Posted on 03/18/2026 3:57:53 PM PDT by TBP
Engineering school taught us to write code. It never taught us to write. Now writing is the whole job.
I went to engineering school at the University of Virginia. I appreciated the education. The engineering program is rigorous. I learned differential equations, thermodynamics, signal processing, data structures, and enough physics to respect what I didn’t understand. (And, I barely made it through.) I now wish I had majored in English if you’d told me that thirty years ago, I would have laughed at you, and then gone back to failing an electromagnetics exam.
You know what I didn’t learn? How to write, or how to read.
I don’t mean I had zero writing assignments. I had a few. Stilted lab reports with prescribed section headings. A technical writing class that taught me how to format a memo. Maybe a humanities elective where I wrote a five-page paper that got a B because it was competent but unremarkable. That was it. Four years of education, and the total investment in teaching me to construct an argument in English was roughly equivalent to the time I spent learning one semester of chemistry, which I’ve never used since.
This isn’t a UVA problem. It’s an engineering education problem. CS programs across the country treat writing as a soft skill, an elective, something the humanities people worry about. The implicit message is clear: your job is to write code. English is for emails, and emails are overhead.
Nobody made me read literature. Nobody taught me how to build a sustained argument across ten paragraphs. Nobody explained that clarity is a structural achievement that requires crafting a narrative — that it comes from understanding your own thinking well enough to sequence it for someone else. That’s a skill. It takes practice. And we gave it approximately zero curricular weight in the discipline that now needs it most — Software Engineering.
My friend Nick majored in English literature. Same school, different universe. His program wasn’t just “read some books and write about your feelings.” It was rigorous in a way I didn’t appreciate at the time. He learned how to read — really read. How to pull apart an argument, understand its structure, identify what was being said and what was being avoided. And then he learned how to produce: how to build a sustained piece of writing that held together, that moved a reader from one idea to the next with purpose. With hindsight, it was a discipline more disciplined than Electrical Engineering.
My friend Brishen studied philosophy. He lived in the next room over. I walked into his room one day and saw more books than I had probably read in my entire life. He hadn’t read all of them, but the difference was he knew how to read them, how to extract what mattered, how to engage with an argument, how to hold competing ideas in tension without rushing to resolve them. That was the training. Not memorizing texts — learning to think through them. Meanwhile, I was across the hall studying materials science.
What Nick and Brishen were learning was how to consume and produce information. Input and output — not the kind engineers respect. They were building the skills of reading critically and writing precisely, and they were doing it every single day for four years. I was building the skill of solving problem sets with a single correct answer.
For most of my career, this was a soft disadvantage. You could get by. The engineers who wrote well stood out, but the ones who didn’t still got promoted, still shipped code, still had careers. Writing was a nice—to—have differentiator, not a requirement.
Everyone jokes that programmers are on the spectrum. I probably am too. But behind that joke is a real thing we’ve just accepted: programmers, as a group, are not great at communicating with other human beings. We’ve built an entire culture around it — the introvert engineer, the awkward standup, the developer who’d rather write a hundred lines of code than one paragraph of explanation. We treated it like a personality trait. It was actually a skills gap we never bothered to close.
That just changed. Because the new problem isn’t just that programmers can’t communicate with humans. It’s that they can’t communicate with models either. And the models are now doing the work.
The Absurd New Reality Developers are now sitting across from a model that has read — or at least ingested — a significant fraction of everything humans have ever written. Philosophy, law, literature, science, rhetoric, history. The entire corpus.
Your new AI programming partner is smart enough to understand a prompt like, “Your class architecture is Kafkaesque, but I need something more like Hemingway.” And the person on the other side of it — the developer — went through a four-year program that might not have ever asked them to read a novel.
Now, is that a problem for most developers? No. We’re at the stage where model providers are asking models to create throw-away 3D games on stage during an announcement. Still, when you start diving into product design and creating applications, you are going to want to understand more about how to communicate requirements. You might think the Kafka/Hemingway prompt from the previous paragraph was a joke, but I’m realizing these cultural shorthands are becoming part of how I communicate requirements.
The irony is brutal. We spent decades telling programmers that the code is what matters. Read the code. The code is the documentation. You are now being asked to write the requirements and context for a model.
The model has exactly what you wrote down, plus the distilled statistical knowledge of humanity encoded in a stack of unknowable multidimensional tensors executing on machines that are more expensive than you can even imagine. Your ability to provide guidance is the critical input.
If what you wrote down is a half-finished README with three bullet points and a TODO, that’s all the model knows. It will happily fill in the rest with garbage and hallucinations.
And, if you can’t write clearly and quickly, these tools are going to be very frustrating for you.
Here’s what should stop every engineer in their tracks: the model you’re talking to has the equivalent of a liberal arts education. It has read the philosophy, the literature, the rhetoric, and the history. It can engage with ideas across every discipline humans have ever organized into a curriculum. It is, by any reasonable measure, the most well-read collaborator you will ever have. And the person trying to direct it — the engineer — went through a program that actively discouraged engaging with most of those disciplines.
You need to learn how to hold up your end of the conversation, and your engineering program never taught you how.
The Liberal Arts Advantage Here’s the statement that’s going to make some people uncomfortable: I think a liberal arts major might now have a structural advantage over a computer science major in AI-assisted software work.
I have an engineering background. I think engineering training matters — the rigor, the systems thinking, the discipline of modeling real constraints. I’m not dismissing it. But I am saying that the differentiator has shifted.
The person who went to school and spent four years learning to read critically, write clearly, build arguments, and understand how ideas are structured across disciplines — that person has been training for exactly the skill that now determines whether AI produces good output or expensive garbage. The CS graduate spent four years learning to write code. The model writes code now. It writes it fast, and it writes a lot of it. What the model cannot do is decide what to build, why to build it, what the constraints are, or how to explain any of that with enough precision that the next person — human or machine — can act on it without guessing. You went to a university. Not a trade school. The point was supposed to be learning how to think, how to communicate, how to capture and interrogate ideas. If your CS program convinced you that the only output that matters is compilable, it failed you — and that failure is about to become visible in a way it never was before. The person who can write a clear, constrained, honest document is now the person who controls the quality of everything the machine produces. That’s not a soft skill. That’s the skill.
What This Means for How We Train Engineers Constructing a clear argument. Sequencing information so that a reader — human or machine — can follow it. Defining terms and being explicit about assumptions and acknowledging what you don’t know. These are the things that a good liberal arts education teaches, and a good engineering education mostly ignores.
Engineering programs are going to have to change. Not because AI demands it — though it does — but because the work has changed. The ratio of code-writing to communication has shifted, and it’s not shifting back. A developer who can think clearly, write precisely, and define constraints in a way that both humans and models can act on is now more valuable than a developer who can write clever algorithms but can’t explain what problem they solve.
This isn’t an argument against engineering rigor. It’s an argument that we defined rigor too narrowly. We treated the ability to express complex ideas in precise, structured English as someone else’s department. We built curricula that produced graduates who could implement a B-tree but couldn’t write a one-page document explaining why they chose it over the three alternatives.
We spent decades optimizing for code output. We’re entering an era where the bottleneck is the quality of human thought, expressed in human language, consumed by machines that take every word literally.
Programmers need to learn to write. Not as a soft skill. Not as a nice-to-have. As a core engineering competency, with the same seriousness we once gave to data structures and algorithms. And the schools that figure this out first are going to produce the engineers who actually know how to use the most powerful tools ever built — not because they can code better, but because they can think and communicate clearly enough to tell the machine what to build next.
Dear FRiends,
We need your continuing support to keep FR funded. Your donations are our sole source of funding. No sugar daddies, no advertisers, no paid memberships, no commercial sales, no gimmicks, no tax subsidies. No spam, no pop-ups, no ad trackers.
If you enjoy using FR and agree it's a worthwhile endeavor, please consider making a contribution today:
Click here: to donate by Credit Card
Or here: to donate by PayPal
Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794
Thank you very much and God bless you,
Jim
Uh… I’m not sure where this guy went to school but both writing and literature were required electives at my college as well as courses in technical writing.
In engineering school in the 70s and early 80s we were taught that if it was not written down it didn’t happen.
We had to write and read. It helped me a lot when I graduated and got an engineering job. Words are tools, Engineers need to know how to use them
And the BO, PR, BS, and the LMNOP degree also.
Very few (almost nobody) get(s) a job who majors in English. Most can’t write when they are done, but they can edit pretty well. The biggest point however is that they aren’t expert or knowledgeable enough to write about anything. So again, useless.
I was taught to write as an engineer. This is a fallacious and stupid article. A techie can learn to write. An English major can NOT learn to be tech savvy in the same timeframe.
Communication skills - written or oral - in the workplace are at an all time low.
Major in HVAC.
I’m not sure where this guy went to school...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
He says he attended University of VA which explains a lot. He should have attended Virginia Tech to get a quality engineering degree. He probably learned enough engineering to understand hot on the left, cold on the right and uhhh...water flows downhill.
I know so many people who can barely speak. They seem to avoid the use of all nouns and all verbs. It’s, basically, you know, the thing, and, ummmm, stuff.
Huh?
AI can outwrite any English major.
I have an engineering degree and I’ll put my writing skills up against any English major.
I got my undergraduate degree over 50 years ago.
It wasn’t in English but in Communications which required quite of few traditional English classes.
I have no idea what colleges are now offering in the way of degrees, but I’d say Communications is still a good field to be in.
I wasn't short changed in my college curriculum. It has served me well for the last 50 years. Interacting with AI prompts is no big deal. I've been able to coax the desired results from the "chat" interfaces. I still practice German daily to maintain and improve my skills. I finished the Welsh course in 2018. It has been improved and would be worth another pass. Evenings are spent learning Mandarin Chinese...just for fun.
I find it ludicrous that they actually have degrees in english or journalism.
If you do well in high school English you pretty much have learned all the English you need.
That’s simply wrong.
If you're an engineer, you just made his point since he clearly stated in the article that he went to UVA. :-)
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.