Hey Dev.to! I'm a 10-year developer and stay-at-home mom. This is about the day I realized everything I learned about coding might be obsolete, and why that's actually exciting.
My 2-year-old knows what he likes. Since before his first birthday, he's been excited by construction vehicles, cars and trucks, and anything else with an engine that goes. He watches videos of trucks and trains, and any game in GCompris with a vehicle in it (there are many) becomes a fast favorite.
So when we told him that he could make his own game, with any kind of vehicle he wanted, it was an unsurprising jump to:
"Make a red car game!"
I explained it like this:
"You type what you want here, and the computer helper will try to make it for you."
His eyes lit up. "Make a red car game! Tell the computer to make a red car game RIGHT NOW."
Why I'm Not Teaching My Kid to Code
I've been in software development for a decade. I've written more functions, loops, and algorithms than I could possibly count. And yet I'm convinced that teaching my toddler how to write a proper if…then would be a complete waste of time (and not because he wouldn't get it—he's already an expert at "If you finish your meatball, then you can have more blueberries.").
The point is that he's not going to grow up in the world I learned to code in. That world is gone. Learning Python will be akin to learning plumbing. Does it make you more knowledgeable and self-sufficient to know how the water gets to the faucet and how to fix a broken pipe? Yes. But 99% of the time your focus is on what you're doing with the water.
LLMs (Large Language Models—or, colloquially, "AI") are already impressive. By the time my son is working-age, they'll be unrecognizable. Well, more accurately, they'll be infinitely recognizable—invisibly embedded into every bit of technology we use, from our wristwatches to our televisions. (For future readers: wristwatches are devices we used to wear on our wrists to tell the time, and televisions are… oh, nevermind.)
If you want your kids to thrive with tomorrow's technology, please don't teach them how to code.
Teach them how to build.
My Toddler, the Prompt Engineer
My son knows his letters but we're still working on keyboard skills, so we helped him type it in.
make a browser based red car game
Visual Studio Code lit up, green lines filling the screen, things happening. He watched excitedly. "Can we play it now? Is the computer done yet?"
Eight minutes later, we opened the browser window and there it was: a little red car, trundling down the road on a sunny day, jumping over traffic cones and obstacles at the press of the space bar. Then,
"Can we make it a digger? Make a firetruck game. Make an AMBULANCE game!"
(He's two, after all.)
My little prompt engineer has been busy: you can see all the games we've made so far at madladstudios.com.
Technical Context
Here's the super simple setup: We're using Cline (formerly Claude Dev) with VS Code. It's essentially an AI pair programmer that can execute terminal commands and write files. The magic is that it understands natural language instructions and iterates based on feedback. Total setup time: ~10 minutes. No configuration needed beyond API keys.
The beauty is that my toddler doesn't need to understand any of this. He just knows that when he describes something clearly, the computer can help him build it.
Forget Syntax. This is What Actually Matters.
My son thinks we made a little red car game that day, but what we achieved was much bigger than that.
He used to think computers were things that showed him other people's stuff. YouTube videos of excavators. Educational games someone else designed. Now? The computer is his collaborator. When he has an idea, he knows he can build it. The computer changed from a tool for consumption to a tool for creation.
But more importantly, he's learning foundational skills for the type of programming that will be most relevant in his lifetime: natural language programming.
I watched him figure out that specificity matters. "Make a fire truck game," left a lot up to the model to assume, but, "make a fire truck game where it sprays water and puts out fires," got him just the game he wanted. When the red car's short jumps landed it on top of obstacles more often than over them, "make it easier for kids," got him bigger, loftier jumps that kept the game frustration-free.
What fascinates me as a developer is that he's essentially learning prompt engineering through play. He's discovering concepts like:
- Context boundaries (when to start a new request vs. modify the existing one)
- Iteration patterns (build, test, refine - our core dev loop)
- Specification clarity (the more precise the request, the better the output)
- Edge case thinking ("make it easier for kids" is user empathy in action)
These are the same meta-skills I use daily, just without the baggage of syntax.
He's discovering that creation can be a conversation. You say what you want, you see what you get, you refine, you try again. It's the same iterative process I used as a developer, minus all the unnatural syntax that gets in the way.
The tools have fundamentally changed. When any child who can describe an idea can also build it, we need to rethink what we're teaching them.
Some parents want to teach their kids Python. We're teaching ours to think in systems and communicate clearly. To break down big ideas into smaller pieces. To iterate based on results. To imagine boldly and describe precisely. To build with timeless skills in tomorrow's world.
Your Toddler Can Ship Software Today
Want to try this with your own kids?
You'll need a computer, an AI assistant like the Cline extension in VS Code, and a kid with ideas. The setup takes minutes—after that, you're limited only by naptime and snack breaks. (If you want specifics, comment and I'll help you out!)
Start with what your kid already loves. Let them describe it in their own words. Type exactly what they say—toddler grammar and all. Watch what happens. The moment they realize their ideas can become real things is when everything changes.
My son still asks for the red car game. But now he also asks, "What should we make today?"
I can't think of a better question to teach a kid to ask.
Have you tried building with AI and your kids? What did you discover? I'd love to hear your thoughts on what skills actually matter for the next generation of builders.
And yes, my toddler really does check the comments with me - especially if you mention vehicles. 🚗
Top comments (20)
I hope this is just a clickbait post. Why should a two-year-old be in front of a screen “shipping” and “vibe coding”? I’m asking as a father of a 2-year-old.
What happened to letting kids be kids? Why are we rushing them into adult concepts when they should be learning through play, building with blocks, and discovering the world around them? The tech industry’s obsession with starting early seems to ignore what’s actually best for child development.
I appreciate your concern as a parent of a 2-year-old too! We actually spend most of our time on exactly what you mentioned: playing with blocks, digging up rocks (his collection weighs more than he does now 😅), reading stories, and he’s super into building 3D wood puzzles right now.
The computer time is maybe 10-15 minutes when it happens, and it IS play for him. It’s like digital blocks he can build with. He's creating, not consuming.
I think the main difference is we’re not teaching “coding” or pushing adult concepts. We’re just using a tool that lets his imagination become something he can see. When he says “make the car green!” and it happens, that joy is pure kid magic.
Kids naturally want to do what they see their parents doing. Since I work with computers, he’s curious about them too. But trust me, he’s way more interested in his 3D puzzles than anything on a screen!
That's thoughtful!
I wonder if "the new way of developing" will attract a different kind of people than "the old way" did/does - less nerdy/geeky/one-dimensional?
You also gotta wonder if there won't inevitably be a shift from software dev being done by "developers" to it largely being done by "business people", or subject matter experts?
Or will it still be 'developers' doing this work, but with a very different approach and mindset ...
Will there still be a place for, not so much remembering programming language syntax, but understanding fundamental system concepts e.g. relating to the internet and the web - HTTP etcetera ... or will even that largely be irrelevant?
My hypothesis is that fundamental concepts (masters in computer science stuff) will become specialized knowledge. Most people working in software development won't ever need this knowledge, but a small handful of people working on specific systems will be in great demand for having it.
Today's top product managers will become tomorrow's top individual contributors. The better you are at communicating (with people or machines) the more success you'll have.
Well I think you're right - all of that "low level" stuff increasingly gets abstracted away, to the extent that it hardly matters anymore - that means the profession changes a LOT, and you could argue it won't be the same profession anymore, but a fundamentally different one ...
Maybe for a lot of "development" we won't actually need "developers" anymore - business people or subject matter experts might be able to do most or all of it ... or, they would still be "developers", but a completely different breed, with very different skill sets, and attracting a different kind of people.
I wonder if a new/different "refuge" will have to be found for the nerds/geeks among us? ;-)
P.S. I think you're completely right in not trying to teach your 2 year old how to "code" - I think it would be a futile endeavor, and pretty much guaranteed to turn him off ... ;-)
As an engineering student, your post is insightful. I'm also learning game development with GDevelop, a no-code engine, so I have to focus purely on logic. when I got into engineering, I had a simple plan: use existing game or app code, modify it, add other functions, and then ship it. Then came AI, which makes such things 1000 times easier because now I don't need a team of 10 people or months to complete my project. However, learning to code is still important because AI models can hallucinate, like creating their own libraries that don't exist
First of all, I commend you for getting your 2-year-old to sit still long enough to join you in this exercise. This is indeed a lovely story about spending time with your kid before they enter an age where they hate everything you might try to get them to do.
I apologize in advance for raining on this parade, but to me, this story is tarnished by a couple of underlying horrors that I think you are ignoring. If you don't want to hear my criticisms, please read no further.
The first horror is that your son (and my daughter!) will grow up in a world teetering on the brink of environmental collapse. Even if it's a collapse we can manage to stave off, the damage we are doing is going to have long-term consequences, and it's not just ignored by the companies providing AI, but in many cases it's directly caused by them. Your son might grow up in a world where he has to purchase clean drinking water because of our rampant consumerism, for example.
Of course the other easy criticism is about the lack of creativity in what you've created. I took a look at the games you've "made", and it's impossible to call them anything but simplistic and derivative. Maybe you already know this, but maybe you don't! Because one truth about AI in development is that it's entirely possible to use it to create derivative works without knowing you're doing so. Aside from the environmental impact, I find this to be one of the most damning shortcomings of the technology. Certainly your son doesn't know he's done something so pointless, but you should know (and almost certainly already do know) that better games exist in the world already.
The joy sparked by seeing something you had a hand in creating is real, and quite palpable. But I think you'd be better off fostering that effect by buying him some crayons. Or let him build things in the dirt outside... while that's still an option.
He's two, so that's fine, and it's a great place to start. But it's only a start.
Does he care that I hacked his game so that I can make my car fly or pass through objects and made the car black?
Because I did.
It's a cute browser-based game, and awesome for a two-year-old to build, and see that he can use a computer to make something cool.
But, it's only a start. Because if all he ever learns is how to prompt the LLM, he's never going to learn to actually build a game that people use for more than a silly diversion.
Please don't misunderstand what I'm saying. Because I think this is awesome. I think it's great that LLMs have introduced your kid to computers at this early an age. But this is not the end of programming. This is a great way to introduce him to programming concepts so that he can actually learn to code in the future. Because as long as there are computers, there will be programmers. We're not getting replaced by prompters.
That's super fun! Can you post a screenshot next time? I bet he'd love to see it
Now he wants to make a flying car game 😂
By the way:
I first introduced my kids to programming by getting them to hack Cookie Clicker, although they were much older than yours.
I have to say this was super interesting.
The idea behind everybody should learn how to code is not only about learning a syntax, but learning how to think in an organized way.
Coding is just enough abstraction to make something quick without consequences.
With plumbing you also need to be organized, but the consequences can be bad if you test it in your own house.
I don't think the syntax is bagage. You can't check the application if you can't read how it works.
It is like reading a comic without text. You can figure out the main story, but none of the side lines and nuances.
I think that education is too focused on data retention. Math should teach logic but for a lot of people, me included, it is an abstraction too far. And that is where coding can come in as a tool to teach logic.
While I think vibe coding is good to learn how to communicate. But once you start "prompt engineering" you took it a step too far. Communication amongst people should be the focus. Talking to a machine should be a side effect.
I think we agree more than we disagree! You're absolutely right that coding teaches logic and abstraction - mine certainly improved from the 10 years I spent doing it.
What I'm suggesting is that AI tools now let us teach those same concepts (logic, systematic thinking, breaking down problems) without the syntax barrier. My 2-year-old is learning iteration, debugging, and precise communication (all the important parts you mention) just through natural language instead of code.
Your comic analogy is perfect. Traditional coding is like making kids learn how to draw comics before they can tell stories. AI lets them start with the story itself.
I love that you brought up communication being key. That's exactly what we're focusing on. When my son has to explain what he wants clearly enough for the AI to build it, he's learning the same systematic thinking that coding teaches. He just gets to see results in minutes instead of months.
Maybe the real shift is from "everybody should code" to "everybody should understand how to think systematically and communicate" - whether that's through Python or plain English?
This website was once good, but now is full of these clickbait articles with the worst type of Linkedin delusional bs. Yes, sure.. your 2 years old kid knows what a browser is and wrote a prompt to an AI mentioning it.
EXACTLY ;)
GENAI is kind of ok but to burst a bubble its not likely GenAI that your son has access to is going to get much better without destroying the earth.
Meanwhile your son can eat mash potatoes and a few years outperform the AI in certain tasks.
So curb your enthusiasm your son may grow up to not want anything to do with tech probably after you grow up you will also have a similar sentiment.
The real world is far more interesting.
I disagree that being a software engineer will be worthless in the future. I believe it will be more important than ever as more and more people forget that fundamental concepts are what matter in this. You cannot properly conduct an orchestra without understanding music and what each player in the orchestra does. Standing up on stage and wildly flailing your arms might seem like it's the same thing, but it certainly is not, nor will the orchestra give you the same results.
It's like the people who only know ORM's and don't know SQL. Sure, that works 80% of the time, but the other 20% of the time is what's going to cause 95% of your serious pain points with the data, and ORM's cannot abstract that away, you will have to get in there in SQL and figure it out. You can't debug what you can't see.
I am not an AI hater; in fact, the opposite. I use it every day at work, integrated into our IDE's in the form of Co-Pilot. I use it to help proofread, brainstorm, and act as a copy editor for my books and blog posts. I use it to help write unit tests, doc strings, and to set up and remove relevant console logs at lightning speed for debugging, which has been awesome to help speed things up.
But at the end of the day, AI is only as good as the person prompting it. If you can't prompt it with precision and you can't understand what it's giving you as output, you are simply putting yourself in a boat of "I guess this is right because it says it is".
And that's not good enough for production and never will be. There are also pretty severe limitations with AI, especially as you get into any kind of complex systems, 3rd party integrations, proxy layers, etc, where it's virtually useless or will have you chasing your tail for hours in a circle if you let it. I'd say it's helpful for roughly 30% of what I do on a day-to-day basis.
Right now, we are at a saturation point with what is possible, meaning we are close to the "best" we can get with the current way things are done, so unless there is some sort of major change in how these work, things probably are not going to improve all that much in the near future. It will simply be a lot of "bells and whistles," adding, like the Madden Video Game series for football, where most years are minor updates and "cool things," but it doesn't fundamentally improve anything in the game itself.