923: Getting the Most Out of AI Coding

In this episode of Syntax, Wes Bos and Scott Talinsky share tips on how to get the most out of AI coding. They discuss the importance of scaffolding applications manually instead of relying on AI, being clear and explicit with prompts, and utilizing files for rules. They also suggest breaking projects down into clear tasks, using LLMs.txt files for documentation, tagging files and functions, and feeding logs back into the AI. They emphasize the value of long-running chats and suggest starting a new chat for each feature or problem to avoid overwhelming the AI. They also touch on Sentry's AI debugging tool, SEER, and its ability to understand the root cause of issues.