We create a job whenever someone with the authority to hire decides the value created is greater than the wages paid.
In my lifetime, we’ve invented 7 billion or more jobs, which is great news. Great for the people who were able to earn a living, and productive for everyone who experienced some of the value created.
“Value” doesn’t always mean profit.
David Graeber defined a bullshit job as “A form of paid employment that is so completely pointless, unnecessary, or pernicious that even the employee cannot justify its existence, even though, as part of the conditions of employment, the employee feels obliged to pretend that this is not the case.”
But over time, there are no pointless jobs. It might be that the value created is hard to measure, creates status for the boss, or is part of a larger system. It might seem pointless to the employee, but someone gets value out of it. Boats have ballast to keep them steady, and many big company jobs serve a similar purpose. Sailing a boat without ballast is difficult, regardless of how pointless carrying dead weight seems to be.
The person who pays for the job is the one who decides if it’s valuable. Calling it pointless from the outside is just substituting your judgment for theirs.
Graeber was right that doing a job that feels pointless is enervating. Even if it has a point, the more obviously connected we are to the creation of value, the more purpose we can find in our work.
And now, here comes AI.
It’s already good enough at many tasks that it’s cheaper and faster and perhaps more reliable to have an AI do those tasks instead of a person.
But we’ve seen this before. The Luddites weren’t anti-technologists. They were defending their livelihoods, a specific job: weavers. People who created clothing.
Today, around the world, we’re all clothed, but it’s quite unlikely you know a weaver. The cost of truly handmade clothing is so high that no one buys it. A handmade t-shirt might cost $500 if you could find one. Machines replaced skilled labor, and all of the weaving jobs disappeared, because the machines created more value per dollar invested.
It took a generation or two for the weaving workforce to fade out. That sort of labor change is now happening in months or years, not decades.
Most people reading this post have a job that didn’t exist a hundred years ago. Some of these jobs create easily measured and obvious forms of value, some less so. And all of these jobs will change as a result of AI.
Jevons Paradox is worth understanding. When the efficiency of coal combustion increased, it didn’t lead to a drop in coal use; instead, more uses were found for coal power, and consumption went up. This will likely happen, at least for a while, in many uses of AI. If it’s easier for programmers to write useful code using AI, then more programmers will show up and write more code, solve more problems, and create more value.
We keep building machines, and the machines give people chances to add value. But that value creation is more likely to happen when people use the machines for leverage, not when they try to do the work that the machine can do.
Bullshit jobs will disappear, especially as the meme of corporations with less ballast catches on. More status will be earned by having a smaller workforce, not a larger one. CFOs will get more value (for a while) from laying people off than from hiring them. The turmoil is certain, the human costs will be real, but the likely outcome is more value created by more people, over time. That’s not much solace for someone who has invested and trained to create value under the old rules, but consumers of value don’t often care about that.
The best plan is resilience. Find a way to create value, more each day. Consider enrolling in the bumpy ride that change brings, because holding on tight to the job we have today is probably going to be insufficient.
History doesn’t repeat itself, but it rhymes. Figure out who is seeking value and create it for them.
💸 Earn Instantly With This Task
No fees, no waiting — your earnings could be 1 click away.
Start Earning