A developer has returned to writing code by hand after two years of using vibecoding tools. The shift, documented in a YouTube video, suggests a growing reassessment of AI-assisted programming workflows. This move highlights a potential fatigue with purely AI-driven development.

The trend of vibecoding—using AI to generate code from natural language—has accelerated with tools like GitHub Copilot and ChatGPT. Many developers initially embraced it for rapid prototyping. However, this return to manual coding points to ongoing debates about code quality, understanding, and the loss of foundational skills in the industry.

What comes next will likely involve a more hybrid approach. Developers may use AI for boilerplate but write critical logic themselves. This personal experiment adds a practical voice to the broader conversation about balancing AI assistance with developer autonomy and deep technical mastery.