Show HN: Lathe – Use LLMs to learn a new domain, not skip past it
Lathe is a Go CLI tool that leverages LLMs to generate hands-on, source-backed technical tutorials for any domain, filling gaps where human-written guides are scarce. It provides an interactive local UI where users actively type code, fostering deeper understanding rather than passive consumption. This project directly addresses concerns about LLMs enabling superficial learning by turning them into powerful tools for active pedagogical engagement in niche technologies.
The Lowdown
Lathe is an innovative project designed to harness the power of Large Language Models (LLMs) not to automate coding, but to facilitate deeper, hands-on technical learning. It addresses the gap in educational resources for niche or rapidly evolving domains by dynamically generating comprehensive, interactive tutorials. The system emphasizes active engagement, mirroring traditional learning methods where users type out code themselves.
- Lathe is a Go CLI that integrates with LLMs (Claude Code, Cursor, Codex) to produce multi-part technical tutorials based on user prompts.
- Users interact with these tutorials through a local web UI, designed for an immersive learning experience with features like a scrolling table of contents, side-notes for deeper thought, and exercises.
- It includes "skills" allowing users to ask questions about the content, verify tutorial code for compilation/runtime, and extend tutorials with new parts.
- Tutorials document their sources, the specific LLM used, and the "voice" (e.g., plainspoken, companion) chosen for the prose.
- The project aims to fill the void where human-written tutorials don't yet exist, enabling learning in obscure or very new technical areas, such as building a 3D slicer or embedded Zig programming.
- Lathe acknowledges LLM limitations, stressing that user engagement (typing code, questioning output) is crucial for identifying and correcting "hallucinations," turning potential errors into learning opportunities.
- Verification is an opt-in feature where the LLM can run through tutorial steps in a scratch directory to confirm functionality.
- The author encourages using human-written tutorials first, positioning Lathe as a valuable tool for filling educational gaps rather than replacing existing resources.
Ultimately, Lathe represents a thoughtful application of LLM technology, shifting the paradigm from automation to augmentation of the human learning process. It champions active, self-directed education, particularly for domains underserved by conventional teaching materials, proving LLMs can be powerful aids in intellectual growth rather than shortcuts.