Abhishek Saikia
38 posts

Abhishek Saikia
@king_rog234
Learning is an adventure and I am in one.


Need a quick reality check on an idea. Over the past few months I've been building a bunch of systems for myself to solve problems I kept running into. Now I'm wondering if they're useful to anyone else. I'm thinking of opening up weekly mentorship + the exact systems I use. Right now there are two: > Gym System – The first version is already on my YouTube, but I've gone through 7+ iterations since then. It's much more practical and backed by better research now. > Learning System – This is the one I'm thinking of launching with mentorship. One thing I want to be clear about. This system won't guarantee you a job. What it will do is help you build a much deeper understanding through bite-sized lessons every day. For example, I learned tRPC in just 5 days using only 9 lessons. But it wasn't just watching lessons. The system constantly made me write code, validated my solutions, suggested fixes, pointed out optimizations, and highlighted what I could've done better. It forces you to learn by building instead of just consuming. I think it'd be useful if you're: > Working a 9 – 5 and still want to keep learning. > Heavy into vibe coding but now want to understand what's actually happening. > Looking for a structured learning framework. > Just getting started with coding. Probably not for you if: > You're only looking for a shortcut to get a job. > You already have 2 - 3+ hours every day to study. > You're currently a student without an income, since you'll also need an AI subscription alongside the mentorship. I'm still validating the idea, so I have one question. What would you consider a fair price for something like this? (System Setup + Weekly Mentorship) I’m not trying to make it expensive, but I also don’t want to make it cheap just to get a lot of people in. I’d rather keep it smaller and focus on quality, since I’ll be spending a good amount of time mentoring and improving the system.

Whatever I learn, I shall learn form labs around it. This way, I can learn more by actually implementing stuff and also documenting for my future self. This is not a novel idea. In fact, many people already do this. Yet, many still don't.


I expected overlayFS to be a filesystem (FS lol). Instead it's a routing layer. The same pathname can point to different physical files over time. That's a much weirder design than I expected. But really interesting.

Built a tiny bash script that polls /sys/fs/cgroup/*.current every second. Set memory.max to 100M, ran a Python script that keeps allocating — watched it get OOM-killed the instant it crossed the line. memory.events even logs the kill count.


Learned cgroups today and manually tweaked cgroup v2 files on the shell--memory.max, cpu.max, pids.max, played with the raw stuff underneath. Fun stuff!



It is my belief that many devs right now are not maximizing what they can do with automatic programming because they still look at the code. Doing it makes you the bottleneck. Your time is better invested in new ideas, QA, design, and asking yourself what is your goal.








