A practical walkthrough of using AI (GPT) to iteratively build a Blazor app feature. The author demonstrates how to guide an LLM through architectural decisions, task slicing, and incremental implementation. Key techniques include breaking work into vertical slices, creating a task-runner 'skill' with a progress log so new LLM chat sessions can pick up where the last left off, and knowing when to do things manually (like project setup) rather than burning tokens on AI loops. The post also covers using AI for component refactoring after initial implementation.
Table of contents
Smaller SlicesYou Can Still Do Things ManuallyKeep Tabs on Your ProgressDid AI Get Everything Right? Not ExactlyIn Summary227 Impressions