Deserialize messy LLM tool arguments into a Rust struct with Laminate
LLM tool-call arguments arrive as messy JSON: numbers as strings, missing fields, invented enum values. Laminate shapes them into a typed Rust struct, and shows every change.
Tagged posts
Laminate: one extract() for every provider's tool calls
OpenAI, Anthropic, and Ollama return tool-call arguments in three shapes. Laminate's provider adapters normalize them in Rust, so one extract() call reads every provider.
A memory for AI sessions that refuses to be clever
I built my AI assistant a memory that stores, inserts, and retrieves, and deliberately never analyzes. That restraint is why it survives daily use.
I set out to build an AI document app. I built a PDF parser instead.
I spent months on eighty thousand words of architecture and a 2,700-line Rust PDF parser. The AI document app they were for never opened a window.
Wake, Sleep, Dream: giving an AI a memory
What would it take to give an AI a memory that lasts? A three-phase idea borrowed from the brain, with a rough Postgres and pgvector prototype underneath.