Why Learning to Code Remains Worthwhile in the Age of AI

I think that, even among those who have had the ability to code since before LLMs, the skill of coding has been in atrophy, and will continue to be so over the next 10 or 20 years. Learning to code, over the same period of time, will not be worthwhile. At the end of that process, cracks in the system will start to show: Namely, we will have nothing but LLM-generated code to train future LLMs on, and codebases for production code will be so low in quality that they will be unintelligible to humans, and LLMs, too, will have trouble maintaining them. At that point, coding will become a valuable skill again. The practical question for someone planning out their career from todays' point of view, is that it's uncertain when that time will come, and these markets can and will stay irrational for longer than any single individual trying to resist them will be able to stay solvent.

As a professional programmer entering the final third of an enjoyable career, I would now place "learning to code" in the same category as "making a living as a poet." As in, it's truly enjoyable art and some people appreciate it, but you'd better plan for a day job.

Senior people who already know how to code are doing OKish for now, from the data I've seen, but the job is increasingly babysitting models like they were junior contributors.

Anecdotally, the people who I know who were not particularly good developers pre-llms still manage to produce bad code even using flagship models now.

I think having solid knowledge/understanding of good architecture and general practices is still crucial, and it's easy to forget that the foundational knowledge and instinct you take for granted now actually took a lot of time and effort to learn when you were less experienced.

> Code is a beautiful form of creative expression, as rich as literature or music

Something I'm trying to do right now is to build something and avoid using LLMs to write any code. I still use it to consult. I'm writing a Dota2 tournament match aggregator in Elixir that takes tournament streams and chronologically orders them in a format that makes it easier to watch them sequentially since I find YouTube hard to use for ingesting series of videos.

I'm building it because... I like programming. I like making things. I find that LLMs are making me intellectually lazy and making things with them feels unfulfilling. I want to build. It's human to want to build.

Please correct me if I'm wrong, but aren't AI agents replacing the coding mostly being done on the outer layers of development? I mean, end user applications, apps, dashboards, business applications? On this "outer crust" people can maybe tolerate things with 99% accuracy, or bloated code. A vibe coded app can be argued to be "good enough." (Even then, look at the disaster that Microsoft apps have become post AI adoption.)

But people are still staying away from LLMs on the critical compilers, frameworks, tools and libraries that people need to really rely on. No one wants to build on code that is 99% accurate or bloated. No one wants to use an AI coded web browser. To really build good building materials, you need to code it and know what you're doing. Where is anybody even getting close to phasing out coding in those critical areas?

Does anyone here know any mathematicians who can’t do long division or basic algebra?

Probably not, because they *don’t exist because learning basic math is necessary to learn higher level math*. Whether or not we have calculators to do basic math is irrelevant if you want to become a mathematician.

I’d argue that “whether or not the average dev will be writing any code by hand in 5 years” is irrelevant to whether or not one should learn to code *if they want to master designing and building complex software* using whatever method they will be using.

Learning to code = understanding a problem, breaking it down into small, manageable pieces, putting all the pieces back together. Debugging. Iterating towards better metrics, etc.

All these are amazingly valuable skills/mindsets that can be highly portable to other "problem solving" domains.

Those are not compelling arguments.

If the best we've got for convincing people to learn to code is that it's like math notation (the most hated part of math for the uninitiated), or pretty like a violin (useless for a new grad), then coding is in serious trouble.

IMO a better argument is it helps you "think like a computer". But if you wanted to learn that there are many video games I'd recommend mastering instead of learning to code. For most people "learn to code" is like telling programmers to "learn asm".

(I've been coding ~30 years)

The arguments put forward are very weak, in my opinion. It made me depressed, not hopeful, as in: if this is all we have left, things are becoming desperate indeed.

To elaborate, if coding is like art, it's the worst art there is. It's far closer to something like legos and the satisfaction you get from completing a noteworthy build. Maybe that's the writer's point - yes it's still worthwhile, in a purely hobbyist sort of way.

I find the "coding helps math" argument to be similarly weak. Yes, it certainly helps with algebra, but overall I would argue the coding flavor of math largely helps more coding type of math. It's a strange type of math with loops and rules and conditionals that wouldn't be mainstream if software wasn't so mainstream to begin with. When you pull back the covers, the argument sounds more circular than something with true meat on the bones.