Simplify Everything, They Said: The Cognitive Load You Just Handed Your Reader
Here is the advice you have heard a hundred times this year. Answer the question sooner. Write less. Strip out the fluff. Make it simple enough for a machine to lift and a skimmer to grab. It sounds obviously correct, the way most tidy advice does, and that is exactly why I want to poke it.
Because there is a quiet assumption buried underneath all of it. The assumption is that simplification removes work. That when you shorten a paragraph or lead with the answer, the thinking that used to live in those words just evaporates. Gone. Everybody wins.
It does not evaporate. It moves.
Simplification does not destroy cognitive load. It relocates it, and usually onto someone with less context than you.
I did not invent this idea. Duane Forrester made the point sharply in a recent piece, and it stuck with me because I have watched it play out in client content for years without having a clean name for it. AI search did not make understanding free. It changed the address where the hard part lives.
What we actually mean by cognitive load
Cognitive load is just the mental effort it takes to get from confused to clear on something. Reading a dense contract is high load. Reading a birthday card is low load. Simple enough.
Now here is the part people skip. The total load in any situation is not fixed to the page. It is spread across a whole chain. The writer carries some. The reader carries some. And increasingly, the machine sitting between them carries some too. When you make one part lighter, you are not shrinking the pie. You are handing a slice to someone else at the table.
So when a marketer tells you to "just simplify," the honest follow up question is the one nobody asks. Simplify for whom? Because the work of figuring out nuance, caveats, edge cases, and "well, it depends" does not disappear when you delete the sentence that explained it. The reader still hits the edge case in real life. They just hit it alone now, without the sentence you cut.
The answer box is a great place to hide the truth
Answer sooner. I get why this became gospel. AI systems and impatient humans both reward the page that says the thing up top instead of burying it under a founder's origin story and a recipe blog's worth of preamble. Fine. Leading with the answer is genuinely good manners.
But watch what happens when "answer sooner" hardens into "answer only."
You get content that gives a confident, clean, quotable line and quietly amputates every condition that made the answer true. "Should you run ads on this channel? Yes." Great pull quote. Terrible advice, because the real answer was "yes, if your margins can absorb the learning phase and you have creative to feed it, otherwise no." The conditions were the actual value. The bare yes is a coin flip wearing a suit.
The reader who takes the bare yes and runs still has to discover the conditions. They will discover them the expensive way, in the market, with their own money. You did not save them cognitive load. You deferred it, added interest, and mailed them the bill later.
A clean answer that drops the conditions is not a shortcut. It is a trap with good posture.
Write less is not the same as say less
This is where a lot of smart people trip. They hear "write less" and start cutting words, when the real skill is cutting words without cutting meaning. Those are completely different jobs.
Compression, done right, is one of the hardest things in writing. You are taking the same amount of understanding and fitting it into a tighter container, which means every remaining word has to work harder. That is craft. That takes more time, not less.
Compression done lazily is just deletion with better branding. You hit a word count, the page looks breezy, and the meaning quietly leaked out the bottom while you were admiring the white space.
Here is a test I use. After you cut something, ask whether a reader could now make a worse decision because the thing you removed is gone. If yes, you did not simplify. You misled them, politely. A hammer is not a house, and a shorter article is not automatically a clearer one. Length was never the problem. Effort per unit of understanding was.
AI didn't remove the middleman, it became one
The reason this matters more now than it did five years ago is that there is a new player absorbing load, and it is not neutral about it.
When someone asks ChatGPT or an AI overview a question, the machine reads a pile of sources, compresses them, and hands back a smooth paragraph. Feels like magic. Feels like the load vanished. It did not. The machine took it on, and the machine makes choices about what to keep and what to drop, and you do not get to see those choices.
So think about what you are optimizing for when you write "for AI." You are writing content designed to be easy for a system to lift a clean claim out of. Which means you are actively encouraging the removal of your own nuance, because nuance is exactly the stuff that does not compress into a tidy sentence. The caveats are the first thing to go.
Read that again, because it is the whole game. The features of your content that protect the reader from a bad decision are the same features the extraction layer is most likely to strip. If your strategy is to make yourself maximally quotable, you may be volunteering to be quoted without your conditions attached.
So who should be carrying the load?
This is the actual strategy question, and almost nobody frames it this way. Every piece of content is a decision about where to park the hard part. You have roughly three options.
- Park it on the writer. You do the deep work, and you present a genuinely digested answer where the nuance is baked into the recommendation rather than deleted from it. Expensive to produce. Best for the reader.
- Park it on the reader. You give them the raw, complete, undigested picture and let them sort it. Cheap for you, exhausting for them, and increasingly ignored because who has the time.
- Park it on the machine. You write for extraction and hope the AI layer summarizes you faithfully. Cheapest of all, and the one where you have the least control over what survives.
Most "simplify everything" advice is quietly pushing you toward option three while dressed up as option one. It looks like you did the digesting. You did not. You just made the machine's job easier and called it customer service.
The best content I have made with clients, the stuff that actually built trust and got recommended, almost always chose option one on purpose. It was short where short was honest and long where the truth needed room. That matters.
The sharper test
Alright, enough diagnosis. Here is the test I would replace "is this simple enough" with, because "simple enough" is the question that got us into this.
Ask: after reading this, is the reader closer to a good decision, or just closer to feeling like they made one?
Those two things feel identical in the moment and could not be more different in outcome. Feeling informed is a sensation. Being informed changes what you do next. A lot of AI friendly content is optimized for the sensation because the sensation is what gets the click, the nod, the "great, thanks." The actual decision quality shows up weeks later, off camera, where your analytics cannot see it and your case study will never mention it.
Ask whether your reader is closer to a good decision or just closer to feeling like they made one. Those are not the same page.
If you want a version you can run on a single sentence, use this one. For every claim you simplified, could you name the reader who is now going to get hurt by the missing caveat? If you can picture that person clearly, put the caveat back. It was load bearing.
What this means for how you write now
I am not telling you to go back to bloated, meandering pages that make people scroll through your feelings to find a phone number. That content deserved to die and I will not mourn it. Leading with the answer is right. Respecting people's time is right. Cutting genuine filler is always right.
What I am telling you is that "simple" is a means, not a goal, and we have started worshiping the means. The goal was always understanding transferred cleanly from someone who did the work to someone who needs it. Simplicity serves that goal right up until the moment it starts eating it.
So the practical move is unglamorous. Before you cut, know why the words were there. Before you compress, make sure you are keeping the meaning and losing the packaging, not the reverse. And before you write something purely so a machine will quote it, decide whether you are comfortable being quoted with the conditions removed, because that is the deal you are signing.
The teams that will win the AI search era are not the ones who simplified the hardest. They are the ones who understood their subject deeply enough to know which complexity was load bearing and which was decoration. That is a knowledge problem, not a formatting problem. No tool solves it for you.
AI search did not lift the burden of thinking off the internet. Nothing does that. It just built a new place to hide it, one layer further from the person who has to live with the answer. Your job did not get easier. It got sneakier. The work is still there, sitting in the room, waiting to see who blinks first.
Make sure it is not your reader.