A developer used Google's Antigravity AI to build a portfolio from scratch, treating it as a practical learning exercise. The project began after discovering a dev.to frontend challenge that aligned with his goal of creating a portfolio while exploring AI tools. He selected Antigravity partly for its name, then followed its structured workflow for planning and implementation.

The process revealed key insights about AI-assisted development. Antigravity proposed plans for implementation and verification before coding, which the developer found helpful for understanding the AI's reasoning. However, he noted the AI sometimes suggested suboptimal choices, like downgrading software versions instead of updating them, and creating overly literal folder structures for Atomic Design concepts.

Ultimately, the developer successfully launched his portfolio using Next.js and Google Cloud Run. The experience taught him about component delegation and the Turbopack compiler. He found Antigravity's approach valuable, especially the clear planning stages, and recommends the challenge for others looking to sharpen their skills through hands-on projects.