AI-First Game Development. Part 6. Launch Day
Last night at 2 am I finally got the message I had been waiting for: App Approved for Submission.
This was the fourth attempt after rejections relating to not working properly on the iPad. (It turns out that even if you explicitly say your App is for the iPhone, they still want it to work on the iPad?) Nevertheless, Bubble Pop Hero — Animals, now works on the iPad. And the larger format works well for the game.
Be sure to download the game and get your dopamine fix from frantic bubble popping! It’s available on the Apple Store here, and coming soon on the Google Play Store.
As a wrap-up to this series on building a game following an AI-First method, I will leave you with my final thoughts and opinions in no particular order of importance.
If you’ve been following this series, I hope you enjoyed the journey. If you haven’t, you can go back to begin and read about the many different learnings I had along the way from a design, creative, learning, and engineering perspective.
Part 0: AI First Game Development (Intro)
Part 1: AI-First Game Development. Training the GPT
Part 2: AI-First Game Development. My AI Teacher
Part 3: AI-First Game Development. The Chasm Between Design and AI-Generated Art
Part 4: AI-First Game Development. The Ideas Machine
Part 5: AI-First Game Development. The Engineer Lives
AI-First Reflections
The AI-Saftey Net. Game development is a challenging and rewarding journey. Coming up with ideas and learning and creative problem-solving your way through to a finished product is fun, frustrating and satisfying in equal measure. Something that’s always put me off doing this is the time required and lack of knowledge to build a proper game. Knowing that AI was there to help me learn, do things faster, and get unstuck motivated me to crack on and get it done.
AI-First Product Management. I didn’t dedicate a section on how Product Management is affected by AI. My main takeaway is that Prompt Engineering is crucial to effective product management, as this is the bridge between the Idea and the Output. We all need to become better product managers, no matter what our discipline, so we can instruct AI to produce output that meets our needs.
If your product isn’t well documented and supported, AI can’t help. The only one challenge I experienced which AI couldn’t help with was trying to get an advertising library called IronSource working on the App. The pain I experienced trying to get this to work in my solution was significant. The documentation and the support were terrible, and when I asked AI to help which included searching the Internet, reading installation documentation, discussion forums, and more, it got lost in the complete mess of information about how to resolve the labyrinth of issues this product had.
Sometimes AI-First can take longer, but that’s Okay. There were many occasions where I thought, ‘I know how to do this’, but forced myself to be AI-First to see if there would be a better way. And often this ended up taking longer than if I had done it the old way. But, I know I won’t learn if I remain stuck in my ways, and paying experimentation tax is worth it over the longer haul.
AI needed me. A finger in the air guess of how much work AI did vs how much work I did was surprisingly small. A rough guess, I would say AI contributed 20% and I (and my design partner, Daz) contributed 80% to the project. I was expecting it to be the other way around at the start. This was probably due to the fact I am still learning how to use AI and the technology is still at an early stage, and will drastically change over time, but it felt good to know I was still useful!
I’m still an AI noob. To the point above, my overriding feeling is that there is so much more to explore in terms of how to use AI to help you achieve your goals, and I’m just getting started.
To wrap this whole experiment up:
Doing as opposed to talking about AI has been an invaluable learning experience. I recommend everyone, young and old, to get stuck in using AI and figuring out how you can make it work for you. Taking on a project as a testing ground is perfect to test the many different ways AI can help you.
I also learned that the value of our skill and experience is still the driving force behind succeeding in a mission. Talented, skilled humans are still indispensable and AI can’t compete, but it can compensate.
But AI is evolving fast, and it’s imperative we start to experiment with this technology to stay ahead. There is much more to come and much more to do.
The AI-Bubble hasn’t burst.
