The Notebook in the Drawer

Somewhere on my computer, there’s a folder with a novella I started years ago. It begins in a remote cove in Mexico, with a woman stepping off a small plane into air so hot it feels like opening an oven door. It’s a good start (okay, I think it’s a good start). It’s also been sitting there, unfinished, longer than I’d like to admit.

I know I’m not alone. Almost every writer I coach has a folder or a drawer in a desk like mine. Maybe yours is a laptop folder named “BOOK FINAL v3,” or a stack of legal pads, or a story you’ve been telling yourself you’ll write “when things calm down.”

So when the AI conversation started heating up in our writing community, I noticed two very different reactions. Some writers were thrilled. Others were furious, scared, or convinced this was the end of real writing.

Wait, haven’t we been here before?

It turns out we have been here before, and many times. Let’s take a little trip through history, because the story of writing is really a story of new tools, panicked critics, and writers who quietly picked up the new tool and kept going anyway.

Socrates Thought Writing Would Rot Our Brains

Before ink, quills, and parchment, stories lived in people. Poets memorized epics, and knowledge passed from mouth to ear around fires and in town squares. It was an effort to pass down these human creations.

Then writing came along, and not everyone was a fan. Around 370 BCE, Plato wrote a dialogue called Phaedrus, where Socrates tells a story about the Egyptian god Theuth presenting his shiny new invention, writing, to King Thamus (Ammon). Theuth promises it will make people wiser and improve their memories. The king isn’t buying it. He warns that writing will produce forgetfulness, because people will stop exercising their memory and lean on marks on a page instead.

Socrates also grumbled that a written text can’t answer your questions. It just sits there, silent, like a painting.

Here’s the delicious irony: we only know Socrates said any of this because Plato wrote it down.

What this means for you: The very first writing technology was accused of making people lazier thinkers. Sound familiar? Every “AI will make writers stop thinking” headline is a 2,400-year-old argument wearing a new outfit.

The Abbot Who Printed His Complaint About Printing

Are you ready for one of the most delightfully ironic chapters in the history of media technology? Johannes Gutenberg built his movable-type press in Mainz, Germany, around 1450. For the first time, books didn’t have to be copied by hand, one painstaking letter at a time.

You’d think everyone would celebrate. Not Johannes Trithemius, a Benedictine abbot who in 1492 wrote a treatise called In Praise of Scribes. He argued that parchment would outlast paper, that a careful scribe was more accurate than a printer, and that copying by hand was a form of devotion no machine could replace. Some of his complaints were fair at the time. Early printed books really were full of spelling errors.

And then (I love this part) he had his treatise printed in 1494, because that was the only way enough people would read it. He also grew his monastery’s library from a few dozen books to about 2,000, many of them printed.

What this means for you: Trithemius wasn’t a villain. He cared deeply about quality and craft, the same things many writers worry about with AI today. But the press didn’t destroy careful writing. It put books into millions of hands and created whole new careers for authors.

Mark Twain and His “Curiosity-Breeding Little Joker”

In 1874, Mark Twain spotted a brand-new Remington typewriter in a Boston shop and bought it on the spot. His first typed letter, to his brother Orion, admits he wasn’t making a shining success of it yet. (It typed only capital letters, which explains a lot).

Twain had a love-hate relationship with that machine. He gave it away, more than once, and asked friends not to tell anyone he owned one, calling it a “curiosity-breeding little joker.”

But he came around. Life on the Mississippi (1883) is widely credited as the first book manuscript submitted to a publisher in typewritten form. Fun detail: Twain didn’t type it himself. A typist worked from his handwritten draft. He still delegated the tedious part, and he still wrote every word.

What this means for you: Using a machine for the mechanical work didn’t make Twain any less Twain. His voice, humor, and stories were all his. The tool just got them onto the page faster.

“The Word Processor Is Erasing Literature”

Fast-forward to 1968. Spy novelist Len Deighton (The IPCRESS File) leased an IBM Magnetic Tape Selectric Typewriter, a 200-pound beast with a tape memory unit, to help him manage the sprawling plot of his World War II novel Bomber. His assistant could pull up earlier text, cut it, insert new material, and jump between sections. Bomber, published in 1970, is generally credited as the first novel written on a word processor.

By the early 1980s, writers were buying personal computers in droves, and the backlash was loud. In 1984, Gore Vidal grumbled that the word processor was “erasing” literature. John Updike worried his machine made writing almost too easy. Some writers at the Iowa Writers’ Workshop even hid their computers, adding fake pencil marks to printouts so their work looked more “real.”

Raise your hand if you’ve ever felt weird admitting you used a grammar checker. Okay, maybe a little.

Meanwhile, genre writers, the science fiction and thriller crowd, jumped in early and happily. Douglas Adams loved his Mac. Amy Tan started a support group for frustrated Kaypro users. Within a decade, the home printer made it possible to produce a clean, professional-looking manuscript at your kitchen table instead of paying a typist.

What this means for you: Today nobody asks whether a novel was “really” written if the author used Microsoft Word. Spellcheck, cut-and-paste, and track changes are just part of writing. The fear faded, and the books got written.

Surprise: You’ve Been Using AI for Decades

Here’s something most writers don’t realize. That little red squiggle under a misspelled word? It was born in an artificial intelligence lab.

In 1961, computer scientist Les Earnest built the first spell checker at MIT, as part of a program that tried to teach a computer to read cursive handwriting. It knew about 10,000 common English words. Ten years later, in 1971, his graduate student Ralph Gorin wrote SPELL at Stanford’s Artificial Intelligence Laboratory. It was the first program that didn’t just flag a mistake but suggested the right spelling, and it spread around the world over ARPAnet, the internet’s great-grandparent, about a decade before most people owned a personal computer.

Grammar checkers came next. In the late 1970s, Bell Labs researchers Lorinda Cherry and Nina Macdonald built the Writer’s Workbench, a set of tools that caught wordy phrases, flagged split infinitives, and scored how readable your prose was. Cherry later observed that tools like these didn’t just fix errors. They changed how people wrote. Microsoft added grammar checking to Word in the 1990s, and Grammarly arrived in 2009.

Think about what that means. For more than 50 years, software has been nudging our word choices, catching our typos, and telling us when a sentence runs too long. Nobody accuses a novelist of cheating because Word underlined “recieve.” We just call it editing.

And honestly? Our writing got better for it. Fewer typos reach agents. Clunky sentences get flagged before an editor ever sees them. Writers who struggle with spelling, including people with dyslexia, get to focus on their ideas instead of their mistakes. Today’s AI tools are the same idea, just a lot more capable.

Google Already Does Your Research for You

Remember how research used to work? Type a question into Google, get ten blue links, open six tabs, skim each one, and piece together an answer yourself.

That era is quietly ending. In 2024, Google began putting AI Overviews, short AI-written summaries, above the regular search results. Then came AI Mode, a chat-style search that answers your question directly. Google says AI Mode now has more than a billion monthly users.

In other words, if you’ve Googled anything lately, an AI probably read the sources and summarized them for you. And people are clicking through far less. An independent academic study that utilized Pew Research Center data found that users clicked a regular search result in just 8% of visits when an AI summary appeared, compared with 15% without one. Links inside the AI summary itself got clicked about 1% of the time.

What this means for you: Even writers who swear they’ll never touch AI are already using it every time they search. Two lessons here. First, an AI summary is a starting point, not a source, so click through and check the original before you put a fact in your book. Second, if fewer people click links, your own website and blog may get less search traffic. That’s one more reason building a direct relationship with readers, through a newsletter or blog like this one, matters more than ever.

Novels Written With Thumbs

If you think writing a book on a phone sounds ridiculous, meet the keitai shosetsu, Japan’s cell phone novels. In the early 2000s, young writers (many of them teenage girls) began tapping out stories on their phones in short chapters, posting them to free sites where readers devoured them on crowded train commutes.

Japan’s literary establishment was not amused. Critics dismissed them as trash, but readers didn’t care. The first big hit, Deep Love, sold 2.6 million copies once it moved to print, and in 2007, five of the top ten bestselling novels in Japan started life as cell phone novels, including the top three.

Around the same time, Amazon launched the Kindle (2007), and self-publishing started shaking off its old “vanity press” reputation. Writers who’d been told “no” by gatekeepers for years suddenly had a way to reach readers directly.

What this means for you: New tools don’t just change how we write. They change who gets to be a writer. Every time, the people who benefit most are the ones who were shut out before.

AI Is the Next Tool in the Toolbox

Which brings us to now. Generative AI tools like ChatGPT, Claude, and Gemini can brainstorm, research, outline, give feedback, and yes, generate text. And the reactions sound awfully familiar: It will make writers lazy. It isn’t real writing. It will erase literature.

Here’s what’s actually happening. In a May 2025 BookBub survey of 1,229 authors, about 45% said they use generative AI in their work, while 48% don’t and don’t plan to. Of the authors using it, 81% use it for research, and drafting prose wasn’t even near the top of the list. A larger 2025 survey of 1,481 writers by Gotham Ghostwriters found 61% use AI at least sometimes, though only 42% of fiction authors do.

So most writers using AI aren’t asking a robot to write their novel. They’re using it the way Twain used his typist, and Deighton used his 200-pound IBM: to handle the heavy lifting so they can focus on the story only they can tell.

Here’s how you can do the same:

Beat the blank page. Ask AI to interview you about your book. What’s the story about? Who’s the hero? What scares them? Answering questions is far easier than staring at an empty document, and every word you type is still yours.

Untangle your outline. Paste in your messy notes and ask for help organizing them into chapters or plot beats. You decide what stays.

Research faster. Need to know what a 1940s bomber cockpit looked like, or how probate works in Colorado? AI can give you a starting point in seconds. (Always verify facts with a real source. More on that in a minute.)

Get feedback at 2 a.m. Ask where your pacing drags, whether your dialogue sounds natural, or if your opening hooks. It’s not a replacement for a human editor or critique group, but it’s a great first reader when no one else is awake.

Stay accountable. Ask AI to build you a realistic writing schedule based on your word count goal and how many hours you actually have each week. Then check in with it.

Handle the business side. Book descriptions, ad copy, keyword ideas, and launch emails are where many authors burn out. Let AI draft a few versions and then make them sound like you.

That book in the drawer? With these tools, finishing it in months instead of years is a realistic goal. Many popular AI plans cost about as much as a monthly Netflix subscription, a fraction of what one round of professional developmental editing runs.

A word of caution: AI detectors are now being used to judge writers, and they get it wrong more often than you’d think. I even had my own pre-2009 writing flagged as AI. I wrote all about it on my Substack, Mind & Manuscript. Link to the Substack detector post.

Now the Honest Part: Should We Worry?

I’d be doing you a disservice if I stopped at the cheerleading. Every one of those earlier technologies had real downsides too (the printing press spread plenty of nonsense along with the Wicked Bible). AI is no different. Here’s what to keep your eyes on.

Yes, there are abusers. Amazon has been flooded with low-effort AI books, fake “summaries” of bestsellers, and knockoffs published under real authors’ names. Amazon’s response has gotten stricter. As of September 21, 2026, KDP lets each account create only 2 new titles per format per week, and since October 5, 2026, it also limits submissions to 2 per format per week. That’s down from 10 a week earlier this year. For a real author releasing one book at a time, it barely matters. For content farms, it’s a wall.

Know Amazon’s disclosure rule. KDP draws a line between AI-assisted and AI-generated. If you wrote the text and used AI to brainstorm, edit, or check errors, you don’t have to disclose it. If AI created the actual text, images, or translation, you must check the disclosure box, even if you edited it heavily afterward. Your book description and ad copy sit outside the book, so they don’t count.

Copyright follows the human. In January 2025, the U.S. Copyright Office confirmed that copyright requires human authorship. Purely AI-generated text isn’t protected, and prompts alone aren’t enough to make you the author. Your own writing, edits, and creative choices still are. Translation: the more of you in the book, the more of the book you own.

AI makes things up. These tools can state wrong facts with total confidence. In the Gotham Ghostwriters survey, 9 out of 10 writers said they worry about these “hallucinations.” Check every date, name, and statistic before it goes in your manuscript.

The training-data question is real. Many authors are angry that their books were used to train AI without permission, and they have a point. In the biggest case so far, Bartz v. Anthropic, the company behind Claude agreed to a $1.5 billion settlement for using pirated books, about $3,000 per title for roughly half a million works. A federal judge gave it final approval in July 2026. Similar lawsuits against other AI companies are still working through the courts. It’s fair to factor this into which tools you choose.

Your voice is your edge. Readers are getting good at spotting generic, flavorless AI prose. The writers who win won’t be the ones who generate the most words. They’ll be the ones who use AI to get more of their own words finished.

What’s Coming Next

So what’s the next printing press? Here are the changes on the horizon that will matter most to writers.

Your phone may stop having apps. On The Joe Rogan Experience in late 2025, Elon Musk predicted that within five or six years, phones as we know them will be replaced by AI devices with no operating systems and no apps. Instead, an AI will anticipate what you want and deliver it. Rogan compared apps to Blockbuster Video, and Musk agreed. He also predicted that most of what people consume will be AI-generated content. He isn’t alone in the first part. In March 2026, Carl Pei, CEO of the phone maker Nothing, told a South by Southwest audience that AI “agents” will replace smartphone apps, though he cautioned it won’t happen right away.

Now, bold tech timelines have a way of running late (remember Trithemius, confidently predicting printed books would never measure up?). But the direction is worth taking seriously. If readers start asking an AI assistant “what should I read next?” instead of browsing Amazon or a bookstore, how will they find your book? The answer is the same one that’s always worked: build a real relationship with your readers through a newsletter, a website, and a community, so they come looking for you directly.

Every book, in every language, out loud. Amazon’s KDP has offered authors free AI narration (“Virtual Voice”) since late 2023, and Audible has said there were already more than 60,000 of these audiobooks by 2025. Audible has also announced AI translation tools, starting with Spanish, French, Italian, and German, including one that translates an audiobook while keeping the original narrator’s voice. Audible’s stated goal is every book in every language. For indie authors who could never afford a narrator or a translator, that’s a door swinging wide open to readers around the world. The quality still varies, and human narrators bring artistry a machine can’t match yet. Listen before you publish.

A flood of AI content, and a premium on the real thing. If Musk is even half right that most content will soon be AI-generated, then human stories, with real memories, real voices, and real stakes, become more valuable, not less. That’s why AI companies have started invisibly watermarking their text and why platforms are labeling AI content. Readers will want to know a person is on the other side of the page.

The tools are about to change faster than ever. The job, telling a true and moving story, stays exactly the same.

The Tool Was Never the Writer

Socrates feared writing would make us forget. Trithemius feared print would cheapen books. Gore Vidal feared the word processor would erase literature. Japan’s critics feared teenagers with phones.

They were all a little bit right about the risks and completely wrong about the ending. Writing preserved Socrates. Print gave us libraries in every town. Word processors helped millions of people finish manuscripts they would never have finished on a typewriter. Phones turned readers into writers.

AI is the same story with a new chapter. The quill didn’t write Hamlet, the Remington didn’t write Life on the Mississippi, and AI won’t write your book either. You will. It can just help you get there faster, cheaper, and with a lot less staring at a blinking cursor.

So go open that drawer, or that “BOOK FINAL v3” folder. Pick one small task to hand to your new assistant this week, and get back to the part only you can do.

Keep writing! The world needs your story.

Want more honest, practical talk about writing, publishing, and AI? Subscribe free to my Substack, Mind & Manuscript.