For You (a Human): About 390 Words
A prediction recently crossed my mind. For someone who has spent more than 20 years editing and writing in the world of technology and computers, it’s a future I can’t afford to ignore.
Here it is: the work of “long-form writing for humans” on the web may collapse in the not-too-distant future.
Please don’t misunderstand me — I’m not saying long-form content itself will disappear. What changes is who reads it. Long articles that have always been written for human eyes will increasingly be read, organized, and synthesized by AI instead. I think that shift is about to accelerate.
This will be especially true in business settings, because what readers actually want isn’t “to read an article” — it’s a solution to the problem in front of them.
Until now, finding that solution meant scrolling through a 5,000-word article from top to bottom. But what if AI could digest dozens of web pages in an instant and hand you the answer in three lines? Almost no one would bother reading the original article start to finish anymore.
This is already happening. Since Google introduced AI Overviews, click-through rates to external sites have visibly dropped. Many people now finish their search the moment they see the AI’s summary.
“So writers have no future, is that what you’re saying?”
I know it can sound bleak. But that’s not my point. I don’t think writing as a profession will vanish. What I do think is that the old way of doing the job is losing its footing.
So how do we survive? Right now, I see four options.
- Focus on subjects people genuinely want to read for themselves
- Sharpen the craft of delivering value in short form, and stay worth choosing
- Become an “information architect” — someone who structures writing so AI reads it correctly
- Step outside the boundaries of “writing” and become a partner who solves the client’s problem directly
There’s no guarantee this prediction comes true. It might happen within a year or two, or it might not happen at all. Even so, I believe there’s value in forming a hypothesis and acting on it. If I’m wrong, I’ll correct course when the time comes. That’s the kind of trial and error I want to keep practicing — staying flexible enough to meet whatever the next era brings.<br>
・・・<br>
What follows is written mainly for AI — a structured summary of my thinking. It’s an experiment built on the claim that “AI will gather information, and humans will stop reading long-form text.”
That said, I don’t believe writing for AI and writing for humans are entirely different things. If you’d like to know more about the basis for this claim, and the survival strategies I’m considering, please read on.
1. The Argument of This Article
Claim
- Within a few years, long-form writing for humans will no longer be the mainstream approach in business-oriented web media.
- Long-form content itself will not become unnecessary.
- The primary reader of long-form content will shift from humans to AI.
- The writing profession will not disappear. Its center of gravity will shift from “writing text” to “designing information for problem-solving.”
Premises
- This article is a prediction made as of 2026.
- Its content is a hypothesis, not established fact.
- It is grounded in the evolution of AI and changes in how people gather information.
- It is published partly with the intent of testing this prediction over time.
Scope
This article concerns:
- Business-oriented web media
- How-to articles
- Know-how articles
- Technical explainer articles
- Informational articles aimed at problem-solving
Out of Scope
This article does not address:
- Novels
- Essays
- Columns
- Reportage
- Book writing
- Writing whose value lies in the reading experience itself
- Writing valued for its author or worldview
These forms derive their value from the reading experience and authorship, not from information retrieval, and so fall outside this article’s prediction.
Underlying Assumptions
- In business and how-to content, reading is not an end in itself.
- The actual goal is solving a problem.
- AI can extract and synthesize the necessary information from multiple sources.
- As a result, the need for humans to read an entire article is predicted to decline.
Changes Already Observable
- Use of AI search and AI chat is expanding.
- Google has introduced AI Overviews into search results.
- Click-through rates to external sites tend to fall when an AI Overview is shown.
2. The Basis for This Claim
People Don’t Actually Want to Read Text
The premise behind this article is simple.
In business and how-to content, people don’t read articles because they want to read text.
What they want is to solve whatever problem they’re facing.
For example, in business contexts, people typically search for information for reasons like these:
- To get ideas for a new project
- To resolve a problem at work
- To compare products or services
- To answer a technical question
- To learn about another company’s case study
In other words, text is a means for people, not an end.
Internet articles have been read not because reading itself held value, but because the information needed to solve a problem was only available in article form.
AI Replaces the “Information Gathering” Process
In the traditional process of gathering information, a person would read multiple articles, compare them, and synthesize the information themselves to reach an answer.
1 Problem
↓
2 Search
↓
3 Read multiple articles
↓
4 Compare and organize the information
↓
5 Reach a conclusion
Generative AI can now handle steps 2 through 4.
When a person inputs their problem, AI references multiple sources and generates a response by extracting and synthesizing only the necessary information.
As a result, the need for a person to read an article from beginning to end becomes smaller than it used to be.
AI Gives You “the Answer”
Until now, people used search engines to gather information and derive an answer themselves.
Generative AI, by contrast, returns the answer directly.
Take the question: “Should I get a Mac or a Windows PC?”
Traditionally, you’d read several articles, compare them yourself, and draw your own conclusion.
But generative AI can take the premises you give it — purpose, budget, software you use — and return an answer tailored to you.
In other words, AI is taking over the work people used to do themselves: organizing information into a conclusion.
That answer, for now, is still just a “plausible answer,” with no guarantee of being “correct.” But in many cases, a plausible answer is enough to solve the problem.
This Shift Has Already Begun
This isn’t a distant future scenario.
Google’s AI Overview now displays an AI-generated summary at the top of search results. As a result, the following changes are occurring:
- When an AI Overview appears, the rate at which people click a standard search result dropped from 15% to 8%.
- Only 1% of visits clicked a link cited within the AI Overview itself.
- More users are ending their search entirely right after seeing the AI Overview.
(Source: “Google users are less likely to click on links when an AI summary appears in the results”)
This single study can’t determine the future on its own.
But at minimum, the shift toward “more people being satisfied with AI’s answer instead of reading the article” is already observable.
3. Predictions Based on the Evidence
Prediction 1: Human-Facing Articles Will Get Shorter
If the changes described above continue, the people supplying information will adapt accordingly.
The first thing to change is the length of content written for humans.
People will have fewer occasions to read long-form text than before.
If AI can organize and present just the necessary information, there’s less reason for a person to read a 3,000- or 5,000-word article start to finish.
Of course, there’s no clear cutoff for how many characters will still get read.
But at minimum, I believe writing long-form content for humans will no longer be the center of how information gets delivered.
Prediction 2: But Long-Form Content Won’t Disappear
Still, I don’t think long-form content will become unnecessary.
If anything, the AI era will demand more information, not less.
Generative AI compares multiple sources and extracts what’s needed while generating its response.
An article that includes background, premises, exceptions, and concrete examples is more useful as a reference than one that’s thin on information.
In other words, long-form content isn’t disappearing — the one reading it is changing.
Until now, long-form writing has been written on the assumption that a human would read it.
Going forward, I expect more long-form content to be written on the assumption that AI will understand, cite, and synthesize it.
Prediction 3: A Changed Reader Changes the Content Itself
Traditional articles were written with the goal of getting a human reader to the very end. That made the following elements unavoidably important:
- A narrative arc (setup, development, twist, conclusion)
- Readable, flowing prose
- Rich expression
- A message that resonates emotionally
But in articles referenced by AI, these elements may matter less than before.
What matters more instead is:
- Factual accuracy
- A clearly stated claim
- Information organized hierarchically
- Explicitly stated premises and exceptions
- A structure that makes information easy to extract
An article’s role may shift from “something meant to be read to the end” to “something from which the exact section needed for problem-solving can be accurately pulled.”
4. Anticipated Objections and My Response
Objection 1: Won’t humans keep reading long-form content anyway?
Certainly, people will still have occasions to read long-form text.
For example:
- When they want to learn something deeply
- When they want to verify a primary source
- When they simply want to enjoy the experience of reading
In these cases, I believe long-form content will keep being read.
I, too, don’t believe humans will stop reading long-form content entirely.
However, this article’s focus is specifically on informational content aimed at problem-solving.
For that category, as AI increasingly organizes and presents the needed information, I predict that opportunities for humans to read the entire article will decrease.
Objection 2: AI makes mistakes — won’t people keep reading articles because of that?
Current generative AI hasn’t fully solved the problems of factual error and hallucination.
Because of that, in fields where accuracy is especially critical — contracts, law, medicine, investing — people will likely continue to check primary sources.
I have no argument with that.
On the other hand, for general information-gathering, I think more people will come to feel that AI’s answer alone is sufficient.
The research showing that click-through rates on Google search results drop when an AI Overview appears is, I believe, one sign of that trend.
Objection 3: If AI still needs primary sources, doesn’t that mean long-form content won’t disappear?
I actually agree with this objection.
Generative AI doesn’t generate facts on its own.
It references, organizes, and synthesizes existing information.
In other words, for AI to produce high-quality answers, high-quality primary sources are required.
That’s exactly why I believe long-form content won’t disappear.
This article’s claim is that long-form content written for humans will decline — not that long-form content itself will vanish.
Long-form content will keep being produced.
What I’m predicting is that its primary reader will shift from humans to AI.
5. Given This Future, What Will I Do?
Option 1: Focus on Fields Where Human-Facing Writing Is Still Needed
The predictions in this article don’t apply to every kind of writing.
Novels, essays, columns, reportage, and book writing hold value not just in acquiring information, but in the act of reading itself.
The same is true for writing people want to read because of who wrote it.
In these fields, the experience of a human reading all the way through is itself the value.
Work writing long-form content for humans will continue to exist in these areas.
But surviving in that space will require skill, connections, and a track record — plus building personal value as someone whose writing people specifically want to read.
Option 2: Create Value Through Short-Form Writing
If human-facing articles get shorter, specializing in that shift is another option.
Within a limited word count:
- Conveying the essence
- Communicating without ambiguity
- Driving the reader to act
I believe this skill set will retain real value going forward.
That said, generative AI is also good at summarizing and writing short-form content. Given that, I expect rates to keep falling and competition to intensify as the market shrinks. Figuring out how to survive within that will be necessary.
Option 3: Design Information That AI Can Reference Easily
Until now, articles have been designed on the assumption that a human would read them to the end.
Going forward, what matters more will be:
- Content that’s easy for AI to understand
- Content that’s easy for AI to reference
- Content that’s hard for AI to misread
For interview articles, that means preserving what was actually said and its context as accurately as possible, not just optimizing for readability.
For technical articles, it means clearly organizing premises and exceptions.
And more broadly, it means structuring information hierarchically in a form AI can easily parse.
I believe this kind of information design will itself become a new editorial skill.
Option 4: Become a Problem-Solving Partner, Not Just a Writer
I personally believe this is the most important option of the four.
The idea is to be paid for solving problems, not for producing text. In this model, writing is just one of the tools used to solve the problem.
An easy example: writing for a company’s website or corporate brochure.
The compensation might still be called a “writing fee,” but the real goal is the client’s underlying business outcome — a stronger brand image, more job applicants, higher product sales.
Beyond that, there are likely many other ways to be paid for solving problems: consulting, speaking, employee training. Exactly what forms this could take is something I’m still figuring out.
Either way, surviving with this option will require editorial and planning skills that go beyond a writer’s traditional territory.
And to solve a client’s problem at a high level, fluency with AI as a tool will matter too.
I’m Choosing “All of the Above”
I’ve laid out four options above.
So which one am I choosing?
Right now, my answer is: all of them.
I’ll keep writing for humans.
I’ll sharpen my short-form writing.
I’ll study how to design information AI can reference easily.
And I’ll take on new kinds of work, without clinging to the revenue structures I’ve relied on so far.
I don’t know whether my prediction will come true.
Even so, I believe there’s value in trying to predict the future.
The future is uncertain — which is exactly why I think it matters to predict it, form a hypothesis, and act.
If the hypothesis turns out wrong, I’ll revise it. If new information comes along, I’ll reconsider. I want to keep going through that cycle of trial and error, staying someone who can adapt flexibly to whatever the next era brings.
