The other day I left a RALPH loop running overnight. The idea was for it to run PHPStan, PHPCS and the tests for a module on its own, fix whatever it could, and flag what it couldn't. I went to bed. The next morning all I had to do was review what it had done, and 70% of the boring work was already sorted.
And that's when something I'd been chewing on for a while clicked. If someone watched me work like that, they might think I'm lazy. And they wouldn't be wrong. The thing is, I think that laziness, mine and that of plenty of people I know, is exactly what's going to be in demand more and more.
The good kind of lazy isn't the one who doesn't work
Let's clear up the misunderstanding right away, because the word "lazy" has a really bad reputation.
I'm not talking about the person who doesn't lift a finger. I'm talking about the one who refuses to do something by hand twenty times when it can be done once, properly, and forgotten. The one who, before starting a repetitive task, stops for a second and thinks: "there has to be a way to make this run itself."
This isn't new, by the way. Larry Wall, the creator of Perl, said more than thirty years ago that the three great virtues of a programmer are laziness, impatience and hubris. They sound like flaws, but he meant them as compliments. Laziness was the effort you put in to reduce your total effort: building something once so you never have to repeat it. Impatience was the anger you feel when the computer is being slow, which is what pushes you to write programs that anticipate what you'll need instead of just sitting around waiting for your commands. And hubris was the pride of writing code nobody will want to complain about, plus that slightly excessive confidence that you can handle any problem. All three have their place, but the one we care about here is the first. You write a script today so you never have to touch that thing again. There's even a line attributed to Bill Gates that goes the same way: for a hard job, pick a lazy person, because they'll find the easy way to do it.
The good kind of lazy works. Sometimes harder than anyone. What they can't stand is working on something that adds nothing.
Until recently, you also needed the workhorse
For years, companies that wanted to automate ended up hiring someone with that mindset. Someone who'd look at a manual process and say "no, I'm not doing this by hand, I'll set it up so it runs itself." That person was worth their weight in gold precisely because they were rare.
But alongside them, and for a long time, you still needed the workhorse. The one who rolled up their sleeves and got the work done without asking whether there was a better way. Head down, task done, next. And it worked. There was so much manual work that someone had to take it on, and whoever cleared it fast and without complaining was a real asset.
The problem is that the world where that was enough is coming to an end.
What changes with AI
What AI has done isn't invent automation. That was already there. What it's done is lower the bar for automating, and lower it a lot.
Before, setting up the script, the cron job or the pipeline that took a tedious task off your plate meant a chunk of work and a certain level of skill. Now you explain to Claude Code or OpenCode what you want, leave it running in a loop, and forget about it. Tasks that weren't worth automating before, because the script cost more than just doing it by hand, suddenly are.
And this is where the lifelong workhorse runs into a problem. Not because they work badly. They work as well as ever. The problem is they've still got the blinders on, looking straight ahead, doing the work by hand like it's ten years ago. And meanwhile, off to the sides, new ways have appeared to do the same thing in a fraction of the time, at lower cost and with fewer human hours invested. They don't see them. Not because they lack the ability, but because they've never had the reflex to look up and ask "isn't there a better way?".
That reflex, which for years was almost a luxury, is turning into the thing that matters.
The other side: laziness done wrong
That said, I don't want to sell this as if the workhorse is obsolete and the lazy person is the answer to everything. That would be oversimplifying.
There's a kind of laziness that's poison. The kind where you hand the prompt to the AI, take whatever comes out at face value, and ship it to production without reviewing it. That's not automating sensibly, it's outsourcing your judgment. There's a recent article by Sean Goedecke that nails it: he argues that in the age of AI the virtue isn't just laziness anymore, it's suspicion. Your default attitude has to be that the model probably got something wrong and it's on you to find it. The good kind of lazy automates, but reviews. The bad kind automates and prays.
And then there's the fact that not everything automates equally well. In Drupal I see it crystal clear. A migration from a D7 full of inherited technical debt, a config you have to keep stable on a high-performance portal, a caching issue you only understand by reading logs patiently... that's where rushing and laziness cost you dearly. Sometimes the value is precisely in the diligence, in sitting down and doing it right even when it's tedious. Laziness done right knows when to automate and when to roll up its sleeves. "I can't be bothered to think" isn't the same as "I can't be bothered to repeat myself." The workhorse's problem was never working hard, it was never looking to the sides.
So what do I do with this?
If I have to take a stance, here's what I'd say.
If you're hiring, stop rewarding volume alone. The person who clears twenty tickets by hand in a day looks more productive than the one who clears five but has built something so the other fifteen handle themselves from now on. The first one solves your problem today. The second one changes your trajectory. All else being equal, I'll take the second one.
And if you're the one doing the work, the exercise is simple even if it's uncomfortable. Next time you catch yourself doing something repetitive and with no enthusiasm, don't just put your head down and grind through it. Stop. Ask yourself whether a machine could do it. These days, the answer is often yes. And if the answer is yes and you keep doing it by hand anyway, that's the problem the industry is going to start looking at closely.
Looking up and glancing to the sides has become part of the job. Not a bonus reserved for the sharpest few.
I might be wrong, for the record. I've been at this for a good few years now and the industry changes its mind constantly, so maybe a while from now I'll reread this and cringe. Could be.
But from what I'm seeing, that's the way things are heading. And I know which side I want to be on: the one who goes to sleep and wakes up with a good chunk of the work already done.