Why VSCode Still Makes Me Feel Alive

Nabil Abubakar — Mon, 10 Aug 2026 17:55:48 +0000

Every time I open Cursor or Antigravity to write code or run codex, claude, grok.
I feel dumb.

Not immediately. But somewhere in the loop prompt, accept, read the diff, push I notice I've stopped thinking. I'm just shipping. Reading code someone else's model wrote, watching reels while it "thinks," accepting suggestions, pushing. Repeat. My brain is on autopilot and I'm just the guy clicking accept.

Then I open plain VSCode. No Copilot, no Cursor, no AI anything. Just me, a blank file, and the problem.

And immediately I feel alive.

It's like my brain took something. Not a drug more like it finally woke up. I have to actually hold the problem in my head. Think through the data flow before I type it. Remember the syntax instead of glancing at a suggestion and pattern-matching whether it looks right. Every line is a decision I made, not a decision I approved.

That's the difference. With AI tools, I'm a reviewer. Without them, I'm the engineer.

I'm not saying ditch AI coding tools that would be stupid at this point obvisiously, they're too useful and too fast. But I've started noticing what they quietly take from me if I let them do everything: the muscle of actually thinking in code. The kind of thinking you only build by struggling through a bug for twenty minutes with no autocomplete whispering the answer.

So now I try to write parts of the code myself. On purpose. Not because it's faster it's not. But because it's the only way I stay sharp enough to know when the AI is wrong, when the suggestion is subtly broken, when the clean generated function is actually solving the wrong problem.

Sometimes the slower way is the only way that keeps you in the game when you don't have a deadline or you could grind leetcode or neetcode problems in your free time to keep your brain active.

I indexed 110+ software laws so AI agents and dev teams can stop breaking architecture

Nabil Abubakar — Tue, 02 Jun 2026 08:58:10 +0000

description: From Conway's Law to the CAP Theorem I built an interactive playbook to integrate classic architectural wisdom directly into your AI workflows and dev teams.

Every experienced software engineer has been there.

You spend months meticulously designing a beautiful microservices architecture. You map out the data flows, write clean code, and set up your pipelines. But six months later, the system is a tangled, unmaintainable mess. The code doesn't mirror your original technical blueprint it mirrors your company’s messy Slack channels and fragmented team structure.

You didn’t just hit a technical bug. You ran face-first into Conway’s Law: Organizations design systems that mirror their own communication structures.

We talk about these foundational concepts Conway’s Law, Brooks’s Law, the CAP Theorem, Lehman’s Laws in passing conversations, tech blogs, or post-mortem meetings. Yet, despite how heavily they dictate whether our codebases live or die, this collective engineering wisdom remains scattered across old academic papers, random blog posts, and obscure GitHub threads.

As an indie developer, I got tired of watching these principles get ignored, especially as AI tools generate more code without any context of architectural constraints. What started as a small personal project to read and track these classic rules has scaled into an AI-powered compliance system.

Yesstarday, I launched DevLaws.xyz.

What is DevLaws?

DevLaws is an interactive, modern playbook and integration hub built specifically for software engineers, tech leads, and technical architects who want to build resilient systems and keep both human developers and autonomous AI workflows aligned.

Instead of reading dry, textbook definitions, it is a centralized platform designed to make engineering wisdom actionable and integrated directly into your development cycle.

Here is what you can do on the platform right now:

Explore 110+ Codified Laws: Deep dives into software design, distributed systems trade-offs, and team velocity dynamics.
AI Agent Integration: Hook your AI agents and coding tools directly into our compliance backend to ensure automated code respects foundational architectural boundaries.
Analyze Real World Case Studies:* Move past abstract theory. Read production breakdowns showing exactly how these laws play out in real companies.
Map Principles to Active Projects: Brainstorm, track, and commit specific structural rules directly to the blueprint of your live repositories to prevent technical debt.
** Align Teams with Visual Cards:** Easily generate and share visual law cards to educate your developers and get everyone on the same page.
Track Trending Concepts: Stay ahead by seeing what architectural mental models technical teams are researching weekly.

Built for Builders (I need your feedback!)

I am actively updating and building upon DevLaws daily to ensure it solves genuine production headaches. Because the platform literally just launched, I am actively looking for engineering teams to jump in, test the AI/integration workflows, and absolutely roast the UI and feature set.

If your engineering team is open to checking it out and helping a solo builder gain some early launch visibility (reposts, sharing with colleagues, etc.), I am offering 3 months of our Enterprise plan completely free to early adopting teams. No hard sales pitch just a builder looking for authentic feedback to shape the roadmap.

Let's chat on x!

Stop leaving your system architecture to chance. Check out DevLaws.xyz and let me know your feedback below.

How are you currently ensuring that AI generated code doesn't violate your system architecture?
What integration or feature would make this a daily tool for your team?

Let's also discuss in the comments!

DEV Community: Nabil Abubakar

Why VSCode Still Makes Me Feel Alive

I indexed 110+ software laws so AI agents and dev teams can stop breaking architecture

What is DevLaws?

Built for Builders (I need your feedback!)

Let's chat on x!