The Optimal Amount of Slop Is Non-Zero

# Summary The author argues that code review rigor should be proportional to risk—sometimes skipping review entirely is appropriate, while other situations demand careful scrutiny. Though LLM-generated code often works as specified, it frequently contains quality issues invisible to external inspection, making human code review a valuable advantage developers shouldn't automatically outsource to other AI systems. The optimal approach requires matching verification effort to the actual consequences of potential failures.