THE 100,000 DOCUMENT GHOST
We vectorized 105,904 legal documents and the model still couldn't find the law. Here is what happens when RAG meets the hierarchy of authority.
We vectorized 105,904 legal documents and the model still couldn't find the law. Here is what happens when RAG meets the hierarchy of authority.
How I built a worktree-based system to run multiple Claude instances safely, eliminating pnpm lockfile conflicts that were blocking builds 4x daily
Vector databases are currently fragile toys masquerading as enterprise infrastructure. A post-mortem of the binary format tragedy that cost us 14 hours of compute.