There is a strange moment that comes when you realise that a machine can do something you have spent years learning to do.

For the Textsmith, that moment may be watching an AI take a handful of rough ideas and turn them into a coherent essay.

You give it five points.

You tell it what you are trying to argue.

Perhaps you even give it a story from your own experience.

A few seconds later, there is an introduction. There are paragraphs. The argument flows. The transitions work. There is a conclusion.

You read it and think:

“That is better than what I would have written.”

And perhaps the most unsettling part is that it is not merely better than what you would have written in ten minutes.

It may be better than what you would have produced after an hour.

That is when the question becomes unavoidable:

What happens to the Textsmith when the machine can write better than the Textsmith?

The uncomfortable discovery

For a long time, writing was a scarce skill.

Having an idea was one thing. Expressing it clearly was another.

Someone might understand an issue perfectly well but struggle to explain it. They might have an excellent argument but produce awkward paragraphs. They might know exactly what they wanted to say but spend an entire evening trying to find the right words.

Writing skill therefore became a form of gatekeeping.

If you could write well, you could communicate your ideas.

If you could not, you needed an editor, a teacher, a colleague or somebody else who could help turn your thoughts into language.

AI changes this relationship.

A person can now say:

Here are the points I want to make. Put them into a coherent essay.

And the machine can do much of the difficult linguistic work.

This is enormously liberating.

It is also enormously disruptive.

What if writing was never the scarce thing?

Perhaps we have misunderstood what writing actually consists of.

We tend to think of writing as putting words together.

But words are only the visible part.

Before the words come questions.

What is worth saying?

Why does it matter?

What evidence supports it?

What assumptions am I making?

What is the strongest objection?

What experience illustrates the point?

What should the reader think about after finishing?

Those are not merely linguistic questions.

They are questions of judgment.

And this is where the Textsmith may discover an unexpected future.

Perhaps AI does not make the Textsmith obsolete.

Perhaps it changes what a Textsmith is.

The machine can write the essay. Can it decide why the essay should exist?

This distinction matters.

AI can generate arguments.

It can combine ideas in surprising ways.

It can propose metaphors, examples and interpretations.

It can produce something that genuinely makes us think.

So I would be careful about saying that AI “cannot come up with ideas.”

It can.

The more interesting question is:

Does the machine know which ideas matter to us?

Consider this article itself.

The idea began with a conversation about printing.

We were talking about why printing has survived despite decades of predictions that paper would disappear. That led to questions about technological disruption. That led to AI. Then came Gutenberg. Then the Luddites. Then universities. Then the uncomfortable possibility that some skills we currently value may become cheap.

The machine did not experience those connections in the way a human being experiences a train of thought.

It did not wake up in the morning thinking about printers.

It did not spend childhood wondering what it felt like to live during Newton’s time.

Those experiences matter because they provide the raw material from which questions emerge.

The Textsmith’s real workshop may therefore not be the keyboard.

It may be the mind.

AI is extraordinarily good at the middle of the process

This is perhaps where I find AI most interesting.

There is a process between having an idea and communicating it.

Traditionally, much of that process was painful.

You had to structure the argument.

You had to find the words.

You had to rewrite the awkward paragraph.

You had to find a better transition.

You had to decide whether the introduction was working.

You had to cut the unnecessary material.

You had to rewrite the conclusion.

AI is extraordinarily good at this middle ground.

And that may be its greatest gift to writers.

The blank page becomes less frightening.

A rough idea can become a draft.

A bad paragraph can become a good paragraph.

A complicated idea can be expressed at several levels.

A writer who knows what they want to say can get there much faster.

That is not the death of writing.

It is the automation of some of writing’s mechanical labour.

But there is a dangerous side

There is also a temptation to surrender the thinking itself.

Someone may say:

“Write an essay about artificial intelligence.”

The machine produces an essay.

The person submits it.

Technically, writing has happened.

But where was the thinking?

This is where the Textsmith philosophy becomes important.

The purpose of writing is not merely to produce text.

Text is the artefact. Thinking is the work.

That distinction becomes especially important in an age when polished prose is cheap.

A thousand mediocre ideas can now be transformed into beautiful paragraphs.

The result may sound intelligent while saying almost nothing.

We therefore face a strange reversal.

In the past, poor writing could conceal a good idea.

In the AI age, excellent writing can conceal the absence of one.

The new literacy may be knowing what to ask

This changes what it means to be good with language.

Perhaps the future Textsmith will not be judged primarily by how quickly they can produce paragraphs.

They will be judged by the quality of their questions.

Can they identify an interesting problem?

Can they notice a contradiction?

Can they connect two apparently unrelated things?

Can they recognise when an answer sounds convincing but is wrong?

Can they provide the machine with useful context?

Can they challenge the answer?

Can they revise it?

Can they say:

“No. That is elegant, but it isn’t what I mean.”

That last ability may become surprisingly valuable.

The machine can produce thousands of plausible sentences.

The human being must still decide which ones deserve to survive.

The Textsmith may become more editor than typist

This suggests an interesting future for writing.

The Textsmith of the past might have spent much of the day composing sentences.

The Textsmith of the future might spend more time directing, questioning, editing and judging.

That sounds like a small change.

It isn’t.

It moves the centre of gravity from production to judgment.

And judgment is difficult to automate because it depends upon context.

A sentence can be grammatically perfect and still be inappropriate.

An argument can be logically coherent and still miss the point.

A technically correct explanation can still be useless to its intended reader.

A beautifully written article can still have nothing interesting to say.

The Textsmith must therefore become the person who knows the difference.

This is not the first time technology has done this

There is a temptation to think that AI is unique because it is attacking intellectual work.

But technology has been doing this for centuries.

The printing press reduced the scarcity of handwritten books.

The typewriter changed clerical work.

The calculator reduced the need for mental arithmetic.

The spreadsheet transformed accounting.

Search engines reduced the time required to locate information.

Word processors transformed writing and publishing.

Desktop publishing eliminated many traditional printing workflows.

Each technology made some human skills less scarce.

And each created new possibilities.

The fact that AI can write does not make this transition fundamentally different.

What makes it particularly uncomfortable is that language has always felt unusually human.

We have attached identity to our ability to communicate through words.

When a machine produces beautiful prose, therefore, it can feel as though the machine is not merely performing a task.

It is trespassing into our territory.

What happened to the Textsmith after the printing press?

Perhaps the historical analogy that helps most is Gutenberg.

Before printing, producing books was extraordinarily labour-intensive.

Someone had to copy them.

Then printing arrived.

The people whose livelihoods depended upon copying manuscripts could not simply tell history to stop.

Their skill had become less scarce.

But humanity did not stop producing books.

Quite the opposite.

Books multiplied.

Literacy expanded.

Publishing became an industry.

New professions emerged.

New institutions developed.

New forms of writing appeared.

The technology did not eliminate the need for people who cared about words.

It changed what those people did.

Perhaps AI will do something similar.

We don’t know what comes next

This is where I become deliberately cautious.

I do not know whether my job, your job or any particular job will still exist in five years.

Anyone confidently predicting that AI will not replace jobs is making a prediction about a future they cannot possibly know.

But the opposite certainty is equally unjustified.

We do not know which occupations will disappear.

We do not know which ones will be transformed.

We do not know which new occupations will emerge.

And we certainly do not know what today’s children will call their jobs when they reach adulthood.

Someone living in Queen Victoria’s Britain could not have imagined the occupations that would become ordinary in our world.

Why should we imagine that our current list of occupations is permanent?

Perhaps we are simply another generation standing at the edge of a transformation.

So what happens to the Textsmith?

The answer may be surprisingly simple.

The Textsmith does not disappear.

The Textsmith evolves.

The old Textsmith might have been valued because they could produce polished prose.

The new Textsmith may be valued because they can recognise a worthwhile idea, develop it, challenge it, give it context, and use machines to express it more effectively.

That requires a different kind of confidence.

Not the confidence that says:

“I can write better than the machine.”

That contest may already be lost in many circumstances.

Instead:

“I know what is worth writing.”

That is a much harder claim.

And perhaps a much more important one.

The machine may force us to become better thinkers

There is an ironic possibility here.

AI may make writing easier while making thinking more important.

If everybody can produce polished prose, polished prose ceases to distinguish us.

If everybody can generate a competent report, the report itself ceases to be remarkable.

If everybody can produce ten thousand words, word count becomes almost meaningless.

What becomes scarce?

Attention.

Judgment.

Experience.

Curiosity.

Original questions.

Taste.

Wisdom.

The ability to recognise something worth saying.

Those are precisely the things a Textsmith should cultivate.

So perhaps we should not fear the machine because it can write.

We should fear becoming people who have nothing to say without it.

That is the real danger.

The Textsmith’s workshop is therefore not closing.

The tools on the bench are simply changing again.

The typewriter changed the workshop.

The word processor changed it again.

The Internet changed it again.

Markdown, Git, static-site generators and AI are changing it once more.

Our job is not to defend yesterday’s tools.

Our job is to remain curious enough to understand today’s tools and thoughtful enough to decide what to do with them.

And if AI can now write better than us?

Good.

Let it.

Perhaps that frees us to concentrate on the one thing that has always mattered more than the words themselves:

having something worth saying.