LLMs and Xfwl4
14 hours ago
- LLMs are a tool with positive and negative aspects, including concerns over training data and copyright infringement.
- Environmental and social impacts of LLMs include high resource consumption and secret deals that burden local ratepayers.
- GPU and RAM supply issues are exacerbated by AI demand, making personal hardware upgrades expensive.
- The author uses Claude extensively for research and planning in developing xfwl4, such as analyzing xfwm4 source code.
- Implementation is mostly done manually, with Claude used for validation, code review, and occasional boilerplate code generation.
- Claude is highly effective for bug fixing, often finding issues faster than the author.
- Unit tests in xfwl4 were written by Claude due to the author's distaste for testing.
- The author avoids vibe-coding and focuses on understanding the code for future manual maintenance.
- LLM usage has saved months of time, particularly in problem-solving and debugging, despite ethical discomfort.
- The author provides transparency about LLM use in the xfwl4 development process.