Back to Journal

AI Game Dev Adventures

Building games with artificial intelligence

Day 72025-08-19
roadways-clone

Mini Motorways Take 2 - The Power of Mini Experiments

Discovering that breaking down complex games into tiny, focused experiments creates much better results than comprehensive approaches - cars driving in circles becomes surprisingly engaging.

Screenshots from Day 7
Screenshot 1 from Mini Motorways Take 2 - The Power of Mini ExperimentsScreenshot 2 from Mini Motorways Take 2 - The Power of Mini ExperimentsScreenshot 3 from Mini Motorways Take 2 - The Power of Mini Experiments

Mini Motorways Take 2 - Breaking Down Problems Smaller

Challenge:

What if I break everything down into mini experiments instead of trying to build a complete game?

Attempt 1: Circular Track

💬 PROMPT
"Create a circular track as a road that has cars moving around it"
✅ OUTCOME
Wow. It kinda works.

Simple house-to-house experiment - it works!Simple house-to-house traffic simulation

Attempt 2: Traffic Rules

💬 PROMPT
"Now make the cars follow traffic laws, i.e. driving on the correct side of the road"
✅ OUTCOME
Worked again!

Attempt 3: Point-to-Point Travel

💬 PROMPT
"Draw two houses on the map and have a street connecting them and then have cars driving back-and-forth between them"
✅ OUTCOME
This is where things fell apart - AI was making too many decisions on its own.

Attempt 4: Traffic Lights

💬 PROMPT
"Traffic lights, roads going through them, and ensure that cars follow the rules of the road"
✅ OUTCOME
Mixed bag. Some good, some messed up.

Complex intersections with traffic lights and roundaboutsComplex intersection experiment - traffic lights, roundabouts, and chaos

Lessons learned:

  • Mini experiments show real promise
  • Just having something moving around the screen changes everything completely
  • Makes it start feeling like something fun you want to watch
  • Still hate the default assets though
  • Feels like doing art in high school: "I hate it, hate it, hate it, hate it again, I love it, this is cool!"

Next:

What if I started with really nice assets first, then try to make something fun?

Ingredients

  • AI Assistant: Claude Sonnet - ~$2 in API credits
  • IDE/Platform: Browser-based game development - Free
  • Assets: Default geometric shapes - Free
  • Approach: Mini-experiment methodology vs. comprehensive requirements
  • Key Discovery: Incremental requests > complex specifications
  • Time Investment: 2-3 hours of focused experimentation

Community & Discussion

Join the conversation through comments or share your thoughts across the web. Discuss this journal entry and share your experiences.

Comments & Discussion

Join the conversation! Comments are powered by GitHub Discussions. Share your thoughts about this journal entry.

💡 Tip: Comments are stored as GitHub Discussions, while web mentions capture shares and mentions from Twitter, Mastodon, blogs, and other platforms. Together they provide a complete picture of community engagement.