Manually retyping LLM-generated code prevents cognitive debt

Prevent cognitive debt by manually retyping LLM-generated code

Despite earlier skepticism, the author uses coding assistants on personal projects but limits them to chat suggestions, manually typing every edit. This slows him down to 2x speed instead of 10x, but builds a deep mental model and spatial map of the codebase, catching hallucinations and adapting code to his taste. He argues this preserves comprehension and fights industry-wide cognitive debt.