So after waiting so long for my 110th subscriber... finally she is here, and I was thinking what I should reward her with... but as I am in my absolute broke era, I wanna promote her channel. So please give her a follow and show her love🫶


https://t.me/my_dev_diary

I’ve always wondered what truly differentiates senior engineers from vibe coders — or high-quality engineering from AI-generated code.

And I think I finally found a somewhat satisfying answer.

Senior engineers had to cook the meal themselves for years before the “AI chef” arrived
So they know exactly what the system is supposed to taste like. They immediately notice when something feels off.


But this generation may not even get the chance to fully experience building everything manually once, because the environment itself is changing too fast.
What senior engineers really built over time is something deeper which is system thinking.
And honestly, they paid a lot to gain that skill.


System thinking is understanding:
👉what you’re building
👉how it works
👉why it works
👉how different parts integrate
👉what breaks if one piece is removed
👉It’s when you can mentally visualize the entire flow of the system — even with your eyes closed.


As juniors, I think these are the kinds of questions we should learn to answer:
👉Where does the state live?
👉Where does feedback live?
👉What breaks if I delete this?
👉How are errors handled?
👉How does the logic actually operate?
👉How do architecture, services, and components connect?
👉How can problem-solving skills (like LeetCode thinking) be applied to real systems?

If you can answer those questions in your head without touching the code, then vibe coding actually becomes powerful.

And another huge advantage this generation has is this:
we can specialize deeply in one area while still being capable across many others using AI.

For example, a backend-focused engineer can now create frontend interfaces with AI assistance. That creates more opportunity to move faster while still mastering a core specialty.

So maybe the future is not just “learning how to code.”
It’s learning how to think like a real engineer.
👉To build systems.
👉To reason clearly.
👉To understand deeply.

That engineering spirit will definitely matter more than memorizing syntax.
A few practices I think are important:

• Design before coding — even with just paper and pencil. Force your brain to visualize systems.
• Write the what and the why before AI writes the how.
• Try the deletion test — remove a component and observe what breaks.
• Don’t blindly accept AI-generated code. Read it. Question it. Understand it.

Because at the end of the day, AI can generate code…
but it still takes an engineer to build systems.

I am falling in love with Claude... Why did no one stop me when I was using ChatGPT before? 🤦🏽‍♀️

So sifen decided to work on this project and share what he is learning along the way, and I decided to code and learn alongside him... so

—> Am I capable? I don't know until I check.
—> Do I think I am capable? No, but I am labeling this feeling as imposter syndrome, and I am not approving it😁. I am serious when I negotiate with my feelings. So..
—> Are we doing it? Yes, we will... I am going to try hard; either I get a fully working system or an MVP with a lot of experience... I think both are a win for me.

So here I am learning in front of your eyes and sharing every little detail of everything I get along the way. So who is in?

Note: As I am in the final exam for the next two weeks, I may move slowly.

In some days, we may not be productive enough, and that is proof of being human... so, no regrets.

Good night, my dear friends🩵

Sunday Reflection

Most individuals seek certainty prior to taking action. However, nearly every meaningful life is constructed in the reverse sequence: action first, clarity second.

A significant portion of modern paralysis stems from over-optimization:
- Waiting to feel ready;
- Waiting for motivation;
- Waiting for the perfect roadmap;
- Waiting to become "the kind of person" capable of undertaking difficult tasks.

Competence is typically a byproduct of repeated exposure to discomfort.

Senior engineers are not calm because they were born with confidence. Founders are not decisive because they possess magical certainty. Disciplined individuals are not consistent because every day feels easy.

They became these individuals after enduring sufficient ambiguity.

The uncomfortable truth:

Your identity follows your actions more than your feelings,
and your mind trusts evidence, not intentions.

Therefore, every time you:
- Study while confused;
- Code while insecure;
- Speak while imperfect;
- Build before you feel qualified;

you are casting a vote for a future version of yourself.

This is not dramatic transformation. It is accumulated proof.

The world often glorifies talent.
Reality rewards sustained engagement with difficulty.

ነፃ ማስታወቂያ እድል! _ Free Advertisement Opportunity!

#Opportunities_Finder_Ethiopia

Companies, organizations, or individuals: Promote upcoming events, workshops, or participant invitations to thousands of engaged Ethiopians (youth, students, professionals).

- What we offer: Free posts with high visibility.
- Your audience: Students, leaders, professionals in Addis Ababa & beyond.

Contact @Benn_amen now

ቤተሰብ እንዴት አላችሁ?

when I was scrolling through x I found out this resource: AI Engineering from Scratch check it out ... + I have previously sent a text based free course site namely: apxml.com and also check that out ...


ig it will be useful

@sudojajos

By the way, do I have any beginners here, or people thinking of starting their web development journey?

I was thinking to share some insights on "what I would do if I had to start learning coding today," but I think most of you are already on the track, so I wanted to ask you.

I am a beginner myself, and this channel is where I share my journey. I wanted to tell how I got where I am so we're all on the same page, which makes sharing daily tips and progress more practical. So, what do you guys think?

I will write tips on how to get started tomorrow, insha'Allah.

I have been thinking all day about how to write something beneficial. Should I write about the strategy that should be followed, or should I walk you through the approach and the roadmap?

You can get the information I give you from a distance of opening ChatGPT or Grok, but anyways, let me share the things I wished somebody would have told me when I was just starting.

**First things we should acknowledge as a beginner.**

- We are in the age of AI, and code is cheap.
- There are people who have started earlier and are doing magical things. These people may be our friends, someone you know, or even your junior by age or class.
- It's not easy, and it won't be easy. You will have to be disciplined and consistent as you never have been before if you want to achieve something satisfying in the shortest possible period.
- Starting this journey with a friend or a community is a plus, but you may not figure out things as fast as your colleagues.
- Finally, you may have to choose a tech influencer for your day-to-day social media consumption, especially if they are someone from around, like school or someone you know. Some people are so skilled at making things seem impossible.

Note: Everything I mentioned above is not a problem. You will learn how to handle things yourself in a short amount of time, and you will start feeling confident very soon. I'm telling you this so you don't have to quit at this stage.

After establishing the right mindset as a beginner, the next thing you need to know before diving into frontend/backend development or HTML/CSS is systems thinking (I have a post on systems thinking above; please read it). This covers how real engineers think, what you should think, and the mental map approach you should adopt on this journey.

At this stage, you can use AI or a senior developer around you (the better approach) to teach you how to think and what you should learn. Keep these concepts at the back of your mind:
- Systems thinking
- Debugging ability
- Architecture awareness
- Knowing what to ask AI
- Learning how to learn
- What you’re building
- How it works
- Why it works
- How different parts integrate
- What breaks if one piece is removed
- Mentally visualize the entire flow of the system—even with your eyes closed.

Believe me, if you always think about these concepts, what you learn will make absolute sense, and it will also be fun. You will start figuring out what you should learn yourself. You won't consume four or five hours of tutorials without building; instead, you will do a simple project and spend your time messing with it—this is real learning. With this mindset, you will begin the learning phase.

The next is the learning phase. Here, you can use AI to give you a real directive roadmap, ask a senior, or use roadmap.sh. Let me tell you something about this phase.

1. Believe me, you don't need that 5-hour tutorial; you only need a 30-minute to 1-hour introduction to the thing you are learning. Then, you will have to open your IDE to write something yourself.

2. For projects, you can ask AI to give you a project idea—not that damn "weather dashboard" or "blog site" thing. You can prompt it to give you step-by-step instructions to really do that project.

3. Personal suggestion: When you are doing a simple project for learning purposes, write the hard code yourself. Use Google to search or to figure out "how to solve this thing." ChatGPT-like AIs may take the thinking part from you, which makes you learn nothing.

4. Practice reading documentation. I know documents are messy junk and you may not know how to figure them out, but the concepts of HTML and CSS are simpler from your next learning phase, so they are the best to practice learning documentation too.

5. When you first ask AI or a senior to give you a roadmap, they will definitely tell you thousands of things you should master. That's a bit frustrating. Hold on, my friend—the only thing you need is clarity and that learn by building strategy. In no time, you will get yourself in the middle of the way already. How? Well, I don't know, but I think it's because you won't learn everything 100%. You will learn only the things that are necessary for the project you are building. Believe me, this will boost your confidence for future learning or for more complex projects.

6. Learn Git and GitHub early. It's not magic, but the greener your GitHub contribution graph, the better chance you will get. So, start committing from your first HTML project. Use AI, or it's better if you learn the commands as they are easy, few, and repetitive.

7. Learn to Google and ask AI about good folder structures, proper naming, clean commits, and code documentation from your first day of coding.

8. Design systems that solve human problems through technology. Starting from your HTML page, try to make it intentional.

9. The magic is being consistent in the right direction and with the right mindset.

**finally current world trends**

The industry is shifting toward:
- AI-assisted development
- agentic workflows
- rapid prototyping
- smaller teams doing more
- product-focused engineering
- automation-first workflows

THERE FOR

The future engineer:
- thinks strategically
- uses AI effectively
- understands systems deeply
- communicates clearly
- ships fast WITHOUT losing structure

So the thing is, if you apply the above tips and you know what you are doing—prompting AI to perform tasks, debugging if something fails, correcting mistakes the AI makes, and guiding it effectively will be a piece of cake.

I hope my suggestion will help someone for real. It's based on things that took me more than a year to understand, and share it for the one you think will need this. 😊

Quick dev security note (VS Code / Cursor / GitHub users)

There’s an ongoing supply-chain security issue being discussed called PolinRider, which targets open-source developers and their GitHub accounts.

The concern is that some developers’ accounts and access tokens may have been compromised, and that access is being used to:

1. modify repositories
2. submit malicious pull requests
3. Even force push & Inject malicious code into projects

Since many modern apps rely heavily on open-source libraries, this kind of attack can quietly spread through dependencies without being obvious at first.

Also important: private repositories are not automatically safe either, if your account, tokens, or CI/CD access is compromised, private code can still be accessed or modified.

Basic things to double-check:

1. Recent commits / force pushes in your repos
2. GitHub account access + authorized apps
3. VS Code / Cursor extensions you don’t recognize
4. Dependency install scripts in projects
5. Run security audits (npm audit, etc.)

Just sharing for awareness, especially for anyone actively building with GitHub + modern dev tools.

@edemy251