Introduction

I recently took a stab at a scalable distributed counter implementation. The inspiration comes from an common System Design interview question: “Design a distributed counter that can handle millions of events per second and provide near real-time read access to the current count.”

I read some about it in theory and design interview questions, but I never actually see it in action. Thanks to the advance of AI, a lot of the implementation burden can be handed off to an AI agent, while I focus more on the overall design.

So now I added it to the Golang Dojo project as I thought it could be a fun and education project to deepen my understanding of distributed systems, and also to explore how AI can assist in such process.

Source code is at: https://github.com/hxy9243/go-dojo/tree/main/src/counter

Read More

Learning Golang with Real Projects

There’s a particular itch I get when learning new tech concepts. Reading docs and watching tutorials helps, but the knowledge stays hazy until I open the hood and build something myself. So I started a project: implement small but interesting systems in Go, working with the garage door open.

This isn’t pure “vibe coding.” I’m using AI to scaffold and accelerate, but the goal is understanding, not just shipping. The first “Hello World” system: a rate limiter. It has enough complexity to be interesting, bounded enough to complete in a short span.

I started off the implementation by auto-complete with an AI copilot first, nothing more. And throughout this project, I’ve learned more than I expected. Now I truly believe that creating something greatly helps understanding. And AI helps much of the implementation burdens.

In the future, I’ll experiment with more sophisticated “vibe-coding” techniques and see how far I’ll go.

Read More