Well, it was just supposed to be another test of free LLMs. I opened Claude AI and described how a simple, SimCity-like game should work. I gave him some exact instructions and let him code an Android app. After providing him with instructions, checks and feedback, I achieved a fully functioning proof of concept in a few days.
This is what it looks like now. It's really nice! This state was supposed to be the final one. However, I really enjoy the game and like how it looks, so I'm in a dilemma. Should I continue working on it alone, or just give up? It is actually on the limit of what the LLM can do, as it has become too complex and the code has grown too much. Now, even small adjustments kill the whole context, and it is almost impossible to make any progress.
So, what do you think? Should I finish it on my own and release it?
Jan "Jenkings" Škoda - blog
Simplifying IT, Amplifying Knowledge