What Will Become of Documentation and Coding Tutorials on YouTube

In late 2019 / early 2020, I wanted to learn how to program computers and figured I would need a programming language. I picked up the C Programming Language book to learn C. I didn’t end up finishing it and ultimately gave up on C (I’ve shared the story here before).

But that was it—books were the primary way for us to learn coding, at least in the space I was in. I remember sharing PDFs of programming books back then in Polytechnic with my friends.

Fast forward two years later. In 2022, a friend introduced me to web development. I had tried doing AI/ML stuff using Python and Julia, but seeing him build cool web projects made me want to try it out. At least AI/ML was boring, I told myself (poor me—if only I had seen the wave coming in 2023).

I asked him, “So which books am I going to read to get started?” He laughed and replied, “Sleep on YouTube, bro.” He went on to say that whatever I wanted to learn—HTML, CSS, or JavaScript—I should just search for it on YouTube, learn, and whenever I ran into trouble with code, read the documentation of the tool involved.

Man, that was a whole new world for me. “How is a 3-hour tutorial video going to be detailed enough to teach me everything?” I asked. But man, I was dead wrong.

Fast forward two weeks later: after putting his advice to work, I was able to build my first website—all on my own, without watching or following any guide. This was crazy and fascinating to me.

For the next three years of my career as a software engineer, YouTube was the first place I went whenever I wanted to learn something new, no matter how hard it was. And truth be told, 90% of the time, I left knowing what I came to learn and started building cool stuff.

Of course, I later found out that some of these tutorials were outdated and that you should also read the documentation of the library to gain a proper understanding of things. But still, YouTube was the first introduction—a way to get a feel for what it was like to use something new.

Do I still read books on coding? Of course, I do—but infrequently. Now, it’s more for fun. I love reading in general and have an insatiable appetite for knowledge, so whenever I get the time, I pick up one of the classics on coding.

Fast forward to today, four years later, and I’m in a completely different world.

A world where coding channels on YouTube are turning into podcasts because there’s lower demand and watch time for tutorials, as newcomers learn coding with a chatbot.

A world where coding channels once dedicated to specific topics—areas the creators were deeply experienced in—are now teaching AI. Sometimes you can tell they don’t yet have deep expertise in these areas. But the market demand is high, so everyone is jumping on the trend, making older tutorials feel outdated.

A world where documentation websites now include files like LLM.txt, SKILLS.md, or a “Copy Markdown” button that you can download and feed to an AI to explain the library and generate code for you.

I get it. I’m not going to sit here and blame anyone. They know coding tutorials and documentation are losing relevance.

Nobody watches videos anymore on “how to create a hamburger menu.” They just ask AI for the code.

Nobody spends three hours watching a video on how to learn React anymore. They skim a bit, then ask AI questions as they go.

Nobody spends a whole week reading documentation before using a library—at least to understand the thought patterns that shaped the creators’ design decisions. They start building and ask AI along the way. Occasionally, when AI hallucinates, they remember there’s actually documentation with up-to-date information.

And let’s not forget technical blogs—articles titled “How to …” or “Deep dive into …”. Fewer people read technical blog posts now. The ones gaining traction are mostly in the AI space.

Deep technical knowledge feels like it’s becoming a thing of the past. I mean, if AI can give me the code within minutes, why go through all that stress?

This isn’t me hating on AI. I would be foolish, as a software engineer, to do that.

AI has 10x’d my productivity. It does this because code is now commoditized and can be produced on the fly. What used to take me weeks to implement—not because I didn’t understand the idea, but because writing clean code takes time—now takes a day. Why? Because I can write specifications and get fairly working, often clean code in a matter of minutes.

It’s truly a game changer.

So why am I writing this? I’m just concerned about the future.

I’m concerned about what the future holds for upcoming software engineers.

I’m curious about the world they will grow up in and how they will be introduced to coding.

I wonder whether Harvard’s CS50 classes will still be relevant years down the line.

I wonder how future software engineers will develop deep technical knowledge and strong foundations in the tools they use before they go out and build software for the rest of us.

I’m genuinely curious.

We live in a time when no one has it all figured out—not even those leading the charge in AI.

I’m just here wondering and soliloquizing about what will be and what won’t.

If you started programming after 2023, I know it’s tough.

It’s tough to read documentation when an AI is right there to generate good code in minutes.

It’s tough to watch tutorials to deeply understand how experts approach tools and the design choices they make, when AI can generate full-blown apps for you.

I know it’s tough.

But I can assure you that if you take your time to go beyond the chatbot and pause to understand the tools you’re using—even to some level—it will help you cross-check your AI-generated code to ensure it is:

• readable
• maintainable
• reliable
• scalable
• safe and secure

and much more.

Don’t let AI make you lazy. Let it empower you.