I used to think I was a "visual learner." I'd pick a video tutorial over documentation every time. Twenty minutes of someone explaining closures while typing in VS Code? Perfect. I'd watch the whole thing, nod along, and close the tab feeling like I understood closures.
Then I'd try to use one and realize I had no idea what I was doing.
This happened enough times that I started questioning the whole setup. Not the specific tutorials, but the format itself. Is video actually a good way to learn programming? Or does it just feel like one?
The Study That Started It All
The idea that multimedia is better for learning comes from real research. Richard Mayer's multimedia learning theory showed that combining relevant visuals with text can reduce cognitive load. A diagram of the event loop next to an explanation helps. That part is solid.
But the industry took "relevant visuals help" and turned it into "video is better than text." Those are completely different claims. And the second one was never proven.
Most programming tutorials aren't even multimedia in any real sense. They're someone talking over a code editor. There's nothing visual about explaining JavaScript closures that text couldn't do better. It's just text delivered slowly through someone's mouth.
Why Reading Is Harder (and Why That Matters)
When you read, your brain does real work. You build the mental images yourself. You pause on a confusing sentence, re-read it, think about it, move on. You're constructing a mental model from scratch.
When you watch a video, someone else did that work for you. The instructor visualized the concept, chose the metaphor, paced the explanation. Your brain is just receiving.
That's why video feels easier. It is easier. And easier means less encoding. Less effort means weaker memory.
I think about it like this: reading a novel and watching the movie are completely different experiences. The reader who imagines a character's face remembers the story differently than the viewer who was handed one. One exercises the mind. The other entertains it.
Programming concepts are abstract. Variables, closures, promises, type systems. None of these are visual. They're symbolic and logical. Text with the occasional diagram is almost always the better format for this kind of material, because it forces you to do the mental work that creates understanding.
Where Video Actually Wins
I'm not saying video is useless. It's the right format when the information is spatial, physical, or auditory:
- Tool navigation. Seeing someone use a UI, understanding spatial layouts. You need to see this.
- Physical skills. You can't learn to swim from a book. Motor learning needs demonstration.
- Spatial reasoning. Circuit diagrams, architecture layouts, things that are actually visual.
- Sound. Pronunciation, music, tone. Text can't capture these.
But for abstract concepts like programming? The video is doing something text already does, just slower and with less effort from you. And less effort is the problem, not the solution.
The "Learning Styles" Thing
"I'm a visual learner" is one of those things everyone says and nobody questions. The idea that people have dominant learning styles (visual, auditory, kinesthetic) has been studied a lot. It's mostly a myth.
What research does support is different. It's not that people are visual or auditory. It's that subjects are. Swimming is kinesthetic. Music is auditory. Programming is abstract and symbolic. You match the format to the content, not to the person.
Saying "I'm a visual learner" when studying TypeScript generics is like saying "I'm a hammer person" when the job needs a screwdriver. The preference is real. The effectiveness isn't.
LLMs Didn't Replace Learning. They Replaced Lookup.
Here's what I think happened. Tutorial videos were never popular because of the video format. They were popular because they were the best guided explanation available.
Before LLMs, if you were stuck on something specific, your options were:
- Read the docs (often poorly written, assumed too much)
- Search Stack Overflow (fragmented, someone else's problem)
- Watch a 20-minute video just to find the 2 minutes you needed
For that use case, LLMs are better. They're instant, specific to your problem, and you can ask follow-ups.
But there's a catch. LLMs work when you already know enough to ask the right question. If you understand TypeScript and you're stuck on a conditional type, an LLM will get you unstuck fast. If you don't know TypeScript at all, you don't even know what to ask. You'll get an answer, but you won't know if it's the right approach, whether you're learning the right things in the right order, or what you're missing.
LLMs are a lookup tool, not a learning system. They don't give you structure, progression, or a way to test whether you actually understood what they told you. They solve the problem in front of you without building the foundation underneath it.
For getting unstuck, LLMs win. For actually learning a topic from scratch, building the mental model, in the right order, and checking it stuck, you still need structured material and active recall. The format of that material matters. And for abstract concepts, text beats video.
So What Does Work?
If video tutorials create the strongest illusion of learning, and re-reading creates the second strongest (I wrote about that in the previous post), what's left?
Reading. Actual reading. The kind where you engage with text, pause to think, and then test yourself on what you just read. It's not glamorous. It doesn't feel as productive as watching a polished tutorial. But the effort your brain puts into processing text is exactly what makes the knowledge stick.
The best format for learning abstract concepts is condensed text followed by active recall. Read the concept. Close it. Try to explain it. Take a quiz. See what you actually retained versus what you only thought you understood.
It's slower. It's harder. It's uncomfortable. That's the whole point.
This is part 2 of a series on effective learning. Previous: You're Not Learning, You're Just Reading