The More You Write, the Less People Remember You

It’s 11 p.m. You’ve just finished formatting and publishing your third article of the day. You spent half an hour selecting the image, rewrote the headline five times, drafted the body with AI, and then smoothed out the word order and added two of your own examples. It looks pretty decent.

You casually check your metrics. Last month, that long-form piece you stayed up two nights for and researched dozens of articles for got just over two thousand reads. The account next door posts five times a day, each a one-click AI-generated “industry speed read,” and its latest one already has twelve thousand.

You didn’t not use AI. You did. Your output went from one piece a week to one a day, boosting efficiency several times over. The problem is, even after that boost, not a single piece is remembered. Readers scroll past, give it a like, and forget. The next time they see your name, they can’t recall what you wrote.

You start to wonder: In an era where anyone can produce ten articles a day, is there still a way to make people remember you?

Content overload has happened repeatedly over the decades

Before you panic, know that this isn’t the first time technology has created this situation.

Every time the cost of “writing” is lowered by technology, a content flood follows.

Before Gutenberg’s printing press, a book had to be copied word-for-word by monastery scribes. Copying a single Bible took months. Books were scarce; being able to write and having something printed was a status symbol in itself. Then the press arrived, and costs collapsed. In 16th-century Europe, the streets were filled with pamphlets of a few pages—religious debates, political rumors, folk remedies—you name it. For the first time, literate people found there was more to read than they could handle, and most of it was trash.

The blog era repeated this. Around 2000, building a personal website required some coding knowledge. Once tools like Blogger and WordPress became widespread, blogging became a few clicks away. Overnight, everyone was a columnist. Then came WeChat public accounts, then short video—every cycle followed the same script: barriers crumble, output explodes, and average quality plummets.

In each wave, a group of people profited first by being “more prolific and faster.” Early content farms did exactly this: hire a bunch of writers to mass-produce articles based on keywords, dominating search results through sheer volume. Before 2011, this worked extremely well on Google. Then Google rolled out an algorithm update called Panda, specifically targeting low-quality, mass-produced content. Overnight, traffic for many of these volume-driven sites evaporated.

The pattern is clear: Every time barriers drop, those who win by volume get the first taste of the spoils, but are also the first to be flattened by the next wave. AI is merely running this cycle again—only this time, the collapse is exceptionally thorough—you don’t even need to hire “writers” anymore.

When everyone passes, passing no longer counts

What sets this wave apart is precisely what AI has erased.

No matter how content proliferated before, “writing well” was still a barrier. Coherent writing, clear structure, data-backed arguments, and apt analogies—these skills were always mastered by a minority. You honed these skills for years; they were your moat.

AI has filled in that moat. Now, anyone can type a few lines into a chat box and, in three minutes, get a well-structured, complete, data-included article. Coherence is the default, structure is the default. Even techniques like “open with a scene hook” or “pair every concept with an analogy”—tricks that took a semester to teach in writing courses—are now the default.

When everyone has something, it ceases to be a differentiator. The skills you spent years perfecting haven’t worsened; they’ve become like air—everyone breathes it, and no one respects you more for being able to breathe.

The harder part is the mismatch between supply and demand. Content is exploding on one side, but the reader hasn’t changed. A person still only has 24 hours in a day, and at most, maybe three or four hours of phone scrolling—a figure that hasn’t shifted much in a decade. Yet the content flooding platforms daily is dozens of times more than a decade ago, now compounded by the tireless money-printing machine of AI. Some estimates suggest that within a year or two, machine-generated text will surpass human-written content on the internet.

Attention is a zero-sum game. When supply multiplies by dozens, the attention each piece receives is diluted by dozens. That long article you slaved over for two nights and the next-door AI’s three-minute speed read are equal in a reader’s feed—equally granted only half a second of scrolling chance.

You want to break through by “writing better.” But the dimension of “better” has been almost freely supplied to overflow by AI.

The paths everyone is trying—which one actually works?

Everyone feels like they’re drowning. Let’s look at how people are struggling to climb out and see which paths are dead ends.

Path 1: Increase output. If one piece can’t grab attention, then publish ten—surely one will hit. This is the most instinctive reaction, and the first to die. Because everyone thinks this way, everyone maxes out their output, and the water level rises together; your ten pieces are buried under another hundred thousand. Worse, platform algorithms and reader thresholds are also rising—readers are spoiled by the flood of content, their appetites becoming duller. Your desperate output is just pouring another bucket of water into a depreciating sea. That’s exactly how content farms worked themselves to death back in the day.

Path 2: Obsess over technique and formatting. Refine the headline further, polish the cover image more, tweak the pacing. This path had some merit two years ago but is mostly ineffective now, because these are precisely the things AI is best at leveling. You research a good headline formula; the next second, it’s a preset button in some tool that everyone can use. Any technique that can be systematized can be copied; what can be copied can’t give you lasting differentiation.

Path 3: Dive into niches. Don’t write generic “AI industry observations”; specialize in “AI product selection practices for cross-border e-commerce sellers.” This path genuinely works. The more specific the field, the less AI’s bland, standardized templates can suffice, because it lacks frontline dirty data and on-the-ground failure details. The problem is the ceiling and replicability. A niche is only so big; it won’t grow a massive following. And once you carve out that path, imitators quickly follow, filling in the very trenches you dug.

Path 4: Ride trends and stoke emotions. Go wherever the heat is, push the emotional intensity to the max. This brings quick returns—one viral hit today, another tomorrow. But it leaves no sediment behind. Readers remember the event, the emotion, not you. Once the trend passes, you and your account are forgotten together, and you have to start from zero chasing the next one. This is rented attention, not your own.

After eliminating these four, what remains is the truly scarce type. They share a common trait: AI can’t write them, and others can’t steal them.

It’s your firsthand experience—that project you personally messed up, how much money you lost, how you misjudged at the time, how you tried to fix it later. These details exist only in your mind. It’s the bold judgments you make and stand behind—when you placed a clear bet on something before it was settled, whether right or wrong, that carries more weight than a hundred “balanced analyses.” It’s your stance and your temperament—why you dislike a certain practice, whom you’ve argued with, the biases in your heart that others dare not voice.

These are things AI cannot provide, because it doesn’t have the brain that tossed and turned last night worrying, or the hand that sweated when making a decision. It can mimic your sentence structure, but not your experience.

The logic is sound, but the platform may not agree

“Just be yourself, write what others can’t” sounds perfectly right. But actually following that advice, you’ll immediately hit a few walls.

The first wall is the algorithm. What do recommendation systems favor? They favor content that can be consumed quickly and sparks immediate interaction. A clear, slightly contrarian but safe viewpoint is more likely to be pushed than your honest but convoluted thinking process that requires patience from the reader. The more you write something authentic and imperfect, the more likely the algorithm will judge it as “low read-through rate” in the first distribution wave, and it won’t even get a chance to reach people.

The second wall is money. Scarce things accumulate slowly. Building a reputation through firsthand experience and honest judgment is a process measured in years. For the first year or two, your metrics will likely look worse than those of the account next door that chases trends and mass-produces daily. You have to endure that period of “doing the right thing, but the numbers look awful.” Many people die here; they can’t hold on and go back to competing on volume.

The third wall is losing followers. Having a stance means you will offend some people. You clearly oppose something that’s been hyped up; those who agree will become more loyal, but another group will immediately unfollow. Compared to the safe path of offending no one and holding steady read counts, this path’s data will be volatile.

So this isn’t a problem of “just figure it out and execute easily.” The algorithm’s short-term preference and your long-term value are structurally at odds. Recognizing this conflict is better than pretending it doesn’t exist—at least you know what your ugly metrics are paying for, instead of assuming you simply did something wrong.

So what should you, staring at the analytics, actually do?

Return to you at 11 p.m., facing the hollow sense of accomplishment after publishing three pieces.

There’s one line that can help you separate what to hand over to AI and what you must bear yourself: For any task where the marginal cost can be driven to zero, give it all to AI. For the part that is non-replicable, grind it out yourself.

Formatting, research, polishing a verbose paragraph, translating English materials, organizing a rough draft’s structure—hand these over to AI without guilt; it does them fast and without complaint. The time saved should be poured entirely into what machines cannot do.

What machines can’t do amounts to just a few things, but each requires you to show up personally.

Write about what you’re genuinely unsure of. Not pretending to be unsure only to circle back to a pre-conceived conclusion, but genuinely still puzzling it out. Your question tonight—“Why is no one remembering me even though my output doubled?”—if you honestly write out that struggle, including how you doubted, to what extent, whether you tried to find counterarguments, that article will make people remember you more than any “in-depth analysis” you assemble with AI. Because it’s a living person’s real state on a specific night; AI can’t fabricate that.

Write the process, not just the conclusion. AI writes conclusions more neatly than you. But “I first tried it this way, found it didn’t work because I fell into that pit, then switched to another approach, which spawned new problems, and finally scraped by” — only you can write something like that. Even you will learn from it if you reread it three years later.

Make clear judgments, and dare to put your name on them. Ambiguity is easy for anyone, AI especially. But placing a clear bet on something before it’s settled, owning the outcome right or wrong—that is scarce. Because it carries risk, and AI will never take risks.

Stop asking “How can I write to make everyone like it?” That’s a remnant of scaled thinking, pushing you to grind yourself into the greatest common divisor, until you and AI are indistinguishable. Pick a small group you truly want to speak to, and write for them.

Tomorrow you’ll probably still use AI, still post those speed reads. That’s fine; putting food on the table matters. But every so often, turn off all the tools, open a blank document, and write something only you can write, something AI can’t touch.

That piece will likely have terrible metrics, and few will like it.

But a year from now, if anyone still remembers you, it will be because of that piece, not the three you published tonight.