[Part 5] Creating a Web Game in 30 Minutes with AI - The End -

Today, I spent a considerable amount of time adding various features. However, I realized that

AI is not omnipotent.

-> As features are added and the code grows longer, the quality gradually decreases, and bugs arise that are not immediately resolved. This ultimately leads to debugging, requiring me to trace each line of code individually.

It's crucial to design the structure carefully from the beginning.

-> Gradually adding features to a simple logic eventually reaches its limits.

This started with simple combat -> auto-combat -> adding a rulebook -> adding pages to the rulebook -> setting conditions on pages to execute combat based on those conditions -> adding settings for normal attacks and skill usage to the rulebook.

Adding features haphazardly made it difficult to get the system working properly...

I also added settings to sell skills in the shop, set skills to appear probabilistically, and prevented resale of purchased skills within the game, etc. After debugging and tracing lines of code until now, I realized...

Ah, I intended to use AI to do this, but it has become completely counterproductive. I will use this project as a valuable lesson and

I will stop here.

Next time, I will need to build a more robust core logic early on before expanding.