Automation Is a Myth

Some HNers might be interested in the book below, which is coming out in a couple weeks with Stanford, and which can be pre-ordered now. It uses a provocative claim to push against the myths of automation but also develop a messier and more interesting portrait of the 'future of work'. Happy to field any questions.

-----

For some, automation will usher in a labor-free utopia; for others, it signals a disastrous age-to-come. Yet whether seen as dream or nightmare, automation, argues Munn, is ultimately a fable that rests on a set of triple fictions. There is the myth of full autonomy, claiming that machines will take over production and supplant humans. But far from being self-acting, technical solutions are piecemeal; their support and maintenance reveals the immense human labor behind "autonomous" processes. There is the myth of universal automation, with technologies framed as a desituated force sweeping the globe. But this fiction ignores the social, cultural, and geographical forces that shape technologies at a local level. And, there is the myth of automating everyone, the generic figure of "the human" at the heart of automation claims. But labor is socially stratified and so automation's fallout will be highly uneven, falling heavier on some (immigrants, people of color, women) than others. Munn moves from machine minders in China to warehouse pickers in the United States to explore the ways that new technologies do (and don't) reconfigure labor. Combining this rich array of human stories with insights from media and cultural studies, Munn points to a more nuanced, localized, and racialized understanding of the "future of work."

There's a lot of work that hasn't been done yet in studying what causes automation from an economics point of view.

A common thought experiment is "Why hasn't fast food been automated?" We certainly have the technology to make burgers automatically, to the same quality that MacDonalds does. This technology is at least 40 years old. Yet we still have kids running these stores. The reason seems to be: kids are cheaper than machines. This seems to suggest that it is low wages and economic precarity which is preventing automation, rather than that these are the jobs most threatened by it.

Fast food is significantly automated, though. Depending on the chain, you get various different parts of the process centrally automated, and standardized food components and ingredients are shipped out to restaurant locations or franchisees. Only prep and assembly are deferred to the very end, from the various standard components. The reason fast food works, and is consistent enough that you can expect the same Chalupa in California and New York City, is absolutely automation and centralization.

I think the bigger likelihood here is that JIT assembly of a few components into many distinct products with differences in technique or ratio is still cheaper to do with manual labor.

> The reason seems to be: kids are cheaper than machines

I'd rephrase that as kids are cheaper to fix than machines. A burger flipper machine is almost definitely more efficient and cheaper per burger, but when it breaks its probably a lot more expensive to fix or replace than a kid.

It's the same idea, but I think that you need to separate the "cost per burger per hour per hour" and the "cost per burger per hour per year" since only one of those will consider the cost of unexpected costs.