# I Created a Rate Limiter Challenge Where AI Gets Confused 🤯

We all think rate limiting is simple.

Just count requests and block when limit exceeds… right?

That’s what I thought too.

So I created a small challenge on VibeCode Arena to test this idea.

And honestly, it’s not that simple.

🚨 The Problem

Here’s the basic logic:

• Count requests
• Check time window
• Allow or block

But when you look deeper, things start breaking:

• Time handling issues
• Reset logic problems
• Not scalable for multiple users
• No support for concurrent requests

This is where most AI-generated solutions struggle.

🧠 What Makes This Interesting

When different AI models try this challenge:

• Some give basic working code
• Some miss real-world edge cases
• Some ignore scalability completely

Very few actually think like a real system.

🔥 Try the Challenge

I created this challenge to test how well AI (and developers) handle real-world backend problems.

👉 Try it here:
https://vibecodearena.ai/duel/57b5c7df-b892-485b-b9c1-c2c684b69328

Curious to see:

Can you fix the bugs?

Can you make it production-ready?

Can you design it for scale?

💡 Final Thought

Rate limiting looks simple.

But real systems are never simple.

The difference between “working code” and “production-ready system” is where real engineering starts.

---

Would you trust AI to design your backend systems?

Let me know 👇