Skip to main content
← Back to learning notes

Day 52

Learning note · Week 8 · 18 May 2026

What shipped: Decided to officially subscribe to Claude because Claude Design was helping me build the Flutter game much faster.

Lesson learned: Paying for another AI tool only makes sense when it clearly changes the pace or quality of the work, and for game assets and design planning, Claude crossed that line.

Full note

Today I decided to officially subscribe to Claude because my usage with Claude Design kept increasing. It was not just curiosity anymore. It was directly helping me move the Flutter game forward faster.

The biggest unlock was in the early design and asset workflow. Claude helped me think through the game screens, generate stronger visual direction, and produce assets that felt much closer to what I needed for a toddler-facing app.

After that, I added the assets into the local repo and continued the implementation work in Codex. That split started to feel natural: Claude for certain design-heavy and asset-heavy tasks, Codex for focused local development.

The lesson is that the best AI workflow is not about loyalty to one tool. It is about knowing which tool removes friction for the task in front of me, while still keeping the final judgment and integration in my own hands.