Stephen Brennan shares his perspective on LLM-based coding agents like Claude Code. While these tools are becoming increasingly capable, allowing both programmers and non-technical users to create features or even entire projects, Brennan doesn't find them enjoyable to use. He argues that his enjoyment comes from the process of software creation, not just the end result.

Brennan values the act of coding and understanding software, which he sees as a worthwhile pursuit. He enjoys representing problems in code and building mental models of systems. For him, the software product's value isn't the only, or even primary, value. In contrast, many developers may value the end product more, seeing software as a means to an end, leading to a different experience with these AI-powered tools.

He suggests that those who focus on the result are likely to embrace these tools, allowing them to iterate and manage AI systems. For Brennan, the learning and understanding gained through the coding process is essential. He also looks at the potential impact on his career, recognizing that the industry is changing with AI advancements.

Brennan concludes that while AI might automate certain tasks, particularly those he doesn't value, he will continue to focus on the aspects of coding that bring him joy and provide learning opportunities. This approach allows him to use AI tools where they are beneficial, while still prioritizing the parts of the process he enjoys.