Working on laconic, a "Ralph loop for Research" API costs for LLM are getting ridiculous. By managing a 4K context window, I can get tiny local models to do ad

Working on laconic, a "Ralph loop for Research" API costs for LLM are getting ridiculous. By managing a 4K context window, I can get tiny local models to do advanced research tasks that require dozens of searches and building a knowledge graph for $0. https://github.com/smhanov/laconic https://github.com/smhanov/laconic