TL;DR: The fear that you missed the window assumes the window closed. It’s still open. AI moves fast, but the skill that matters is the judgment to use any tool well, and that judgment builds on itself from wherever you start. Authors who feel hopelessly behind are usually one honest week of effort from competent. More on how I help non-tech authors use AI. The window isn’t last year.
It’s this week, and starting today puts you on the same curve as everyone else who’s going to matter in a few years.
The feeling of having missed it
Everyone else seems years ahead. They talk about tools you haven’t heard of, in shorthand you don’t speak, and the gap looks too wide to close.
The conferences are full of people demoing workflows that read like science fiction to you. The articles assume you already know what a context window is and what an agent does. So you tell yourself the moment to get on board already passed, and you settle in to be left behind. That feeling is real. The conclusion under it is wrong.
Look closely at what makes someone seem ahead of you, and most of it is tool familiarity.Share on X
I hate what this feeling does to people.
It talks them out of something useful before they’ve even touched it. Catching up on AI works differently from most technical fields, because the skill that matters, the judgment to use any tool well, grows from any starting line. The gap is a lot smaller than it looks from where you’re standing.
What does being behind on AI measure?
Look closely at what makes someone seem ahead of you, and most of it is tool familiarity.
They know which buttons to press in this season’s app. That sounds like a lead until you remember the apps turn over every few months. The tool everyone is fluent in today will be replaced before you could have caught up to it anyway. Last spring’s hot model is this fall’s deprecated one. The workflow that won conference talks twelve months ago is the workflow nobody bothers to mention now, because there’s a better one.
That chase is a treadmill, and falling off the treadmill isn’t the same as losing the race.
The people who look like they have a permanent lead mostly restart that lead every six months, along with everyone else. They have simply gotten louder about doing it. None of that is a moat. None of it lasts. A year of obsessive tool-chasing produces a person who knows last year’s tools well and is starting from scratch on this year’s, alongside the person who started this year cold.
I find the loudest early adopters a little tiresome for this reason. Plenty of them are selling the lead they claim to have, in courses and newsletters and conference slots, and the anxiety they stir up is part of the product. Knowing where this season’s buttons are has never made anybody a better writer.
Which skill does not expire in the AI era?
Judgment: knowing what’s true, what matters and what to do next. Tools change every year; that skill doesn’t.
From the podcast
Martin Ricketts has spent thirty years predicting technology trends, and on Leaders and Their Stories he put it this way: “Being good at prompt engineering is way more important than jumping around to every different tool.” The person starting today is behind on tools and level with everyone on the skill that outlasts them.
The people who look like they have a permanent lead mostly restart that lead every six months, along with everyone else.
Underneath the tools sits the thing that matters, and it doesn’t reset every six months. The ability to ask a sharp question. The instinct to check the machine instead of trusting it, the judgment to know when an answer is confidently wrong and the discipline to verify it before publishing. Those skills carry from one tool to the next, and they compound. A person who builds them starting today passes the person who only memorized last year’s buttons, usually faster than either of them expects.
The tool tutorials skip over those skills, because they’re harder to monetize than a course on the latest app, and they’re the entire game. The professionals who get useful work out of AI, the ones whose output is worth shipping, know when the machine is lying, when its output needs a complete rewrite, and when to throw away the draft and start over. Knowing every feature has little to do with it.
That’s judgment, and judgment is built one careful interaction at a time. The hallucination survival guide walks through the verification habits that matter most, and those habits are the same regardless of which model you’re using.
Why this week is the window
The window was never a fixed date you either made or missed.
It opens the moment you start, because the judgment grows from your first hour, not from some hour two years ago you can’t get back. If you start today, you’re not permanently behind the people who started earlier. You join the same curve they’re on, gaining the part that lasts while the part they memorized expires around them.
The further back you trace this, the clearer it gets.
People who were hopelessly behind on AI in 2023 and started seriously in 2024 are now ahead of the people who chased tools obsessively in 2022 and burned out before the foundation skills set in. The pattern repeats every twelve to eighteen months, and there’s no reason to believe the next cycle breaks it. Start this week and the same dynamic works in your favor.
If you wait another year, you’ll be asking the same question next fall and getting the same answer, minus a year of practice. I’d hate that for anybody with a book in them. The tools won’t settle down while you wait, so waiting only costs you time.
My take
Forget the tools you missed. They reset every few months anyway. The skill that lasts is judgment, knowing when the machine is confidently wrong and checking it before you publish. That builds from your first hour, whenever that hour is. Start this week. Wait a year and you’ll be asking the same question next fall with a year less practice.
What should you do if you feel behind on AI?
You don’t need a course or a certificate. Take one real task you already do, a chunk of research, a messy set of notes, a rough outline, and run it through a tool this week. Watch where it helps and where it lies. Check its work by hand. Notice what you had to fix and what was useful. That single honest hour teaches you more than a month of reading about the future, and it moves you from “hopelessly behind” to “started” in an afternoon.
One real task, checked by hand, teaches more than a month of reading about AI.
I don’t care which tool you open first. Once you’ve started, every hour builds on the one before, and the fear of being behind turns into a short list of things you’ve learned to check.
If your specific worry is the book you’ve been putting off because the AI question feels too overwhelming to face first, the answer is the same. Start the book the same way: take one real chunk of material you already have, run it through whatever tool you have access to, see what happens, fix what’s wrong. The act of doing that resolves the AI question into something concrete and workable, instead of a vague cloud of inadequacy.
The cornerstone piece on whether AI can write your book covers the line between what to delegate and what to keep, and that line is where the actual work starts. It makes me angry that so much AI coverage tells capable authors they’ve already missed out. They haven’t, and the coverage saying so is selling urgency. Open a tool this week, try it on a real piece of your book, check everything it hands you, and keep writing the book.
This week is still early enough to get on the curve. The Birth of the Augmented Human is my book on what goes right when people engage with AI properly instead of handing over the thinking.
