Building the App Wars Scoring App with ChatGPT :/

A Journey from Idea to Execution

At Blue Otter, we’re always exploring ways to integrate new technologies to make our projects more efficient. Recently, we used ChatGPT to build a scoring app for our card game, App Wars (theappwars.com). The experience was both exciting and enlightening, showcasing how AI can be a powerful tool in software development. Here’s how it went from a simple idea to a functioning app in just a few hours.

The Prompt That Started It All

I began with a simple idea: create an app that allows players to keep score while playing App Wars. The game revolves around players collecting users, with milestones at 1,000,000, 3,000,000, and 5,000,000 users marking sprints and potential game endings. With no specific plans for the app’s design or functionality, I started with this prompt:

Help me create a ReactJS app for players to keep score for App Wars, a card game where players collect users to win the game. The game is divided up into sprints with each sprint ending with 1,000,000 acquired users or more. Once a player reaches 3,000,000, the app will ask the player if they want to end the game or continue to the extended game ending at 5,000,000.

Within five minutes, ChatGPT generated a basic ReactJS app. It wasn’t perfect, but it gave us a functional starting point that I could tweak and refine.

Iterating with AI

I spent the next 2-3 hours making adjustments, fixing bugs, and enhancing the app. I’m not going to lie. There were times I cussed it out because it was so dumb. I took my frustrations out on the AI but it always responded with, “I’m sorry for the oversight, here’s the solution.” After my blood cooled, I thought to myself, I ust yelled at an AI. I hope it doesn’t keep it in the collective and when SkyNet wakes, it doesn’t take revenge on me. Ahh, enough with the side thoughts…

While ChatGPT was incredibly helpful, there were moments when it made solutions more complex than necessary. For example:

• Over-engineering Logic: At times, the AI suggested intricate solutions for simple problems, requiring us to clear the chat and refocus it on the task at hand.
• Repetition Loops: Occasionally, ChatGPT got stuck in an endless loop, repeatedly suggesting the same solution. This taught us the importance of asking specific, clear questions to guide the AI effectively.

Despite these hiccups, the process was surprisingly smooth. The key was diligence: knowing when to reset, refine the prompt, and steer the AI in the right direction.

No Code Written by Hand

One of the most striking aspects of this experience was that I didn’t have to write a single line of code. Every piece of functionality—from score tracking to prompts for ending the game—was generated by ChatGPT. All I did was copy, paste, and test.

What We Learned About AI in Development

The experience reaffirmed a few key points about using AI in software development:

1. AI as a Companion, Not a Replacement: While AI can kickstart a project and help with repetitive tasks, it still requires human oversight to ensure the solution aligns with the project’s needs. Developers remain essential for guiding AI and making critical decisions.
2. Specificity Matters: The more detailed and precise your questions, the better the results. Broad prompts yield generic solutions, while specific prompts lead to tailored responses.
3. Efficiency Boost: AI speeds up development by handling boilerplate code and offering quick solutions to common problems, freeing up developers to focus on higher-level tasks.

We will continue to improve the app using the help of AI, specifically Cursor.

AI at Blue Otter

At Blue Otter, AI isn’t just for coding. Every team member—whether in QA, development, project management, or even leadership—uses AI to enhance productivity. It’s a tool that helps us think through solutions, manage tasks, and deliver better results for our clients.