Should you learn how to code in The Age of A.I.?
Or has artificial intelligence turned it into a commodity?
The other day I ran across this Tweet and it got me thinking… is it even worth it to learn how to code anymore?
A year ago, I would have said that one of the most important skills you can have is knowing how to code. You can bring your ideas to life as a founder or, if you'd rather work for an existing company, you're effectively guaranteed a well-paying job. On top of the personal security, it's a rapidly evolving industry so there's always something new to learn or play with.
But it seems that the "rapidly evolving industry" part of that is a little too true, eh?
Fast forward 10 years - heck, even 5 years - do you foresee a world in which non-technical folks won't be able to execute on their ideas with code written by an AI? It doesn’t seem like it… at least not to a crippling degree. The impact that Github’s Co-pilot or Replit’s Ghostwriter has had on the productivity of developers is nothing short of amazing. And with GPT-4, I can’t imagine the progress slowing down anytime soon.
Which begs the question...
Is it still worth the effort to learn how to code?
I have some thoughts - let's dive in.
Welcome to Making Product Sense
I’m Jacob, a product-focused founder in the trenches sharing product lessons from the best companies on the planet.
🫡 Stories from the Trenches
Last year, I went through Lambda School (now called BloomTech) and spent 6 months getting a crash course in programming. We learned HTML, CSS, Javascript, React and Node, built countless micro-projects and got a rudimentary grasp of computer science.
After graduating, I built a task management app (a programmer's right of passage), a recipe app, a photo of the day app, and a contacts app. All tiny projects, most run locally, and none with any sort of practical use case. Just projects to hack around on and practice thinking through data structures, functions, and UI.
It's a little embarrassing to admit this but... I haven't used my newfound knowledge for anything "worth it." And that's a bit disappointing. In my role as a co-founder and Head of Product, there’s a lot on my plate that isn’t coding but still, I had dreams of materially contributing to the code. Maybe that'll happen one day, I don't know. All I know is I've submitted a grand total of two pull requests, both of which were some basic HTML/CSS. Nothing to write home about.
So as I attempt to answer the question, "should you learn how to code?" please know that I'm going to be doing my darnedest to NOT use the advancements in AI as a way to justify me not writing any code. Because let's be honest... I haven't contributed any code with AI either.
We’ll circle back to my story later, but for now let's get into the arguments FOR and AGAINST learning how to code in the age of AI.
The Bull Case: I could change the world…
Sahil Lavingia, founder of Gumroad, was recently on the "Where It Happens" podcast with Greg Isenberg and shared his thesis that the people who will actually build things that matter will be the founders who know how to code. In his eyes, it’s a coder’s world and everyone else is just playing in it.
After listening to this for the first time, I made myself feel better by mentally running through a list of famous founders like Steve Jobs (Apple), Reid Hoffman (LinkedIn), Brian Chesky (AirBnB), Ben Silbermann (Pinterest) and Sean Rad (Tinder) who were not programmers, yet built culture-defining companies. Then, after I regained my composure, I had to admit... despite a short list of incredibly famous non-technical founders, the vast majority of influential tech companies were (and are) built by founders who know how to code.
Why is this? In short, leverage.
Let's say you have an idea and you have to explain it to someone who then has to figure out how to code it. Not only can the more subtle nuances of the idea be lost in translation, but the rate of building and shipping that idea will be much much slower.
But if you can build the ideas as they come, not only will you build more true to your vision, you'll out-ship the competition. And that competitive advantage is often enough to win.
Sahil hacked together the bones of Gumroad in a weekend and, despite a rocky journey, has built a service that has processed hundreds of millions of dollars in sales for creators across the world. That's the power of knowing how to code.
But even if you’re not hacking away on a project on the weekends, he argued that knowing how to code is essentially knowing how the world works which will make you better at just about everything else. His analogy is that, if you were an artist, understanding anatomy is critical if you have any hopes of becoming the best artist in the world. Because it’s about understanding the way the world works. So if you run a tech company, it’s similarly critical to understand programming because that’s how the (digital) world works.
A strong argument, no doubt. But what if it was abstracted away enough that you didn’t have to understand it? After all, modern day programmers don’t write in 1’s and 0’s. So what if future programmers don’t write in code but in chat prompts?
The Bear Case: …but then again, why bother?
In the early days of computers, only the world's brightest minds like Ada Lovelace and Alan Turing could work out the difficult math to create a computer from nothing. Fast forward a few dozen years and it still required some pretty brainy folks to develop a program in BASIC or C.
But the world of software is becoming increasingly more approachable as the complexities of code get abstracted away. What began as binary code - a bunch of 1's and 0's - slowly evolved over the last 80 years until what’s left is something easy enough for an average guy like me to learn in a 6-month crash-course. And with no-code or low-code tools like Webflow, Bubble, and Zapier, talented indie hackers could spin up a product in a weekend without knowing a lick of code.
So when I see a Tweet like the one at the top of this article where someone uses ChatGPT to help them bring a website to life in less than 2 hours using code they literally can't read, I can't help but be optimistic. I can't help but get excited for the sea of non-technical founders who will be empowered to bring their ideas to life.
I was listening to the All-In Podcast where Jason Calicanis made a sound argument for the role of “Prompt Engineer.” In the future, why wouldn’t there be someone who is skilled at talking with an AI in a way that gets the desired outputs? In a world like that… who needs programmers?
So, was it worth it?
To bring it full circle, allow me to add some color to my story. Sure, I haven’t materially contributed to the product’s codebase. But that doesn’t mean learning to code was a waste of my time.
As a founder, I have a greater appreciation for the work my team produces.
As a designer, I can build our design system in a way that works with my team instead of demanding the impossible.
As a PM, I can understand, prioritize, architect, document and communicate the features and initiatives that we need to pursue with a finer level of detail than before.
And I still have hopes of being able to contribute meaningfully someday. In fact, I’m working towards doing just that as I write this article. But until then, I know that I’m a better founder, designer and PM because of it.
So what about everyone else?
As 2023 rolls in, I'm reminded of the hard problems still left to solve in the world. Hard problems that probably won't be solved by the digital equivalent of duct tape. And because I know someone will point this out, yes there are problems that can be solved without software, but keep in mind that this newsletter is for the founders and product leaders in tech. So that’s the perspective I’m writing from.
We need ambitious founders who are willing to get in the trenches and do the work. I believe AI can help us do that better than ever before. It can point us in the right direction, help us think about a problem in a new way, and possibly even hand us a solution on a silver plater. But for now at least, I believe that learning to code is no less important than I thought it was last year.
I’ve begun to think of it like writing. Is it important to know how to write well if your job is a programmer? Well… it’s not imperative. But I know it’ll help you become a better programmer, because writing well is an exercise in thinking well. And learning how to code is much the same in other areas of life.
As we see this rising trend of small, nimble teams creating highly crafted, beautiful software, the more founders, designers, writers, PMs, researchers and sales people who know how to code the better.
What do you think? I’m really curious to know. Drop your thoughts in a comment below and let me know. Worth it? Or a waste of time?
Catch you in the next one.
—Jacob ✌️
❤️ Smash that heart!
If you enjoyed this article, smash that heart icon to show some love! 🙏
Great read. Somewhat in similar position in that I have taken a few programming classes, done small projects but never contributed anything substantial that would qualify me as an actual programmer.
I think what's key is learning to think from the computational lens. Understanding conceptually how things work under the hood (e.g. how there's a client server relationship in web apps and how certain content renders at different points of the rendering journey) provides a good translation layer between myself and my engineering team.