69. The unbundling of humans (in agriculture)

A look at unbundling of physical and cognitive skills

Hi. If you are new here, I am Rhishi Pethe, and I am excited you’re in the “Software is Feeding the World'' community. Every Sunday, you will receive this free newsletter at the intersection of technology and agriculture systems. I am a product manager at Project Mineral (focused on sustainable agriculture) at X, the moonshot factory. The views expressed in this newsletter are my personal opinions.

The unbundling of humans (in agriculture)

If you asked someone involved with agriculture in the US Midwest whether robotics, automation, and algorithms will be able to do all the tasks involved in farming, the answer will be a strong “NO.”

A human will always be involved.

Agriculture is complex, with an open system, unpredictable weather patterns, variety of soil characteristics, localized decision making, and decision making based on experience. Due to these factors, it is believed that a human can never be taken out of the equation.

At a high level, this is mostly true, not just in the US Midwest, but almost anywhere in the world.

The question of which tasks will be automated and which tasks will always require a human is complicated

Throughout history, we have found human tasks and outsourced them to machines, and now more and more to algorithms and models. For example, a combine automates many tasks, makes harvesting easier, safer, and more efficient. We continue to unbundle physical and rote mental tasks so that we can focus on higher-order bits requiring creativity and imagination.

Take the mundane example of booking your travel. I am old enough to remember to make a phone call to book my tickets for travel. Now, 99% of the time, I don’t need to talk to anyone to make a reservation.

Take the harder example of self-driving cars. We want to unbundle our driving skills, outsource our fatigue, (and our drunk driving) to an algorithm. We want an algorithm to take us from point A to point B in a safe and efficient fashion. It is a much harder problem to solve, but we are working on it.

So the question is:

Do humans want to be unbundled?

Unbundle us (but conditionally)

I posit that most of us want to be “unbundled” to a large degree. We don’t want this unbundling to threaten our livelihoods, but to improve our life and to create new opportunities.

According to MIT economists Daren Acemoglu and Pasqual Restrepo, (“Automation and New Tasks: How Technology Replaces and Reinstates Labor1”), if productivity growth in the future continues to be “automation”, the importance of labor can decline. The creation of new tasks (for example, automation has created new jobs for data analytics, scientists, engineers etc.) and continued wage growth will depend on skill sets, demographics, markets, policy, and innovation ecosystems.

Simply put, as we unbundle human capabilities, we will create new tasks, and whether the labor participation goes down or not will depend on the balance between the human tasks that are automated away, and new human tasks that get created.

In the steel industry in the US, the number of human tasks (and humans) that were automated far outnumbered the number of new tasks that were created. This resulted in a 42% drop in employment, big political changes, and no material drop in production.

Graph from the St. Louis Fed

Something different was observed in the banking sector. When hundreds of thousands of ATMs were introduced over the last 30 years in the United States, conventional wisdom was that it would reduce the number of human tellers drastically. This expected outcome did not pan out, and in fact the number of human tellers increased. Cash handling became a less important and smaller part of a teller’s job, but their ability to use interpersonal skills to work with bank clients became more important.

The automation of the cash transaction created time and space to do new higher order tasks.

What gets unbundled?

So what are the aspects of today’s farming in the developed world that can be unbundled?

Traditionally, tasks that are repeatable, more or less predictable, and do not have stringent human skill requirements are good candidates for automation and unbundling. There are many examples in agriculture.

Auto-steer: A GPS guidance system that steers agricultural equipment with centimeter accuracy.

Prior to auto-steer, the operator of the tractor had to focus and judge where they were driving the tractor. With auto-steer, they can free up their mind (a bit), and focus on higher value-add activities or rant on #AgTwitter! Auto-steer was adopted very quickly by farmers. From a 2016 article,

John Deere AutoTrac™ went into production in 2002. The guidance system has become so popular that today 60 to 70% of the crop acreage in North America is farmed using AutoTrac or similar systems. In Australia, that number exceeds 90%.

The above figure shows data from a survey of Nebraska Farm Business, Inc producer members and included in Castle’s thesis2

Recently Raven industries, a leader in agriculture automation released an automation platform called OMNi.

OMNiDRIVE is Raven's aftermarket technology kit that transforms existing tractors into driverless machines. It is the first farming application that allows the farmer to monitor and operate a driverless tractor from the cab of the harvester so the harvester can offload on-the-go, then return the tractor and grain cart to a predetermined unloading area - all without a second driver.

You just unbundled another worker!

OMNi enables holistic, hands-free precision; it gives the farmer total control of their time and maximizes operational productivity and efficiency.(Highlighted by me)

If we take the example of auto-steer technology, it has reduced operator stress and fatigue, increased accuracy (row shutoffs, section control), and improved efficiency (reduced fuel & input costs). 

Physical or mental unbundling?

Within agriculture, most of the talk around autonomy and automation has been around the unbundling of physical tasks. Based on the definition of autonomy from Cash IH, we are already on our path towards full autonomy. The different levels of autonomy are still heavily focused on unbundling the equipment operator and some of the human infrastructure that goes with it.

What about tasks that require experience, special knowledge or skills, and trust?

We see examples of highly difficult tasks like conversations on the autonomy spectrum, ranging from a bot with no autonomy (Level 0: Live Chat with a human) to human-like autonomy (Level 5: Fully Autonomous Bot), to  handle all situations without the need for a human. Level 5 means the bot will learn and teach itself autonomously, continually improving and evolving.

Within agriculture, it is not too much of a stretch to imagine a machine learning model, which chooses the right crop type and seed variety based on your past history and agro-ecological conditions. It creates a plan to plant the right amount of seed in the right parts of your field at the right time. The grower or the agronomist can review (or not) the plan.

It will search for the best price for the seed (or a relevant substitute) and place an order. The seed will be delivered in time for your planting operation. The seed is loaded into your autonomous planter. The planter has an accurate map of your field and will plant accurately to the plan created by the model.

If the model can do this consistently, and deliver results, the trust in such models will continue to improve. You might still need an agronomist or an advisor, but their role will have changed dramatically.

So human beings are being unbundled not only from the low end of physical skills, but also the high end of decision making.

Paul Ford writes in Wired,

Billions of us need help making millions, billions of decisions. Decisions about whether to upgrade HVAC systems, or how to fuel our shipping, or what to plant in the backyard. Sometimes it feels like the paradigm has inverted. Technology was the mold growing across human systems. Software was eating the world. Now it feels like humans are the mold growing on technology. (Highlighted by me)

Are humans just mold growing on technology in the future? It might feel like that in a few years.

Farming’s new definition

Depending on the rate of technology adoption, and what gets unbundled, what it means to “farm” might change in the future.

Mark Young, ex-CTO of The Climate Corporation (digital farming subsidiary of Bayer Crop Science) and I discussed this in edition 52 of the newsletter.

The notion of what it means to be in agriculture will change over time. You could be a “farmer” and work in a warehouse. Even in broad acre production it'll change because as autonomous equipment and automation takes over, what it means to be a farmer will change. (Highlighted by me)

What can agriculture look like in the future?

Noted political strategist, Peter Zeihan provides one view for the US Midwest. He imagines a future in which automated equipment with cameras will take photos of each individual plant. The equipment will identify if the plant is a weed, or a crop and assess the health of the crop. It will give it a little jolt of whatever is appropriate (herbicide, pesticide, fertilizer, water etc.). He believes we are on the verge of production increasing by a factor of 2-3 (“on the verge of” is subjective, though within the realm of possibility.) It will turn conventional farming into conventional gardening, with a lower pollution rating and a far lower carbon footprint. (Highlighted by me)

I am all for the autonomous tractor, or my Rumba, or automated temperature control. I want these advancements, but in a way which frees us up to work on higher order tasks. I want us to have more time to be creative and unique. I want us to spend time with our family and friends. I want to fulfill what it means to be human, without sacrificing our livelihoods.

Jim Barksdale, the former CEO of Netscape, once famously said,

There are only two ways to make money: Bundling and Unbundling.

Maybe there are only two ways to be human: Bundling and Unbundling.

So, what do you think?

💗 If you like “Software is feeding the world”, please hit the “Like” button at the bottom or share with a friend.

🙏 If you don’t mind answering 3 questions anonymously (2 are optional), I would love to get your feedback.


Acemoglu, Daron, and Pascual Restrepo. 2019. "Automation and New Tasks: How Technology Displaces and Reinstates Labor." Journal of Economic Perspectives, 33 (2): 3-30.


Castle, M.H. (2016). Has the Usage of Precision Agriculture Technologies Actually led to Increased Profits for Nebraska Producers? (Masters thesis). Retrieved from University of Nebraska-Lincoln Digital Commons.