The author also differentiates between generic web development and specialized algorithmic work. For the former, they argue that it doesn't matter if code is written by a human, a library, or an LLM, as the goal is to create a workable component. For the latter, a solid understanding of the subject is beneficial. They also note that high-specificity queries are more helpful than broad ones, and that thorough comments improve the advice given by LLMs. They prefer ChatGPT over Copilot for its superior performance.
Key takeaways:
- Language Learning Models (LLMs) like ChatGPT can be a valuable resource for learning and problem-solving, even if they sometimes make mistakes.
- Whether or not to rely on LLMs depends on your role and goals. If you're a web developer dealing with common patterns, LLMs can save time and effort. If you're aiming to specialize in complex algorithms, a deeper understanding is necessary.
- Web development is often about reusing code and finding efficient solutions, and it doesn't matter whether you wrote the code yourself, used a library, or got help from a coworker or LLM.
- High-specificity queries tend to yield better results from LLMs, and providing thorough comments can improve the advice you receive.