A Pragmatic Approach to LLMs
6 hours ago
- The author chooses not to use LLMs for coding, not even tools like Claude Code or Cursor.
- The author values the craft of programming, arguing that friction in code-writing encourages thoughtful design.
- Using LLMs is seen as surrendering autonomy, creating dependency on corporate tools that may increase costs.
- LLM-generated code is criticized as 'signal-shaped noise' that degrades open-source projects with unmaintainable contributions.
- LLMs are considered the wrong approach for human-computer interaction due to lack of determinism compared to compilers.
- Legal risks exist with LLM-generated code due to unclear copyright and potential derivative work issues.
- Environmental concerns are noted, as excessive LLM usage and training contribute to energy consumption.
- Limited legitimate use-cases for LLMs include navigating poorly documented codebases and vulnerability research, but only when run locally and with fully open-source models.
- The author supports policies like Zig's ban on LLM contributions to protect project integrity.