Better Times Are Ahead. Bring Boots.

TL;DR — Every dark stretch in history ended because the people living through it built the way out, not because time passed. The world is shaped less by the powerful than by what the rest of us accept without reading — and today that includes answers from AI, pasted as-is. Better times are ahead. They just won't be easy, and they will be muddy, like every spring.

There's a line on my About page that says I believe better times are ahead. I wrote it a couple of years ago. Since then I've re-read it more than once with the news open in the next tab, and each time it sounds a little more like a joke I'm telling myself.

I'm keeping the line. Not out of stubbornness — because when I stop reading the news and start reading history, the joke is on the news.

Three cliches, and why I still buy them

You've heard all three. They get printed on posters and posted under bad days.

Hard times make strong people, good times make weak people. The funny part is that this one isn't ancient wisdom at all — it's from a post-apocalyptic novel published in 2016, and it went viral because it sounded old. That doesn't make it wrong. Everyone I know who is genuinely hard to break got that way in a period they would never volunteer to repeat. We are in one of those periods now. Whatever else it's doing, it's forging.

Even the darkest night ends and the morning comes. This one is at least old: a theologian wrote it down in 1650, and he was probably repeating something older. It's not a promise that the night is short. It's a statement that night is not the permanent state of the sky, and that somewhere out there the switch exists, whether or not you can see who's holding it.

Things have to go bad before they can go well again. That's just the shape of a cycle, and cycles are the one thing history is reliably full of. The question is never whether there's a way out. The question is where in the cycle you are and what you do while you're there.

What the past actually says

I don't want to name anything from today here. Not a country, not a leader, not a company. Partly because I don't want to make anyone angry, and partly because it's beside the point: the pattern is older than all of them.

In the fourteenth century a plague killed somewhere between a third and half of the people on the continent. It is hard to imagine a darker night. And yet the survivors found themselves, for the first time in their lives, scarce — and scarce things get paid. Wages doubled and tripled. The rulers of the day passed a law freezing pay at pre-plague levels, and it was ignored so thoroughly that it's mostly remembered as an example of a law that didn't work. The worst thing that ever happened to those people was followed by the best bargaining position their class had ever had.

Five hundred years later, weavers looked at the new machines that made their skills cheap and started smashing them. The state's answer was to make machine-breaking a hanging offence; seventeen men were executed in one session. The looms stayed, and the mills that grew up around them kept hiring for the rest of the century — just not at hand looms, and not the same people.

Around 1900, four in ten workers in one large industrial country grew food. Today it is under two in a hundred, and the country grows more food than ever. Nobody planned the other thirty-eight into new jobs. The jobs came anyway, most of them in categories that didn't have a name yet.

"Computer" used to be a job title. Rooms full of people doing arithmetic by hand, and doing it well enough to put rockets into orbit. The machine took the name and the arithmetic. It didn't take the people who understood what the numbers meant; it handed them better tools.

And the last century's worst stretch — two world wars with a depression between them — was followed, within twenty years, by the richest and most peaceful period the same places had ever known. Not because anyone waited it out. Because the people who came through it rebuilt, and had learned exactly what not to build again.

Every one of those dark stretches ended, and if you look at how, the same detail shows up each time: somebody was building while it was still dark.

The world does not change by itself

That's the part the posters leave out. "Better times are ahead" is not a weather forecast, because times don't get better by themselves. People make them better, or worse, and everything we have right now — every rule, every price, every tool on your desk — is there because people chose it.

We like to think the choosing is done by the powerful few, and sure, they get the biggest votes. But most of what shapes a society isn't decided by anyone important. It's decided by the masses, quietly, through what they accept without reading. The contract signed without opening it. The headline forwarded without clicking it. The default kept because changing it takes a minute. Nobody decided that; everyone just applied things as they came.

This is a blog about AI, so let me say where I see that same reflex today: in answers pasted straight from a model into a place where they matter.

Lawyers have already been fined for filing court briefs that cited cases which never existed — the model made them up, and nobody checked. A consulting firm refunded part of a government contract after its report turned out to quote a judge on words the judge never said and cite papers nobody had written. And that's just the cases that made it into the news because someone downstream did read. Most of the unread output doesn't make the news. It just becomes the codebase, the policy, the plan.

None of that is the model's fault, any more than the loom was at fault. The tool did what it does. The failure was a human deciding that reading is optional. I've written before that code you can't read is code you don't own. Scale that up. A society that applies answers it doesn't read is a society that doesn't own its decisions — and then it wonders who's making them.

The good news hides in the same sentence. If the world is shaped by what people accept without thinking, then thinking is not a passive act. Reading the thing before you apply it is the smallest possible vote, and it's the one that's always available.

The news is not the weather

About that next tab. Most media doesn't sell truth; it sells content, and bad news is the content that sells best — measurably, down to each extra negative word in a headline. Bad things do happen all the time, the way weather does. That's not a reason to watch every cloud, and it's definitely not a reason to read the forecasts piled on top, which cost nothing to make and nothing to get wrong, because nobody ever checks. Journalism is about as broken as the world it reports on - the industry, not the people in it, who mostly do honest work under bad incentives.

The odd part is that the same tool flooding the feed can drain it. A model that reads a hundred sources in a minute will tell you what actually happened, what is just guessing, and who said the same thing last year and was wrong. That's better filtering and better forecasting than any pundit, and it sits on your desk.

And after the filtering, you still have your own brain. Spend it on what this means for you and what to do tomorrow, not on how to feel tonight. The news is built for the second question, and the second question never built anything.

Spring is muddy

I want to be careful with "better", because it's easy to hear it as "good".

It doesn't work like that. It has never worked like that. When winter ends, the first thing that happens is not flowers — it's mud. Snow melts and the roads turn into everything the snow was hiding, and for a few weeks it's honestly uglier than the winter was. Nobody sane looks at the mud and concludes that spring has been cancelled. The mud is the spring, just the early part. You wash it off the roads, and you keep going.

That's the shape I expect. Not everything will get good. Some things will be worse for a while in ways the winter made invisible. But the direction changes, and the direction is what matters when you're deciding what to do with your hands.

The jobs will go. They always do.

AI will replace a lot of work. I'm not going to pretend otherwise; I'm using it to replace my own, every day, and the part it replaces is real. That's exactly what automation did to weavers, farmhands, and human computers. The pattern is not new. What's new is that this time it's coming for the desk jobs, so this time the people writing the eulogies are the people who own keyboards.

Two things the past says about it.

First: the count usually goes the other way. When cash machines arrived, the obvious prediction was the end of the bank teller. Between 1980 and 2010 a few hundred thousand machines went in — and the number of tellers went up, from about half a million to nearly six hundred thousand. Machines made a branch cheaper to run, so banks opened more branches, and the teller who was left stopped counting cash and started doing the parts of the job that needed a person. The job didn't vanish; it changed shape, and the people who changed shape with it were fine.

Second: complaining has a terrible track record. The weavers who smashed looms were right about everything except the outcome. Their pain was real, their wages did collapse, and the machines still won, because the machines were better. The people who did well were not the ones who fought the loom. They were the ones who learned what the loom couldn't do.

I already see the same story replaying at company scale: a firm announces a model doing the work of hundreds of support agents, then a year later is hiring humans again, not for the routine tickets, which the model handles fine, but for the ones that need judgement, patience, and someone to actually own the outcome. The volume went to the machine and the responsibility came back to people, which is how it has gone every time so far.

So: fit, don't complain. Not because complaining is beneath you, but because it has never once worked, and fitting has worked every single time.

How to get ready

The new times will not be easy. They will just be different, and different rewards preparation more than it rewards strength. Here's what I'm doing about it, roughly in the order I'd tell a younger version of myself.

Start with the cheapest one: read what you apply. Whatever the source — a model, a headline, a contract, a diff — if you're going to act on it, read it first. That's the whole lesson of every incident above, and it costs nothing but a few minutes of attention. Make the machine explain itself until you could explain it back, because if you can't, you're not using a tool, you're being used by one.

Then own your tools instead of renting the answers. Learn what the models are actually good at and where they fall over, from your own hands, on your own problems, not from someone's thread about it. The teller who understood the cash machine kept the job; the one who saw it as an enemy, or as magic, didn't.

While you're at it, move toward the parts the machine can't hold. I've said before that what stays with the human is the thinking, the taste, and the refusal to accept a bad answer; add responsibility to that list, and knowing why rather than just what. Every wave of automation has paid those things more, not less, because they're what's left when the routine is gone.

Build things, now, while it's cheap. The cost of trying an idea has never been this low, and the people who come out of a hard period ahead are usually the ones who spent it making something rather than refreshing feeds.

Keep a buffer — health first, then money, then skills, then people you can call. Hard times punish anyone running with no margin. This is boring advice, and it's boring because it has been right for every generation so far.

And don't outsource the thinking. The one thing you cannot delegate is being the person who decides. That was true when the deciding was done by kings, and it's true now that a model will happily decide for you if you let it. Don't let it. That's the vote.

The line stays

So the About page keeps its line. Better times are ahead — not because the calendar guarantees it, but because every night on record has ended, and every one of them was ended by people who got up in the dark and started building. This is a hard time, and it is forging people. Some of them are reading this.

Bring boots. The roads are going to be muddy for a while, and that is exactly what spring looks like.