Haven't tried it, but applications like these are the best uses of LLMs. Translation from an informal human language to a formal language that the human can review, and can be manipulated formally further.
I believe the attempts to have LLMs "write the code" or "reason" are wrong. What we need is to move to more formal languages to replace the ambiguity of natural language.
one thing i personally admire/appreciate about simonw's work is how beautifully his tools, services and research "stacks" to unlock higher and higher levels of abstraction. so many of the little one-off scraper/notifier/transformer repos on his github showcase the amazing composability of the tooling he built previously.
DuckDB is a more convenient and performant tool for analyzing large CSV/JSON files. Shameless plug: if you’re interested in combining DuckDB with Claude using the recently announced Model Context Protocol (MCP), check out the MyDuck Server project (https://github.com/apecloud/myduckserver) that I’m working on. You can follow this guide https://github.com/apecloud/myduckserver/blob/main/docs/tuto... to get started. The guide includes a conversation history showcasing how the free-tier Claude Haiku model successfully analyzed the Our World in Data Energy dataset.
MyDuck Server adds a Postgres frontend for DuckDB, allowing seamless interaction with DuckDB via the official Postgres MCP server.
I once told my son when he was 4 that if his tablet crashes, or a video stops working, he needs to ask it nicely. It usually worked - the issue was lack of patience. Still, it made me smile when I saw him every now and then putting his most polite voice and saying, "please, can you play my video?".
But soon... Won't be able to do anything without interacting with an LLM!
I believe the attempts to have LLMs "write the code" or "reason" are wrong. What we need is to move to more formal languages to replace the ambiguity of natural language.