SCIOPS 01.34: Fully Automatic Machine Fun

Do you want to know how soon a robot will take your job?

It’s not a question of “if”. Robots have been taking our jobs since the start of robots. The whole science-industry-progress narrative is based on automation, both in the form of machines (cotton gin, calculator) and in the form of human control systems (“scientific management”, Prussian schooling). Humans are replaced with machines or organized into machine-like assemblages. Even the word “robot” comes from the Czech word for “worker” – or “slave”.

How soon will your job be automated? You can get a rough estimate by assessing the amount of craft and art it requires of you.

Craft is about the convexity of your learning curve. Is your job easy to learn, with not much reward for high levels of practice? Something like mowing lawns, for instance: someone with 40 years of mowing experience isn’t that much better than someone with only one year. Versus a job like computer programming, where the first year of experience makes a much smaller difference to your overall skill level than the 40th year. The amount of craft, the distance between apprentice level and mastery, has an inverse relation to automatibility. Things with low craft will be automated sooner.

Craft is about low-hanging fruit. If I can build a robot with a 1-year skill level at either of these jobs, it makes way more sense to automate lawn-mowing than programming. It can mow anything, except highly advanced lawns. The 1-year programming bot will barely be a beginner, and will struggle to stay afloat, probably falling in with some bad influence-bots and taking up a life of automated crime.

Art refers to the uniqueness and “touchiness” of each particular task in the job. A job with a high art level might be edible landscaping, where each garden has to be custom-designed to site factors, and where the feel-good factor of human interaction is important. Cashier is a low-art job: the process for each transaction is basically the same, and the human interaction is less valuable than efficency.

When someone’s making permanent changes to your landscape, you want to engage with that person. You want a bond of trust and a customized experience. But when you’re buying a cardboard box full of aluminum cans full of carbonated water, you don’t need warm-fuzzies. You just need your cold fizz.

So the less art and craft your job requires, the sooner you’ll be automated away. There’s no place in capitalism for inefficient workers. Even the professions, jobs we think of as being highly skilled, will be replaced with computers and robots.

Lawyers will still exist in 10 years, but (I would guess) an order of magnitude less of them. At least. Because the low-hanging fruit will get picked: writing wills, standard contracts, and the like. And the advanced lawyering, actual “Your Honor may I approach the bench”-type stuff, will be done by human/robot hybrids: centaurs. The human will do the smooth talking and the wearing of expensive suits, and their computer will do the researching of precedents and the finding of loopholes.

It’s not a question of “if”. It’s a question of where the profits go.

Next time, I’ll explore the ways we Puny Humans could turn automation into paid vacation, instead of forced resignation. See you in a week!

Thanks for reading,

– Max


is a cogsec newsletter. If you liked this issue, please forward it to a friend. The virus will only cease when it has breached all inboxes in the world. You can also share your thoughts, critiques and questions by responding to this email. Or, write me at .

If you’re not signed up, you can sign up at