Interviews with Shawn · No. 2 · Mark Twain

Never a Better Time to Not Know What You're Doing

The interviewer is an AI playing Mark Twain. He lost more money on new technology than most men ever make, and has opinions about it. The answers are all me.

---

TWAIN: My qualifications first. I spent two years memorizing twelve hundred miles of the Mississippi, every bar, bend and snag of it, by daylight and by dark, to earn a pilot's license; proudest credential a man could hold in that country, and the war and the railroads turned it into a souvenir. Later I poured the fortune I made writing books into a typesetting machine of surpassing genius, eighteen thousand parts, every one ahead of its time. A simpler, homelier machine ate its lunch and mine with it. So when a man tells me his trade moves too fast to keep his footing, I don't laugh. I check his arithmetic.

I've seen the ledger you keep on the wall over there. It says the whole respectable world of your February, the retrieval contraptions, the prompt-engineering trade, all of it, was in the discard pile inside a year. Bold claims attract me the way a snag attracts a steamboat. So: of what you first sat down and learned, how much is already worthless? And what has held?

SHAWN: I don't know that I'd call any of the learning worthless. I got value out of it, especially since I have no software or engineering background of any sort. But things have definitely changed. When I first started, it was just a chatbot. Me asking it something, it answering, and me copying and pasting out of it into Word documents. And I'd say pretty early on, I figured out that trying to get a formal education around something like this, for someone like me with no computer science degree, probably wasn't going to be worth it.

TWAIN: That's a convenient answer, and I want you to notice that it's convenient before I let you keep it. "None of the learning was worthless" is what every man says about the years he can't get back. I said it about the Paige machine for a while, right up until I had to look at the ledger. Now, I happen to believe you, because every foot of river I memorized went useless the day the war closed the trade and yet the river-sense never left me. But you'll grant that the claim can't be checked from the outside, and a man ought to know which of his beliefs are the unfalsifiable kind. We'll come back to it.

But first, the thing you dropped at the end there and walked past like it wasn't loaded. You figured out "pretty early on" that formal education in this trade wasn't worth it for a man like you. That's backwards from human nature. A fellow with no degree, standing at the door of a technical trade, runs toward the schoolhouse hat in hand: get the certificate, get the paper, get respectable. You walked the other way.

So I want two things from you. First: what did you see that told you the schoolhouse couldn't help you? And second, and be honest, because I'll know: did you ever test that judgment? Did you ever actually sit down in one of these courses and find out? I'm told there's a story there, and I didn't come back from the grave to be denied a story.

SHAWN: There was a sitting down before the walking away. I did a decent amount of research on where I could get a certification that would be respected, something people would look at and know I had accomplished something. But it also needed to matter for the work I was doing. I'm a little too old to be collecting pieces of paper or pedigrees if they aren't going to get me closer to what I want to do.

So yes, I sat down. I took a course from IBM that was going to explain everything about AI and how best to use it. And there was a lot in it that was genuinely helpful, especially the ground-level basics of how AI works and the different kinds of AI. But beyond that, the methods it was describing I already knew were being passed over for other things. After maybe a hundred hours in the course, I dropped it, realizing my time was better spent following technical blogs and, ironically enough, asking AI to go research the topics I wanted to know more about, and learning from it how to use it better.

TWAIN: A hundred hours. You sat down before you walked away, which is more than I'd have done. But let's not build a cathedral on it. One course from one house, and you returned a verdict on schooling entire. That's a sample of one, sir. What you proved is that that map was last season's river; whether every schoolhouse sells old maps you don't know, and I notice the conclusion happened to be the one you already preferred.

Then you apprenticed yourself to the machine. In my day you learned piloting beside a man with the river in him, and he cussed you until you had it too. The cussing was not incidental. He could see what you couldn't and had no reason to flatter you. Your engine has every reason to agree with you and no eyes at all. So answer this one silently if you like: in eighteen months, who told you that you were wrong? If it was only ever the work failing, that counts, but it's a slower teacher than a man who has already run aground where you're headed.

Wider, then. It isn't only the schoolbooks that won't hold still. While you were learning this trade an entire profession was born, salaried, titled, argued over at dinner parties, and buried, in about the time it takes to break in a pair of boots. "Prompt engineer," they called it. My license needed a civil war and the railroad barons to kill it, and that took the better part of a decade of dying.

So here's my question, and I want you to answer it as the operations man, not the enthusiast: you've run warehouses and factories and companies. You've watched trades change before. In all your years of working life, have you ever seen an environment do this? And if not, what does a working man do differently when the ground itself won't hold still?

SHAWN: The prompt engineer, and I'm sure there are others I don't even know of, and ones still coming, people trying to capitalize on this new technology and make a way for themselves and their companies. But to your question, have I seen this before? I'd have to say not really. In most of my time across different industries, change was about progress. A new idea or skill set gets bolted onto an existing one and changes it a little, because of the technology you have to learn to keep doing the same old thing in a new environment. Evolution. Not an entire career type being born and killed in that short an amount of time. No, I haven't seen that. And I don't know that anyone else has, really.

TWAIN: So you've no precedent for it. Neither have I, though I'd caution us both: every generation that ever lived through a fast season swore its season was the fastest, and the graveyards are full of men who mistook their own novelty for history's. I watched the telegraph get called the end of distance. What you described, the old way, is how it's supposed to go: the trade evolves, a new skill gets bolted on, the blacksmith learns the automobile, painful but survivable, one bolt at a time. Maybe this is only that, running hot. Maybe the river didn't shift a bar or two but looked around one Tuesday and decided to be a railroad. I lean your way. I just won't have us pretending we know.

But I notice, sir, and I say this with affection, that you answered the easy half of my question and let the hard half float on by. I asked what a working man does differently when the ground won't hold still, and you gave me a very fine description of the ground. That's the enthusiast's move, and I promised the readers the operations man.

So let's have it, and I'll take it with receipts. Two parts. First: what does a man actually do differently? Not a philosophy, a practice. The fellow reading this runs a warehouse or a claims department, he's got two kids and no hundred spare hours, and he now knows the certificate won't save him. Monday morning. What does he do? And second, prove you follow your own advice: what work of your own have you already had to bury? I've heard rumors of a whole contraption of yours, a thing of eighteen thousand parts if you take my meaning, that the industry made obsolete while you were still oiling it. A man who's lost a Paige machine can spot another man who's lost one across a crowded room. Tell me about yours.

SHAWN: Different men are going to do different things. Some will carry on and hope the times don't change so fast it takes them out. My advice is to find some of those extra hours. We all get the same twenty-four every day; it's a matter of priorities. Instead of watching that favorite show of yours, take an hour and think about how you do your job. What parts of it don't really require your expertise or all your years of experience? Figure out how to get those done faster using these new tools. Once you've done that, you've given yourself more hours, and you spend them learning and staying up to speed on the changes that keep coming. So what does he do Monday? Monday morning, he goes to work and does his job. Monday night, he spends a little time planning ahead.

As for what I've already buried. Early on, I realized that what you could do with AI depended a lot more on you and the way you work than the hype let on. It was letting me write little programs, code, things I never would have attempted before. But without process, without operational structure around what I was building, mostly what it produced was frustration. So the first thing I really poured effort into was something I called IntentOS. The idea was a framework that broke work into small pieces and handed them off, each piece with one specific job. I didn't even know the word "agent" back then. I called them bots. Building it taught me everything. APIs, MCP, what GitHub, Supabase, Vercel, DigitalOcean, Cloudflare, Sentry, and PostHog even were, and how to wire them together. How to get different AIs talking to each other, and when a step needs a human before it acts. I got a crash course in the tech world, all the things real developers use to build real things. And here's the punchline. About the time I finally got it to spit out anything useful, the big frontier models started doing it on their own. Innately. What I had spent months building a machine to do, you could now just ask for.

TWAIN: Now let's compare scars honestly, because I notice a temptation in the room to make yours the bigger one, and it isn't.

My Paige machine cost a hundred and ninety thousand dollars and a lecture tour of Europe to pay for it. Yours cost evenings. That is not the same wound, and you shouldn't let anybody, least of all a dead humorist warming to his subject, charge you for a bravery you never had to buy. What I grant is the part that transferred: you built the contraption before you had a word for it, the big engines swallowed the idea and did it natively, and you came away holding the protocols, how to break work into pieces and make the pieces talk, when a machine ought to stop and ask a human. That is real, and it's more than I came away with.

Which lands us back where you started. The material rots; the discipline holds. Good line. It's also the line every man tells himself while the water comes over the gunwale, and the trouble is you can't tell the two apart from inside the boat. I told myself the Paige machine was teaching me the printing trade. It was teaching me I'm a poor judge of machines.

So let's land this boat. I bet on a machine once and machines have a way of becoming last year's machine, so I'll offer the wager on the man who keeps relearning the river, and I'll note that I have no better record at this than you do. Last word is yours, and make it short, because I've seen your ledger and brevity is not among your assets: for the fellow standing where you stood eighteen months ago, what's the one bet that pays?

SHAWN: There has never been a time in history when you could take a real run at something you weren't born into or trained for with better odds than right now. Spend a little time learning to use the tools, be willing to put in the effort to stay current, and have the perseverance to suffer the failures and keep moving forward. People would be surprised at what they can achieve right now, and it's only going to get faster.

And for those who would read that and scoff, who think "not me": at least latch onto this new open world of education. You can have AI help you learn almost anything at this point. If you don't want to do something new, then use it to get better at what you do, do.

---

The eighteen months in question

← All interviews