(Apologies in advance. IÂ know blog content has a bad rep for telling you the author's whole college study abroad experience before getting to the freaking pasta recipe, but I've got something to say. If you just want the numbers, scroll to the charts. But here it is.)
‍
Despite being a lifelong literature nerd, I started college in engineering. Why? I was absolutely 100% you-could-not-have-changed-my-mind-if-you-tried certain that I wanted to study chemistry. I loved it outright, the way some people love springtime or violin concertos or sitting by a campfire. Not just as a field of study, but as a framework for understanding everything. Everything. All the beauty in the world, every emotion you've ever felt, it all comes back to chemistry.
I used to tell anyone who would listen -- or rightfully grill me about someone who had been saying her favorite author is Edgar Allan Poe since third grade wanted to go into STEMÂ of all things when literature was right there -- that chemistry was remarkable because every beautiful thing in the universe is made of the same 118 elements. Every star, every river, every person, every piece of art. And if you break chemistry down far enough it becomes physics, and if you break that down it becomes math, and if you break that down it becomes philosophy -- our tiny, human attempt to quantify and understand the edges of what's knowable. It captured my imagination completely.
And then they put me in an intro computing class and made me put HTML inside angle brackets, and so began a love affair with coding that eclipsed my big feelings for chemistry entirely.
I was making templates before the assignment was due. Writing JavaScript beyond what was in the textbook. Practically doing my classmates' assignments just to have more code to write. Java, C++, assembly, I loved it. The logic. The structure. The way all of the pieces fit together just right, like human language, but more precise. With a cleaner framework to build off of.
I didn't finish the CS major.
Not because I lost interest. Because the environment made it very hard to stay. Classes of 50+ students where maybe two or three of us were women. Group projects I completed myself because my peers outright told me "no" when I asked to work with them, or (in a misguided attempt at chivalry?) offered to do all the work so I didn't have to worry about it -- as if I weren't shelling out thousands of dollars in tuition to work alongside them, to learn the same things. I had professors who were brilliant, but whose expertise I couldn't access because it was so far abstracted from the level of knowledge I was starting from. The constant, low-level feeling of being behind classmates who had been tearing apart and assembling computers for 10+ years when I could not have told you Donald Knuth from Donald Duck a mere 10 months before. If these were my peers, the people I'd be heading off against for jobs, how could I ever hope to land something? How would I pay off my student loans? How could I expect to fill in the gaps in my knowledge, my network, my path forward?
So I switched to English. Kept the CS minor, kept the fundamental coding skills.
And then, without ever going back for the degree, I have gone on to manage databases, host websites, become the de-facto technical person in every group I've worked in. Responsible for the technical implementation on our clients' websites -- schema, canonicals, HTML embeds, the whole stack. I interviewed at Google and had an engineer tell me "wow" about my analysis of a regex algorithm. That the solution IÂ mapped out during the code-writing part of my interview was "elegant." I didn't get the job. But I got a lot of other things.
I know it's a long, winding story, and maybe I come off as a little wishy washy -- but I would argue that sorting out your professional ambitions is exactly what the college years are for. I tell you this story because it's my honest answer to the question people are actually asking when they wonder whether learning to code is worth it. It's not about learning to code, really. It's about the career path. What does learning to code get you? Is there a door still open, and if so... where do you find it?
So is it still worth it, now that AI writes code?
The honest answer is: it depends on what you mean by "learning to code" and what you're trying to accomplish.
AI has changed this calculation in real ways. GitHub Copilot, ChatGPT, Claude, and a dozen other tools can now write functional code from a plain-English description. They can debug, refactor, explain, and suggest. A developer with AI tools in 2026 is dramatically more productive than a developer without them five years ago. And a non-developer with AI tools can now build things that would have required hiring someone a few years back.
But "AI can do this" is not the same as "humans no longer need to understand it." And here's why.
What about just using AI to write the code for you?
This is what people are calling "vibe coding" -- describing what you want in plain English and letting the AI generate it. And honestly? For a lot of use cases, it works well enough.
Vibe coding can build low-res prototypes, draft schema markup, generate HTML tables, write tedious but basic scripts that would have taken an hour to do manually. The output is often good. Sometimes it's even excellent. But not for everything.
Here's where it breaks down: anything that needs to live in the real world for more than a week.
The code AI generates is often technically functional but not particularly good code. It's not thinking about maintainability, scalability, or the five weird edge cases that will emerge the second that real users get their hands it. It produces working solutions the way a non-native speaker can produce grammatically correct sentences in their second language: technically right, but missing the instincts that come from really being fluent. When something breaks -- and, inevitably, something will break -- you need someone who can read the code, understand what went wrong, and fix it. If you vibe-coded your way through the build, that person often can't be you.
For proof-of-concept projects, solo tools, quick automations, and things like drafting schema or building a one-off template? Vibe coding is great. It's democratized coding in a way that actually makes it possible for people without a technical background to stand up their ideas, demonstrate their unique perspective, and join the conversation. Tech needs that. But as the cornerstone of a product that will have to scale, be maintained by multiple people, and handle the complexity of a real-world rollout? It's a foundation with cracks in it.
The more useful question than "should I learn to code?" might be: "Do I need to understand what code is doing, or do I just need the output itself?" For a lot of people and a lot of jobs, you do need to understand it. And AI tools are actually better used in the hands of people who do.
The door to tech is not the single, shining entrance to Silicon Valley you think it is.
Here's the thing I wish someone had told me when I left the CS major: technical skills themselves are way more portable and more learnable on the job than the degree implies. What you actually need to get into -- and stay in -- most technical roles isn't a CS degree or even mastery of a particular language. It's familiarity with technical concepts, the ability to learn new tools quickly, and a portfolio that shows you can make things work.
For some jobs that absolutely isn't true. Jobs like backend engineering, systems-level programming, anything that requires deep computer science fundamentals -- those doors require the full thing. The degrees and certifications are worth it.
But for a surprising number of technical and technical-adjacent roles like front-end development, technical marketing, data analysis, product management, developer relations, UX engineering, technical writing -- you can carve out a real position in tech with conceptual literacy, some project-based learning, and the willingness to keep learning. I've watched this happen countless times working in a startup environment. Some version of this is the story of my whole career, start a new project, rinse, and repeat.
Perhaps surprisingly, the AI era actually makes it more true than ever, not less. Because now the question isn't just "can you write code" -- it's "can you work with evolving systems? Ones that include AI tools? Can you understand what they're doing, and know when to trust them and when not to?" Which is a totally different skill. It's adaptability and good judgement, not rote technical ability.
Using AI works to learn things faster, but not as a shortcut to avoid putting in the work.
The best use of AI coding tools if you're learning is not to have them write the code for you. It's to use them the way you'd use a knowledgeable colleague sitting next to you -- someone who can explain what a piece of code does, suggest why something isn't working, show you an alternative approach, and answer the "why does this work this way" questions that tutorials often skip.
And maybe, on occasion, help you find that missing semi-colon you've been combing the program for 45 minutes to find.
I've gone back to coding concepts multiple times over the years using free resources -- you remember how much I liked it, after all. And I was surprised to find that the gap between wanting to understand and actually understanding those concepts was way shorter than I expected. The fundamentals of web development have not changed that much. HTML is still HTML. JavaScript is still JavaScript at its core. The ecosystem around them shifts constantly, but the underlying logic is logic.
Which means if you want a way into coding, or even just understanding the basic concepts of coding and computer science, don't be afraid to start with the path of least resistance (and a way smaller price tag than a college degree):
- freeCodeCamp: Genuinely excellent, easy-to-follow lessons for total beginners and up. Covers everything from responsive design to python to relational databases and gives you a clear path through, plus community and opportunity to build your portfolio, too. We stan.
- HackerRank: Interview prep + coding puzzles to keep you sharp. It's like the NYT crossword for tech nerds.
- MDN: Trusted coding resource that's been serving the community for 20+ years.
- Coursera: If you want a certification to dip your toes in, Coursera Plus is like $35/month and lets you sign up for unlimited courses
- EdX: Hosts Coursera-style online classes from MIT, Harvard, and other top universities, available to you for free. Including CS50, Harvard's legendary intro to computer science, which again, is totally free to audit. ‍
- Khan Academy: It's not just for passing algebra anymore. Khan Academy has dedicated courses on electrical hardware, CS theory, specific languages, and more. Quizzes and interactives help you test what you know. That's all free. But they also rolled out AI study companion Khanmigo for $4/month if you want an AI companion specifically designed to help you learn. ‍
- Youtube:Â Honestly, the gold standard. If there's something you need to learn how to do, there's almost certainly someone on Youtube who can teach you. And their office hours are wayyyyy more flexible than your average TA's.
So what AI has really changed is the pace at which you can test your understanding. You can describe a problem, see a solution, try to replicate it yourself, and get feedback in minutes. That's a better, more focused learning environment than you'll find in most classrooms.Â
Regardless of how you feel about AI -- including the environmental and ethical implications that I for one am extremely concerned about -- AI is a part of the landscape now, particularly if you want to be involved in tech, and it doesn't seem to be going anywhere. But I'm not telling you to just deal with it. I won't give you the advice a hiring manager friend recently gave me ("AI may not take your job, but someone who is willing to learn AI sure will"). But I will point out that if AI truly is a fixture in our world now, the only way to get a handle on any of those concerns is to have smart, empathetic, ethical people who know how to code. People who can weigh in on regulations. People who can optimize the processes that AI utilizes and put necessary constraints in place. And if AI-assisted learning helps those people get up to speed with the breakneck pace at which AI has been evolving? Good.
The proper regulations and safety precautions are not in place yet, but the people who will establish them are almost certainly going to use AI to get to the point where they can make a meaningful difference.
What I'd Tell Myself Going Back
I don't regret switching majors. The inhospitable environment in that CS department genuinely was affecting my ability to learn what I was paying to learn, and I think staying in it would have cost me more than the degree was worth. But I also think I undersold myself out of that space. I assumed walking through the door to a tech career required credentials I didn't have, when in practice it only required skills I had or could build.
So if you're sitting there now, wondering if tech is accessible to you, if it's too late for you to start, if AIÂ has closed the door before you reached the handle -- the answer is no. The path to that door and the path on the other side of the door look different than they did a few years ago, but it's open. The question is more if it's the one you're actually trying to walk through.
The person who wants to build a product alone in a weekend? AI tools plus some basic coding skill is probably enough.
The person who wants a career in engineering and is willing to put in the real work? Actually learning to code still matters, and the investment in the degree still pays.
The person who's in marketing or design or product and wants to stop being the non-technical person in every technical conversation? You don't need a CS degree. You need enough working knowledge to hold your ground, and that's more doable than you think.
Good luck out there. And (I hope) welcome to the knowing-coding-basics-club. Until next time :)
