I spent 3 years being afraid of AI math. Then I built NPCs that actually learn.

Hey PH community 👋

Long-time lurker, first-time poster. I've been watching launches here for years, always impressed by what people build. Finally have something of my own to share.

The backstory:

I wanted to add AI to games since forever. Every tutorial? "Let's start with linear algebra and gradient descent." I'm not a math person. I quit three times over two years.

Then I discovered something: modern game engines (Unity, Unreal) have AI tools that hide the complexity. You can train an NPC to learn from player behavior without writing a single line of calculus. Nobody told me that. Every course assumed I wanted to become an AI researcher.

So I built what I wished existed.

AI Powered Game Dev For Beginners

What it actually is:

• No PhD required. Zero "let's derive backpropagation" energy.

• You build NPCs that learn, train reinforcement agents, make worlds that react

• 50+ hours of "watch me fail so you don't have to" content

• Unity ML-Agents, Unreal behavior trees, real projects not toy examples

Why I'm posting here (and not just launching):

PH's forum guidelines say to post genuine questions or helpful insights . So here's my honest question for this community:

For those who've built game AI before; what was the first moment you felt it "click"? For me, it was watching an agent finally navigate toward the player after 2 hours of failing. Felt like teaching a Roomba to have emotions.

Kickstarter link if curious: https://www.kickstarter.com/projects/eduonix/ai-powered-game-dev-for-beginners

Happy to answer questions, argue about whether behavior trees beat state machines, or just lurk more productively than before. 🎮