In this April 1, 2026 live dev catch-up, I walk through what’s been happening in my AI game dev journey—especially how I’m escaping the classic solo dev trap (perfectionism, burnout) while juggling a full-time studio job and my own prototype. I’m a 3D artist first, but lately I’ve been pushing hard into Unity, tooling, and game design because I’m building a casual puzzle game with “flower bosses.” The big focus is my level preset/creator system: I can drag-and-drop AI-generated pixel-art heads (owl, sheep, pig, etc.), extract colors, tune flower counts (I’m usually capping around ~600), scale, connect bosses, hide colors with mystery boxes, and rapidly iterate levels without killing FPS. I also show my practical AI asset pipeline: generating UI and sprites in Flow/Whisk (Nano Banana 2), then doing quick cleanup in Photoshop (selection tool + Ctrl+J layers, hue/saturation, export). On the broader AI side, I talk about a fresh open-source workflow that auto-separates 2D character art into animation-ready parts (eyes, ears, etc.)—the kind of thing that used to take days of pen-tool grinding in Spine prep. Finally, I share the studio workflow I used on “Fit the Object,” where I generate scenes, split them into paintable parts, and keep improving the process inside Flow/Gemini + Photoshop generative fill.