"But There Won't Be Any Programming Jobs in Five Years!"
I field this objection all the time. My company, Jippity, teaches kids to code on a new kind of editor, one with a full embrace of AI, specially-designed for kids and beginners. Often the most technical parents worry that coding, the hottest, most marketable skill for so long, will become obsolete — worthless in the face of fast and cheap LLMs that can spew Python like nobody's business.
They are wrong. There has never been a better time to pick up some programming. AI demolishes so many of the barriers that kept half-curious kids and adults out. If there's one skill to acquire in 2025, it is foundational coding — comfort with basic reading and writing — and yes, AI will help you get there. Let me show you why. But first, some charity for the haters. Misguided though they may be, these parents' fears hold a kernel of truth. It used to be that name-dropping a few programming languages on your resume caused an instant 50% bump in available salaries. "Software Engineers," a broad and often nebulous class of workers, pulled in the biggest bucks, right out of college, so that a straightforward parent-conclusion arose: learn to code. But now the market has saturated enough that software engineering salaries are starting to fall, and moreover, AI, the great multiplier, means that fewer programmers can get more done for less money. It sounds nice, in fact, for those "already in the door:" replace all your junior engineers with Anthropic AI's Claude, cut down on meetings, negotiate an even-more inflated salary for the sharks and the graybeards who do the real heavy lifting. As the saying goes, "What one programmer can do in a month, five can do in five."
But even if such a scenario unfolds in some engineering departments, this is a narrow-minded objection, stuck in the old paradigm of Smart Kid → Computer Science degree → Software Engineer. Let us imagine some characters across a spectrum of age, skill, and interest, so that we can dig deeper into the supposed bleak fate of professional coding in the age of AI.
Rebecca, The Entrepreneur
Becky is fresh out of college. She's bright but no genius. High-energy but not some kind of twitchy freak. She has been nursing an idea for a couple years — an app that will fill a real niche. She knows it could be big; she can see the path. But it's an expensive path, with one line-item dwarfing the others. Hiring a competent programmer, who can handle iOS and Android, who can stay light on their feet, is a prohibitive expense, even for six months of development. Even if Becky wanted outside funding, she couldn't secure any without showing something. A huge personal loan is too big a risk to take...
In 2015, Becky's prognosis is not so positive. She could enroll in some kind of Coding Boot Camp which half-aligns with her vision. She could beg a student of dubious skill for their weekends, possibly giving away more percentage points than she'd feel comfortable with. As great or not as her idea may be, Becky does not get a real opportunity to try in 2015.
Fast forward to 2025. Now, Becky gets a fair shot. A few weeks back-and-forth with an AI coding assistant produces something like a mock-up skeleton. Maybe so too with design! Either way, two positive outcomes, both unlocked because of AI:
On the concrete side, Becky can take her next step as an entrepreneur. She can bring a robust demo to potential investors and early users. She can see what's working and not working — she can iterate. She can react and grow. Maybe it turns out to be a flop, but at least she gets an honest attempt.
Bigger picture, Becky learns to code. Even if she starts in the most naïve, misdirected way, she will be swimming around in code, learning through immersion. She will pick up high-impact skills and skip the boring academic stuff. Maybe she digs it and she can go deeper, looking ahead to a career in software. Maybe not, but by picking up some code-literacy, she unblocks herself. Her future is that much brighter.
Harry, The Designer
Harry's been bouncing around, contract-to-contract, designing web pages, app layouts, and graphics. He's got a sharp eye for color and he knows Figma like the back of his hand. He communicates; he delivers beautiful work; he doesn't waste time. But that full-time job simply won't materialize. When he hears the dreaded questions, he knows he's done for.
"Do you know Svelte?" "Can you make a React component?" "How's your HTML and CSS?"
These questions frustrate Harry. "Are the programmers really that much smarter than me?" he thinks. His usefulness gets bottlenecked, over and over again, by the arcane language of code. Harry sucked at math and finds no joy in deciphering the hieroglyphs of HTML; in other words, he was never destined to be a software engineer. The 2015 conclusion would be that Harry can't code. But 2025 tells us that Harry can write simple English, rely on his sharp eye, and employ GPT to handle the code side of Web Development. A smart firm will realize that a design-brain like Harry's, with some AI coding in its pocket, is a far more valuable and scarce resource than an old-school software engineer trying to intuit the principles of design for the first time. (Before the objections begin — AI is better at code than design! And what would you rather have: safe, generic, industry-standard boilerplate code, or safe, generic design?)
Harry learns to code. Not to pass an exam or to handle database IDs for a bank, but in support of his complimentary skills. Coding becomes part of his skill stack — at first, a narrow focus, close-to-home for a designer, but the confidence to read, prompt, and ultimately write some code without fear changes his life. AI un-bottlenecks him and transforms him from a "starving artist" type to an employable force.
Usha, The Kid
Usha is not unique among her friend group: she loves Pokémon and Minecraft. She loves the style and the substance. Her communities coalesce around these games. She does well in school, and in moments of inspiration, even manages to blend her interests into her schoolwork: a history project about architecture, built in Minecraft; a mathematical analysis of Pokémon's Experience System. But Usha's parents pressure her as they deal with an anxiety about the future. She's going to need a job. Dad is an accountant, and he knows his daughter would never have the patience to sift through forms for hours and hours — she was born into a different, high-speed world. Mom is a teacher and her salary even after twenty years of built-up seniority fails to keep pace with the ballooning cost of living. Their friends have a kid who went to Harvard years ago — "Take CS50 online. It's the best coding course in the world." Usha makes it 15 minutes before her eyes roll back in boredom.
"What is she going to do? She's almost fifteen." "Minecraft isn't a job." "Even if we could afford med school, she faints at the first sight of blood."
But one day, on the recommendation of a friend, Usha tries out Jippity. Her friend made a homepage for her favorite Pokémon, Kingler.
“It’s actually pretty easy.”
Usha types an innocuous prompt: "Help me make a game like Pokémon." Jippity responds with some follow-up questions. Usha answers. Next, an outline of some code. Usha clicks a button to pop it into the editor. 'Run.' A little window appears: a meager character walking around a little scene, a digital Pinocchio, all from a couple back-and-forths with the AI. Usha smiles ear to ear — I made that.
Within a few months, Usha has coded up her own hybrid of Pokémon and Minecraft, all with her own cute, naïve pixel art graphics. Her parents can't believe it — night and day from the soul-draining boredom that overtook her during the CS50 lecture. She spends hours drawing characters, writing scenarios, reading up on biomes, architecture, and design. Then she writes code as the glue that holds her project together. The AI makes coding a painless experience — as long as you can describe the problem, you can solve it — and in fact, Usha doesn't think of herself as a coder. She codes as part of her creative practice. When she lands the summer internship at Microsoft's Xbox Game Studios, her parents breathe a sigh of relief. Not only do they swell with pride — she's going to be fine — but Usha even helped them out, fearlessly, in their own careers. She threw together a Python script for Dad that cuts out hours and hours of spreadsheet-wrangling. She took inspiration from the 'Generate Questions' feature to help Mom produce individualized worksheets for all of her students with the click of a button. None of the people have "coding jobs," but thanks to playing-field-leveling of AI, they touch code as part of their jobs. Code fills in the cracks of their work.
~ ~ ~
You can imagine more tales like these. The larger story is that AI breaks down the barriers to coding. Not coding in the old sense of stomaching syntax and tricky patterns, but in the more general and more profound sense of solving a problem with a computer. It may be that the gross number of people who call themselves "software engineer" decreases, but I predict — and you should quote me — that the number of people who dip into code as part of their workflow will skyrocket across tons of domains, tons of careers. The often-snarky quip "learn to code" used to contain a hidden predicate: "so that you can get a job as a coder." Not anymore. Learn to code so that you can be a better ________ ← fill in the blank! And, let's be real: there will still be heaps of software engineering jobs in five years. Someone has to get their hands dirty. The code doesn't spring up from absolutely nowhere. But, if software engineer becomes a narrower role for those with deep interest in the inner machinations of computers, rather than just a one-way-ticket to a high salary, so be it. Good! There will be jobs in five years, and jobs benefit from programming, and programming benefits from AI. Take the plunge, whoever you are! I promise it isn't as hard as you think. "What language to start with?" — English. Get GPT to write a script that handles your monthly budgeting and fiddle around with it. Make a simple game on Jippity in an hour. Don't learn to code for its own sake. Pick it up like a new language: first as a tourist, then as a local. You have the world's best-ever guide, who only gets better every day.
A final note, aimed at parents: the AI revolution, risks though it may pose, really cracks open the door for kids. Five-year-olds are building games on Jippity. Nine-year-olds are programming multiplayer web servers. Teenagers are prepping for their AP exams. The frankly shocking success of AI coding in the hands of kids leads me to conclude: If you can type, you can code.
 | Jacob Shulman is a teacher, programmer, and musician from Los Angeles. He is the cofounder of Jippity, the only AI-powered kid-friendly coding workspace on the web. He puts out records on Endectomorph Music. |
