I’m planning to work on a few personal projects in the (hopefully) great year of 2026. This is a rough plan for now. I don’t know exactly what shape it’ll take, but it feels good to start somewhere.
The core idea is simple: learn something new every week and share what I learn on the internet. The format doesn’t matter much blogs, videos, short posts whatever feels natural at the time. The important part is learning consistently and putting it out there.
Learning in Public
I want to begin the year by revisiting some well-known technical books. I’ve read a few of these before, but this time the goal isn’t just to finish them it’s to understand and retain what’s inside.
Re-reading with the intention of teaching or sharing forces you to confront the gaps in your understanding. That’s exactly what I want.
Books on the List
- Designing Data-Intensive Applications
- Database Internals
- Building Microservices in Go
More books to follow
Projects: Back to First Principles
Over the years, I’ve built quite a few projects. Most of them focus on using existing tools and frameworks to solve real-world problems and I still think that’s valuable. I strongly believe in not reinventing the wheel.
That said, this year I want to shift gears.
From my experience working on production systems and debugging real prod issues, I’ve learned one thing clearly: deep, internal knowledge of how systems work matters. When everything looks correct on the surface and all the safeguards are already in place, the real issue often lies deeper in the fundamentals.
So in 2026, I want to build things from scratch. Not because it’s practical, but because it’s educational.
Back to Go
I’ll be spending most of my programming time in Go again.
At work, I don’t get many chances to use it these days, and I miss it. I genuinely enjoy working in Go, and I strongly feel it’s going to dominate large parts of backend and infrastructure development in the coming years.
This year is about getting comfortable with it again writing more Go, understanding its internals better, and using it as my primary language for experiments.
Thoughts on AI (Briefly)
There’s no denying it—AI agents completely dominated the past year, and that trend isn’t slowing down anytime soon.
I consider myself an optimist when it comes to AI, but I’m also realistic. It’s not perfect. Still, the pace of improvement is impossible to ignore. I’ve worked with Claude Opus 4.5, and honestly, it’s the most intelligent coding assistant I’ve used so far. Of course, in this race, today’s best model is usually replaced within months.
That said, the AI space has become overwhelming. New tools, agents, workflows keeping up with everything is exhausting. I think it’s time to slow down, settle a bit, and focus on fundamentals again.
I’ll continue using Claude Sonnet 4.5 for my personal projects, but not in full agent mode. Since these projects are primarily for learning, I don’t want to delegate everything. Speed matters when building consumer apps; learning matters when you’re building for understanding.
I’ll write more about my thoughts on AI separately—this post isn’t the place for that.
Wrapping Up
The main goal for 2026 is simple: be more intentional and organized than I’ve been in previous years.
I have a proven track record of starting things with enthusiasm and abandoning them just as quickly. This plan is my attempt to break that pattern by learning steadily, sharing publicly, and focusing on depth instead of novelty.
Let’s see how this goes.
Hopefully, this blog post will still be here in 2027.
Thanks.