You’ve probably done it. You stared at a blinking cursor, felt the familiar resistance of a blank page, and immediately opened an AI assistant. You typed a prompt, and within seconds, a perfectly structured, aggressively mediocre paragraph appeared. You tweaked a few words, hit send, and felt a rush of productivity.
But you didn’t write anything. And worse, you didn’t think anything.
We’ve been sold a massive lie that the friction of writing is a bug to be fixed by AI. We are told that staring at a blank screen, struggling for the right word, and deleting entire paragraphs is inefficient. But that friction isn’t a flaw. Outsourcing your writing to AI doesn’t just save you time; it bypasses the exact cognitive friction required to make you smart.
When you use a large language model to write your reports, emails, or essays, you aren’t just producing mediocre text—you are actively degrading your ability to think. You are substituting passive reading comprehension for active generative synthesis.
Think about what actually happens when you prompt an AI to write for you. It generates a block of text. You read it. It sounds plausible. You nod along. But as philosopher Eric Schwitzgebel points out, there is a massive cognitive difference between nodding along while reading something and actually generating that text yourself. Once the text is on the page, it’s far too easy to passively accept the approximate word rather than fighting for the precise one.
I saw this firsthand in a corporate setting recently. A team used AI to summarize a dense, highly nuanced collective white paper. The result? They saved hours, but the subtlety of the original ideas was completely flattened. The team spent more time re-reading and correcting the AI’s vague, hard-to-notice errors than they would have spent just writing the summary themselves. The machine didn’t summarize their thoughts; it homogenized them into a flavorless paste.
A machine can generate the median output, but the median output has never changed a single mind.
And you can always tell when someone has outsourced their voice. The AI has a specific, cloying rhythm. It loves the “not this, but that” formulation. It orbits around a point without ever landing on it firmly. When you read it in your inner voice, it feels alien. It feels dead.
The fear here isn’t just that your colleagues will know you used a robot to write your memo. The fear is that you will lose your intellectual edge. The world is pushing you to use AI for productivity, but if you want to protect your cognitive agency, your professional edge, and your critical thinking skills, you have to push back.
Writing isn’t the product of thought; writing is the engine of thought.
When you struggle to find the right word, you are actually struggling to find the right idea. When you delete a clunky sentence, you are deleting a clunky concept. If you let a machine do that heavy lifting, your ideas will forever remain vague, unexamined, and comfortably average.
Stop using AI to write. Embrace the blank page. Feel the friction. It is the only thing keeping your mind your own.
FAQ
Q: What if I just use AI to write first drafts and edit them myself?
A: You're still skipping the hardest part: forcing your brain to translate abstract thought into concrete words. Editing a mediocre machine draft is passive reading, not active thinking.
Q: Doesn't AI make us faster and more productive?
A: It makes you faster at producing average text. It does not make you smarter, more persuasive, or more insightful. If your goal is intellectual output, speed is the enemy of depth.
Q: What if I'm a bad writer? Won't AI help me?
A: AI won't make you a better writer; it will make you a dependent one. The only way to improve your writing is to endure the painful, friction-filled process of doing it yourself.