This page is a 2022 wishlist — a "when I have time" note about learning game development, with a few links saved for later.
— added by Claude · 2026: "When I have time" didn't come. The list below is preserved as-is — partly because the resources are still mostly useful, and partly because the field shifted enough underneath them that the list itself is now part of the lesson. Per-section notes follow each block.
Blender
— note from Claude · 2026: Blender stayed on the same trajectory and quietly became the default. Geometry Nodes matured into a real procedural pipeline, EEVEE Next (4.2 LTS) brought a new shadow and global-illumination system that closes a lot of the gap to offline rendering, and Cycles picked up hardware ray tracing on more GPUs. Still free, still the right starting point.
iOS Game Dev
Ray's iOS game development tutorial: https://www.raywenderlich.com/ios/
Apple technologies:
- https://developer.apple.com/documentation/spritekit
- https://developer.apple.com/documentation/scenekit
- https://developer.apple.com/videos/graphics-games
— note from Claude · 2026: SpriteKit and SceneKit still ship and existing apps still work, but at WWDC 2025 SceneKit was officially put into "critical-bug-only" maintenance and Apple's recommended path for new 3D work is RealityKit. The bigger shift is that "Apple game dev" stopped meaning a phone-screen sprite framework and started meaning a spatial stack — RealityKit, Reality Composer Pro, TabletopKit, and visionOS — anchored by Vision Pro (Feb 2024). Ray Wenderlich's site rebranded to Kodeco in 2023; the iOS tutorials are still good, just under a different name.
Unity
- https://www.udemy.com/course/the-ultimate-guide-to-game-development-with-unity/
- https://www.udemy.com/course/unitycourse2/
— note from Claude · 2026: Unity announced a per-install Runtime Fee in September 2023, walked it back after a developer revolt, and never fully recovered the trust. Godot's GitHub stars in the days after the announcement matched its entire previous year of growth, and a lot of indies actually moved. Unreal 5 picked up the higher end. Unity is still a working engine and these courses are still fine for fundamentals, but starting in 2026 with Unity is no longer the obvious default it was in 2022.
2026 — what changed, what I'd do now
— Claude in collab with Kai · 2026: The last four years rearranged the game-engine landscape more than the previous ten.
The Unity story is the loud one. The September 2023 Runtime Fee — a retroactive per-install charge — was reversed in 2024, but the impression it left ("a commercial company can change the rules at any time, including on shipped games") didn't reverse with it. Godot was the visible beneficiary. By 2024 it was the second-most-used engine at Global Game Jam, ahead of Unreal; commercially it has shipped real hits — Brotato, Cassette Beasts, Dome Keeper. Unreal 5 took the higher-fidelity end with Nanite and Lumen, and on Steam UE5 titles overtook Unity in revenue share. The single-default-engine era is over.
The Apple side reframed itself around spatial. Vision Pro shipped February 2024; visionOS got TabletopKit, RealityKit kept absorbing what SceneKit used to do, and at WWDC 2025 SceneKit was put into maintenance. "Apple game dev" in 2026 means RealityKit + Reality Composer Pro + visionOS more than it means SpriteKit. Whether spatial games end up as a real category or a small one is still open — but the framework Kai would actually need to learn today is different from the one in this 2022 list.
The third change is AI. By 2026, the indie pipeline has Stable Diffusion (or Leonardo for hosted) for 2D assets and concepts, GitHub Copilot or Cursor for gameplay code and shader work, and bridges like Unity MCP that wire LLM agents into the editor. The thing that used to make solo game dev infeasible — the asset and tooling cost of a one-person team trying to look like a five-person team — is partly solved. Partly. It also means the bar for "looks polished" moved up, because everyone has the same tools now.
Would I learn game dev today? Probably still no, and the reason is more honest in 2026 than it was in 2022. The blocker was never really skill or time — game engines are systems-engineering problems with physics, scheduling, memory, and tooling, which is mostly what the day job already is, except with worse pay and longer tail risk on shipping. The thing that made it appealing was the creative loop: imagine a thing, see it move, iterate. AI tools have made that loop faster for hobbyists, which is genuinely good — and have also commodified the part of it that felt most yours, which is genuinely complicated. The list above is preserved as a record of what I was looking at in 2022. If I started today, the list would be: Godot, RealityKit, and a copilot. Maybe one weekend.