The difference between trying to write and writing has never been wider.
There has always been a distinction between people who try writing and people who do it. The distinction is not about talent or success. It is about regularity and commitment. Someone who tries writing buys notebooks with good intentions. They sign up for a creative writing workshop and attend three sessions. They start a blog and post twice, then let it sit dormant for two years. They talk about the novel they want to write the way they talk about getting in shape or learning Italian. The fantasy of being a writer is more vivid than the actual work of writing.
Someone who does writing has a different relationship to it. It is not aspirational. It is not something they might do someday if inspiration strikes or circumstances align. It is something they do. They write on Tuesday mornings before work. They have opinions about what they are working on and what they have written. They've experienced the particular frustration of a sentence that refuses to behave, and they have a method (however private or strange) for wrestling it into submission. They read their own work with something like a critical eye. Writing, for them, is not a someday thing. It is a this-week thing, a this-month thing, a this-morning thing.
This distinction has existed for as long as people have wanted to write. It predates every technology that has ever promised to make writing easier. The typewriter did not eliminate it. The word processor did not eliminate it. The internet did not eliminate it. Throughout all of these revolutions in how we can produce text, the gap between trying and doing has remained. If anything, it has widened. The barrier to producing words has fallen consistently, but the barrier to developing a writing practice has barely moved.
And then something changed.
The collapse of friction
For the first time, there is now a kind of writing activity that requires no struggle at all. You can open ChatGPT, describe what you want, and receive something that reads like prose. It may be generic. It may be overwritten or underwritten. It may miss the specific thing you were trying to capture. But it is undeniably text, produced instantly, without you having to do the actual work of making language behave.
This has created a new middle ground. A person can now "try" writing in a way that was never before possible. They can prompt an AI, read the output, feel some combination of impressed and guilty and relieved, and experience something that approximates the satisfaction of having written something. They have not sat down and produced sentences. They have not wrestled with word choice or rhythm. They have not experienced the particular frustration of knowing what they want to say and being unable to find the words. But they have created text. They have a tangible output. The boundary between trying and doing has collapsed entirely.
This is different from hiring a ghostwriter, which at least requires intentional collaboration and real conversation. This is different from outsourcing to a creative agency, which at least involves a real person or team investing attention and judgment. What this is: the possibility of experiencing the fruits of writing without any of the friction of writing. You can feel like a writer without doing the work that writers do.
And yet, something strange is happening beneath this frictionless surface. The people who are actually doing something with AI are not the people who are satisfied with prompting and publishing. The people who are seriously using these tools look fundamentally different from the people who are dabbling in them.
What doers and triers actually do
Watch someone who is "trying" AI writing. They often approach it with vague hope. They ask for something general (write a social media post, write the intro to an article, write a funny email) and publish what they get, or nearly what they get. The output feels like something that happened to them, something that emerged from the machine, rather than something they made. If you ask why they phrased it that way, they often cannot say. They didn't. The AI did.
Watch someone who is "doing" AI writing, and it looks almost like a different activity. They approach the tool with specific opinions. They might start with detailed notes rather than a vague prompt. They might run three different versions and have genuine preferences about which one captures the idea better. They might use the AI to generate a first draft and then spend two hours rewriting it, changing the angle, restructuring the argument, replacing entire sections. They read it after with the same critical eye they would bring to something they had written entirely by hand. They can articulate why they made each choice. The AI was a tool in a process; it was not the whole process.
The difference is not about the AI. It is about what someone brings to it. Someone who is only trying writing with AI is doing something similar to what they would do with a blank page: hoping that the right thing will appear without them having to do the difficult work of thinking it through. The friction has simply relocated, not disappeared. It has moved from writing to thinking.
The person who is actually doing something has thought first. They know what they are trying to say. They have opinions about whether the output matches what they intended. When it does not, they know how to adjust. They are not waiting for inspiration or perfection; they are applying judgment and taste to imperfect raw material. This is not a new thing. Writers have always done it. The AI has simply changed which tools are available for generating that raw material.
The uncomfortable middle
But there is a strange middle ground that did not exist before, and it is worth examining. There are people who are now producing a stunning volume of text with the help of AI. They might be writing a newsletter every week, or a series of blog posts, or pitches to investors, or social media content. Measured by output, they are doing. They are producing more words than most writers produce. They have consistent work, regular deadlines, outputs that go into the world.
And yet, if you look at their work over time, nothing has really changed. The newsletter from week one reads the same as the newsletter from week fifty. The blog posts are indistinguishable from each other. There is no evidence of development, no sense of learning, no progression toward something. They are using AI tools to produce, but they are not developing a practice. They are not progressing as writers. They are simply generating volume.
Is this "doing"? They are certainly doing something. They are not trying in the sense of dabbling. But they are also not doing in the sense of developing a craft, building taste, making choices that reflect deepening judgment. They are busy, but there is a difference between busyness and practice. A person can run five miles every day without ever becoming a runner in the sense that matters. They can produce fifty articles without ever becoming a writer. Activity is not the same as intention.
This middle ground is invisible to the people inhabiting it. Someone producing high volume with AI tools often feels productive and legitimate. They are making things. They are shipping things. They have outputs. The fact that those outputs do not reflect growth, that they are not learning anything about the craft, is not obvious from the outside. It is a sophisticated form of trying masquerading as doing.
What separation looks like
The try/do distinction is becoming a more reliable signal than it has ever been. When anyone can instantly produce passable text, the thing that separates writers from text producers is not the ability to generate words. It is not even access to tools. It is something much older and much harder to measure: the development of taste, judgment, and intent.
People who are doing writing are the ones who have thought about what good writing is. They have read widely. They have opinions about what works and what does not, what feels true and what feels false, where a piece is working and where it is falling apart. These opinions are not innate; they are developed through reading and through the hard work of trying to produce something and seeing how it lands.
People who are trying are often the ones who have read less, who have thought less about the craft, who treat writing as something that happens when inspiration arrives or when a tool makes it easy. They are waiting for the magic to happen rather than creating the conditions for it to happen. AI has made it possible to create an enormous volume of text without ever thinking about any of this. This is not gatekeeping. It is simply observation. Writing has always been hard for the people who were actually doing it, and easy for the people who were not.
What has changed is that ease. For the first time, someone can produce professional-looking text without ever developing the foundations of a writing practice. They can skip the hard part entirely and go straight to output. This might be great for marketing. It might be fine for newsletters. But it is not the same as becoming a writer.
The question of commitment
All of this points to something larger, which is that AI has not eliminated the try/do distinction. It has made it more important. When friction was the thing that separated trying from doing (when writing a paragraph meant actually writing it), the barrier was somewhat arbitrary. AI has removed that arbitrary barrier. Now the distinction is clean. It is the difference between using a tool and having a practice. Between generating text and making choices about what you want to say. Between producing and progressing.
Someone who tries writing with AI is not bad or illegitimate. They are not a "real writer" in some gatekeeping sense. They are simply doing something different. They are using tools to produce output. This has value. It might be efficient. It might be what the situation calls for. But it is not the same as doing writing in the way that develops a person as a writer.
Doing writing requires something that tools cannot provide: attention, intention, and the willingness to sit with something until it becomes what you want it to be. It requires caring enough about what you are saying that you will rewrite it ten times. It requires reading enough to know when something is working and when it is not. It requires commitment to the thing itself, not just to having written the thing.
AI did not create this distinction. It simply made it visible. For decades, the try/do line was blurry because the friction of writing itself was doing some of the sorting. People who actually wanted to write had to want it badly enough to overcome the resistance of producing text by hand or on a keyboard. The lazy tried, the committed did. AI has removed that resistance. Now, if someone is going to actually write, it has to be because they chose to. The practice has to come from something other than the lack of alternatives.
In a world where anyone can produce text instantly, the writers are the ones who choose not to. They choose to think before writing. They choose to revise after writing. They choose to care about the specific way something is said, not just whether something is said. These choices are harder to make when they are not necessary. And that, perhaps, is the real consequence of AI writing tools. Not that writing has become possible for everyone, but that it has become possible to produce text without ever becoming a writer. The barrier has fallen, but the climb remains.
Some will take the easy path. Some will do the work. The difference has never been clearer.