I Asked ChatGPT with Agent mode to Rebuild an Old RTS Game with Me
📜 Video Description:
In this video, I take a deep dive into the classic RTS game Baldies (1995) and show how I worked with an AI agent (GPT-5, agent mode) to recreate the game step by step in Unity.
We start with research into the game’s history, gameplay mechanics, and development background — and then move into a strict, AI-driven sprint plan where every system is implemented sequentially:
✅ Project setup & GitHub integration
✅ GameManager, cursor input, grid + A* pathfinding
✅ Units with roles (Workers, Builders, Soldiers, Scientists)
✅ Building placement, upgrades, breeding & resources
✅ Research tree (technologies), barracks, enemies & combat
✅ Enemy AI waves, level progression, and results screen
✅ Audio system, UI polish (tooltips, highlights, save/load)
✅ Performance optimisations (object pooling, active lists)
✅ CI/CD pipeline with GitHub Actions for automated builds
⚡ What you’ll see:
By the end, we’ve got a fully playable RTS framework — not yet the complete game, but a systems-complete prototype. It’s ready for content authoring (levels, tilemaps, art, audio, balancing) in Unity Hub.
👷 What’s next:
I’ll show how to design the first levels, hook up sprites and sounds, and push a release build. If you’re curious how to use AI agents as coding partners for game dev, this is the full journey from idea → research → implementation → build.
🔗 Resources & Links:
• Baldies research & history → [Wikipedia/MobyGames]
• Unity LTS & Tilemap docs → [Unity Documentation]
• GitHub Actions Unity Builder → [game-ci/unity-builder]
👉 If you enjoyed this breakdown, like & subscribe for more AI-assisted game development and C# / Unity tutorials!