building @marqui, a tool that gives anyone that kind of power, but through a simple, intuitive, hopefully fun experience a personal website that actually does things for you a knowledge engine hiding under a playful UI if you’re curious, follow @marqui we’ll be onboarding users soon in ~early jan~ excited to see how it can work for you

12 replies

10 recasts

44 reactions