What Writing Still Adds Value in the Post-LLM Era

Creating content in the era of large language models has become, for the most part, a commodity. Not all types of content, of course—but much of what’s informative in nature. If you’re trying to explain a technical concept or distill complex knowledge into a digestible format, LLMs are extremely effective. Any style, any depth, any audience—just ask, and it’s done.

So the question is: what kind of writing is still worth doing?

The answer likely lies in the difference between what humans can do, and what machines—at least for now—cannot.

Here are a few thoughts.

Original thought still matters

Most of what an LLM writes is a reassembly of existing content. It can be clever, well-written, even nuanced—but it’s largely derivative. It draws from what already exists. Truly original thinking—something that brings a new perspective or opens a door others hadn’t noticed—is still something human minds are better at producing.

That’s not to say machines can’t be creative in some sense, but it’s a different kind of creativity. They don’t originate ideas in the way a person does when connecting past experience, intuition, and a personal sense of curiosity.

Opinion is uniquely human

Today’s LLMs don’t have opinions in the way we do. They aren’t motivated by anything. They don’t have experiences, feelings, or personal history. They don’t care, and they’re not meant to. In fact, they’re explicitly trained to avoid taking sides or sounding too firm in any direction. This is partly to prevent bias and partly to avoid legal risk.

But humans do care. We take positions. We feel strongly. We change our minds. Writing with a clear, honest opinion still cuts through, precisely because it’s personal and alive. That kind of voice is something machines don’t convincingly replicate.

Vision comes from experience

LLMs can predict. They can extrapolate based on data. But a real vision of the future usually includes more than that. It includes feeling. It includes a sense of purpose, of what should happen—not just what might. When someone shares their view of what’s coming next, we’re interested not just in the logic behind it, but in the person behind the vision.

A machine can simulate an answer to “what will happen?” But “what do you see coming, and why does it matter to you?”—that’s still a question for a human.

What about factual or academic content?

There’s still room for it. But there are a few things to keep in mind.

Timeliness is one. LLMs have a delay in absorbing new knowledge. It takes time to collect, clean, and train on new data. In fast-moving fields, this gap matters. Some tools use retrieval techniques to bring in fresh information, but that doesn’t fully solve the problem.

Structure and point of view are another. Even when the facts are available, deciding how to present them—what to emphasize, what to leave out, what story to tell—requires judgment. And judgment comes from experience.

In short, writing that adds value now is writing that does more than inform. It reflects who you are. It carries your thinking, your voice, your sense of what matters and why. Machines can help you write. But they still can't be the one writing.

Not quite yet.