Your kids probably won't work 40-hour weeks

In 1926, Henry Ford put his plants on a five-day, eight-hour week. Before that his people were working 48 hours a week, six 8-hour days, and he found the extra hours weren't producing much output. Tired workers slowed down and made more mistakes, and when he cut the hours, output held up. By 1940 that was the federal workweek, and it's been the same ever since.

What Ford ran was a compression experiment, cut the hours and see if the output holds. It worked partly because the output was so measurable. You could count the cars. Time became a proxy for output because the two were fairly one to one, and that proxy is the assumption this whole conversation is built on.

Computers didn't shrink the week

The challenge is that as the economy changed, the one-to-one got less and less true. Think about when personal computers showed up. What one person could produce changed completely, and the hours didn't move. All the new productivity got absorbed into the same 40. Nobody said great, we're making the same output as before, let's cut the week. People just did more.

Part of why is that white-collar work is hard to measure. In my experience it's tough to put a metric on success in most of these roles, so a lot of them come down to feel and hours. I don't think most companies really measure output at all. The measurement is people work their 40 and the company is growing or it's not, which is rudimentary, and a lot of whether the company grows sits outside any one person's control anyway. For most people the real measure is whether their direct manager says they're doing OK in a performance review.

Another compression experiment

I think what's in front of us, with everything changing this fast because of AI, is that we basically have to run another compression experiment. Everyone is getting more effective with less of their time, and now there are agents running in the background on top of it. The question is what that experiment looks like for knowledge work. Can you cut the hours and hold the output? If you can, is that the goal, or should the same hours just produce more?

It's a hard question because nobody's settled how to measure people right now. Is it output? Is it time? Is it effort? Is it quality? Every job is different, there's so much nuance in a modern company, and roles change so fast that I think it's almost an impossible answer.

Keynes already made this bet

Keynes was thinking about this in 1930. He wrote that his generation's grandkids would be working around 15 hours a week, because output would keep climbing while the time it took kept falling. As AI works its way into everything, the conversation that keeps coming up is UBI, the idea that machines do more and more of the work, people do less and less, and eventually you don't work for income at all. That's a pretty mainstream conversation at this point. The logic makes sense, but I still think we're a long way from there, and I don't think we go from where we are today straight to UBI. If we head that direction, I'd bet it looks more like what Keynes proposed, the week working itself down first.

That wouldn't even be a new trend. The week shrank for most of 150 years. In the late 1800s, a factory worker in the industrialized world put in 60 to 70 hours a week. Today, averaged across the year, the US is in the mid-30s, and some countries are down near the mid-20s. The US number hasn't changed a whole lot since the 1980s, while other countries kept adjusting.

There's a catch, though. The week doesn't seem to shrink on its own. Productivity gains never did it by themselves, or the PC era would have done it. The 40 came from Ford running the experiment and writing down the result, plus decades of unions demanding shorter weeks and a Depression-era Congress trying to spread the work around. The countries that work less now kept deciding the week should come down. Keynes's century is up in 2030, and the American week is still 40. Maybe he was just early, or maybe nobody's run the experiment for knowledge work yet.

The new social contract

The big one I keep coming back to is what this does between employers and employees. Everybody on salary is technically hired for an output. That's supposed to be the deal with salary work. So say AI gets you to where you can do all of your work in 10. Should you work the 40 and turn out four times the output? Or does the company say the output's there and call it a week? The capitalist in me says we all work our 40, and if you've got extra hours, more output goes in. The human in me says maybe it's OK if some people work less, spend more time with their kids, more time with the people they care about.

On top of all this, the eight-hour day is already kind of a farce. Lots of office time isn't work, it's lunch, water-fountain conversations, travel, all the stuff in between. The real work is maybe 3 to 4 hours of it.

What I don't know yet

There's a lot in this I'm curious about and don't have an answer for. How the nature of work changes from here. How you measure people and success when every role is different. How you keep driving growth and rewarding innovation and speed while taking care of people. How you do all of that without losing the social fabric or the engagement. I try not to think about it too long, because I start to spiral.

I don't really know where any of it lands. My own work is going to change in the next five or ten years, and I don't have a position on how, other than it feels like something is going to shift. I just think there's an interesting conversation to be had about what the next work week looks like, especially for knowledge work. Just like Keynes bet his generation's grandkids wouldn't work as much, I'm willing to bet my kids won't be working as much either. The 40-hour week came out of an experiment. I think the next one will too.