AI made it much easier for me to build games. It didn't make figuring out what's fun any easier.
retained text
I've been developing a small indie game where players build their own aircraft by connecting different components. I have over 20 years of experience in game development, including programming, game design, and production. But this is my first time relying almost entirely on AI to write the code for one of my own game projects. And the experience has been pretty eye-opening. **AI has dramatically reduced the cost of implementing ideas. But it hasn't reduced the difficulty of figuring out which ideas are actually fun.** Over the past few weeks, I've experimented with several different directions. **1. I tried making it more like Snake.** The idea was to keep collecting components and growing your aircraft. But once I added guns, the game stopped feeling like Snake. Shooting changed the whole experience. What seemed like a small addition completely changed the core gameplay. **2. I tried a Vampire Survivors-style approach.** Lots of enemies, automatic shooting, numerical upgrades, and constant rewards. It worked mechanically, but I didn't enjoy playing it as much as I expected. **3. I experimented with a Brotato-style approach.** Short combat encounters, upgrades, and build progression. Again, I could make the systems work, but something was missing. It felt too focused on stats and upgrades rather than the experience I originally wanted. **4. I experimented with free exploration and scrolling shooter gameplay.** Free exploration gave players more freedom, but controlling the aircraft was difficult to get right. Scrolling combat felt much better to control and made bullet patterns easier to design. But it also changed the feeling of the entire game. I also tried giving enemies the same movement speed, weapons, and component rules as the player. That turned out to be a disaster. Enemies chased me too closely, their bullets were too fast to dodge, and fighting multiple enemies became overwhelming. It reminded me that symmetrical rules don't necessarily create enjoyable PvE combat. **5. Then I replaced numerical upgrades with physical components.** Instead of choosing something like "+20% damage," enemies randomly drop actual weapons, engines, and other parts. Players can attach these parts to their aircraft and immediately see the results. And suddenly, I started enjoying the game much more. I realized I didn't just want to make numbers go up. I wanted to build something, experiment with different combinations, and see my creation become stronger. For example, a few forward-firing guns mounted at different angles can create a spread-shot pattern without needing a separate shotgun component. I'm now exploring the idea of a rotating component that can rotate an entire attached structure. Connect a laser to it, and you get a sweeping laser. Connect several guns, and you get a rotating bullet pattern. The long-term possibilities are exciting, but I'm trying not to get carried away with scope. **What surprised me most was how much of the game I had to rethink after actually playing it.** In traditional development, implementing and then discarding a gameplay system could be expensive and frustrating. With AI, I can prototype ideas much faster, play them, discover what doesn't work, and change direction. But AI can't play the game for me and decide what feels right. I still have to make those judgments, and that's often the hardest part. I used to think game development was mostly about turning a design into a working game. Now I'm increasingly convinced that the real process is: **Build → Play → Discover problems → Redesign → Build again.** And a lot of the actual game design happens during those iterations. I'm still far from having a finished game, and I'm not even completely sure what the final gameplay loop will look like. But I think I'm finally starting to find something I genuinely enjoy playing. I'm curious how other developers approach this: * How much of your original game design survives the prototyping process? * Have you ever implemented a mechanic that sounded great on paper but completely changed the game in an unexpected way? * If you're using AI for development, has faster prototyping changed how you approach game design? I'd love to hear other developers' experiences. Solo development can be pretty lonely, and hearing how others deal with the same problems really helps.
document fetched 2026-10-13 00:27 · canonical https://reddit.com/r/gamedev/comments/1x3rz0k/ai_made_it_much_easier_for_me_to_build_games_it