I’m not 100% certain I love this article. I made myself write it in less than one hour. It’s core point is supposed to be around written language being the dominant skill going forward which replaces today’s primary skill of knowing how to do stuff (visual skill). I will likely revisit this topic in the future as I think about it more.
For at least the next 5-10 years, professional work is going to become more language-based, not less. We are entering a period where being able to express yourself clearly in writing will be more critical than ever.
That runs against most of the technology narratives of the last decade. We’re supposed to be heading toward voice, visuals, and immersive interfaces. But the reality created by AI – especially LLMs – points in the opposite direction.
AI lives in a textual world. And that changes how work gets done.
AI Doesn’t “Use” Software
LLMs don’t open apps.
They don’t click buttons.
They don’t follow human workflows.
They translate language into logic. And that logic is language, itself.
When you ask an AI to do something, it doesn’t try to imitate a human sitting at a laptop. It finds the most direct, abstract path to the outcome – through code or structured instructions. It takes language and turns it into language to do something.
The response isn’t visual, it’s linguistic. And while that language may ultimately create something visual, think of a mermaid.js diagram or a chart drawn with plotly.py, it is still written language.
A Small Example With Big Implications
I’ve been using this recent – relatively trivial example – to explain a lot of my current thinking, so I might as well use it here. I recently needed to merge approximately twenty-five PDF files into a single document (I wanted it to be easier to print.) The rules were trivial:
- Same folder
- Alphabetical order
- Single output file
In a traditional workflow, this is manual work. Open a file. Save As. Insert pages. Repeat. Slow, tedious, easy to skip a file or make some other error.
Instead, I told an agent (Claude Code) what I wanted, in plain language.
It didn’t ask how to open Adobe.
It didn’t simulate mouse clicks.
I didn’t upload the files somewhere.
I opened a terminal. Typed Claude. And entered “Merge all PDFs in this directory into a single PDF in alphabetical order by filename.”
It wrote a script and ran the script. And I had a single PDF file.
Same outcome. Radically different execution.
But also a radically different work style. There was no mouse, no app knowledge, etc. There was no reason to outsource the work to another person. I didn’t need the right version or license of Adobe Acrobat. I didn’t need to know how to copy pages from one to another in Acrobat.
Once I had clarity on the instruction, it was just done. All the value was in the instruction clarity, not in knowing how to carryout the instruction.
AI Bypasses Human Process
This is the core shift most people are missing. And that many will be challenged by. So much of our work today is knowledge of “how to do something” not so much “what to do”.
Which means the skills that matter are changing.
What matters less:
- Navigating complex UIs
- Memorizing tool-specific workflows
- Being “good with software”
What matters more:
- Clear written intent
- Precise constraints
- Structured thinking
- The ability to describe outcomes unambiguously
If your job involves a computer – and most white-collar jobs do – you are now collaborating with systems that only works with language. Yes, you can give them an image here or there, but what do they do with that image: follow your written instructions.
The Skill Gap Is Linguistic
This will create a sharp divide.
High performers will:
- Understand processes in systems-thinking language
- Process their thoughts to remove ambiguity
- Turn clear thoughts into a well-defined written statements.
- Iterate on problems through language clarity
People will argue that LLMs don’t need clarity. I see this in lots of videos and hear it from many. They are correct for the basic stuff. But that doesn’t translate into complex work.
Others will keep asking:
“Why didn’t it do what I meant?”
AI will faithfully reflect the quality of the input it’s given. That feedback loop compounds fast.
The Reality
We are not moving into a post-language world. We’re moving into a hyper-linguistic one. And a mastery of language, especially system thinking language will give people superpowers. Those who lack the ability to turn their thinking into language clarity will be without that superpower.
For the foreseeable future, written language is the highest-leverage professional skill – not because writing is fashionable, but because the most powerful systems we’re building require it.
The future belongs to people who can think, reason, and work in text.
Everyone else will still be clicking – moving slower and getting less done.
Full disclosure: I used AI to help edit this article.
