Vibe Coding: 12 Months with Claude Code, Case Study
summary
A year-long practical experience using Claude Code to generate code from natural-language prompts is examined, focusing on what works and what doesn’t for tasks such as legacy maintenance, large-scale refactoring, new feature development, validation and testing. The talk uses the "8 Levels of Agentic Engineering" framework (Bassim Eledath)—from simple autocomplete to autonomous agent teams—plus context engineering and automated feedback loops, and discusses human challenges like team adoption, evolving code review practices, and trust in AI-generated code.
AI-written summary · about
