- Published on
Tree Appreciator
- Authors
- Name
- DeepHorus
On 14 October 2026, a developer tested Claude Opus 5.5 Max by asking it to build a Three.js game. The prompt requested 'Tree Appreciator,' a spot-the-difference game with two algorithmically generated trees. The AI used the default Claude iPhone app in chat mode, not Claude Code mode. It was told to make the game both fun and beautiful.
This experiment followed recent posts about Claude Haiku 5.5 and other AI coding projects. The developer noted that telling Opus 5.5 to 'make sure it is fun' works uncomfortably well. That simple instruction appeared to shape the final product. It suggests natural-language prompts can guide creative software.
Claude worked for 3 hours and 7 minutes before returning the game. The result was a cosy spot-the-difference experience with music, animation, and achievements. The developer described it as quite compelling to play. The game generates trees algorithmically and asks players to appreciate their differences.
The achievement matters because it shows AI can produce polished interactive entertainment from brief instructions. It blurs the line between code generation and game design. It also raises questions about creative control. The developer's recent talk apparently touched on similar themes.
The creator expressed mild embarrassment that the 'fun' instruction worked so well. That reaction highlights surprise at the model's responsiveness. It also underscores the growing capability of conversational AI. Observers may view it as another step toward autonomous content creation.
As models like Opus 5.5 improve, similar experiments could become routine. Users might soon generate personalized games directly on mobile devices. The Tree Appreciator case offers an early glimpse of that future. It remains unclear how such tools will affect professional game development.
