CODE SMELLS FOR AI AGENTS: Q&A WITH ENO REYES OF FACTORY

Summary from the TLDR Dev newsletter

Building good coding agents isn't just about the model, but also about the infrastructure layer that manages context, tools, and hundreds of validation signals like linters and tests. Stanford research found that AI agent adoption doesn't predict productivity, but code quality does. High-quality codebases accelerate with AI, while low-quality ones actually decelerate.