A Map is Not a Blueprint: Why Fixing Nature Fails
Ozempic, Fertilizer, Lobotomies, and the dangers of hubris
Hi friends, today’s essay is about a mental model to protect you from the unintended consequences of scientific hubris. It might just save your life. If you’re new here, join the 45,000 other readers to get my next essay in your inbox.
We should expect interventions and innovations ranging from Ozempic to hydroponics to embryo freezing to oat milk to fail, or at least have significant unexpected side effects, for a common fundamental reason.
A way of seeing the world which, once you’ve added it to your visible wave spectrum, will induce a healthy skepticism of our many attempts to outwit biology.
It starts with the Coastline Paradox. Say you hire three people to measure the coastline of an island. A few weeks later, they all come back and have wildly different answers! One of them thinks the coastline is 10 miles long. One thinks it’s 20 miles long. And another thinks it’s 30 miles long. Who is right? And how is this possible?
It turns out that they’re all right, and they’re all long. The length of a coastline depends on how finely you measure it. Imagine a very jagged coastline like this:
If you measured it straight across, you would get one distance. But you would get a different distance if you zoomed in and measured each jagged edge.
A coastline is not a simple geometric shape like a square or circle. It cannot be precisely measured. It is a fractal, an infinitely complex shape where you would have to drill down to the atomic level to get a precise measure of it and maybe not even then.
Drawing a map then becomes a question of how precisely you want to represent it, and how much space you have on the map to do that representation. If the island needs to fit into a 1-inch diagram, you will have to sacrifice considerably more precision than you would if it were fitting into a 10-inch or 10-foot drawing.
A map of a coastline can approach accuracy; it can get infinitely close to accurately representing the coastline, but it can never fully represent it. There is an infinite amount of subtle detail that the map will have to leave out.
This doesn’t matter for normal navigational purposes. It will still help you find the beach, even if it’s only 90% precise. But imagine you had to rebuild the coastline from the map. Now the precision matters quite a lot! The more precise you measure the coastline, the more accurate your reconstruction will be. But here’s the important part: you can never successfully map the coastline well enough to accurately rebuild it. Your reconstruction based on the map will always be deficient. A tourist with their nose jammed in their phone won’t know the difference, but a local who walks the beach every morning will step into the surf and say “no, something isn’t right.”
A map is very different from a blueprint. A blueprint for a house, bike shed, car, or whatever starts in someone’s mind and then is sketched out as a guide on how to produce the thing. With a moderately detailed blueprint, say of a mid-size sedan or a silicon processor, you can produce something that is the thing. The blueprint is complete. Each item produced according to the blueprint will of course have subtle differences, they are not literally the same but they are functionally the same. And there is no ultimate representation behind the blueprint that they’re approximating. Unlike the Map, a Blueprint succeeds in fully representing what it wants to create.
Some technologies you interact with, like a chair or a car, effectively started in a human mind and then we bent nature to our will to create it. They are blueprint-based. Other technologies started in nature, and then we attempted to understand and recreate that natural phenomenon. They are map-based. You ignore this distinction at your peril.
Let’s talk about Trans Fats.
Proctor and Gamble had a thriving candlemaking business in the 1800s, but as electricity started spreading, the demand for candles fell dramatically. They primarily used cottonseed oil to make their candles, but as the demand for candles fell, they looked for a new use for their cottonseed oil.
In 1901, a chemist developed a process called “hydrogenation,” which turned liquid oils into solids to help preserve them for longer. It also happened to give cottonseed oil a consistency very similar to lard, the dominant cooking fat of the time, so Procter and Gamble bought the rights to the hydrogenation process, started using it with their cottonseed oil to create a room-temperature stable fat, and then branded it as a cooking product called Crisco.
Crisco was a remarkable product. It stayed solid at room temperature and could stay good for much longer than lard. It was also cheaper and much easier to produce, and it was kosher-friendly since it was a vegetable product instead of an animal product. They could also lean on it being vegetable-based in their branding, the first vegetable-based oil that makes your food just as delicious and creamy as animal lard.
It became an incredible business line for P&G and started the entire alternative-fats industry we have today. But I suspect you know where this story ends: it turned out that hydrogenated cottonseed oil was predominantly composed of Trans Fats, a type of fat so toxic to our health that it was eventually outlawed in many countries around the world starting in Denmark in 2003 and eventually in the United States in 2015.
The saga is a tragedy. People worldwide were eating a toxic fat engineered as a soap byproduct to help save a P&G business line. We could point to various causes of the tragedy: profit incentives, dishonest marketing, and poor regulation of new food groups, but I think the fundamental problem is this misunderstanding of the Map and Blueprint dichotomy.
In 1911, P&G thought they had a proper map of cooking fat. They believed that if they made something that was composed of fat, which tasted good, which performed the same in baking, then they would have made something just as good as lard. And that’s certainly how it was presented, in fact, it was presented as being better than lard. They thought they had a blueprint, but really they had a map scribbled in crayon. They had no idea how many parts of the coastline were missing. And it took a hundred years and countless deaths from heart disease for the error to be discovered.
Once the problem with trans fats was realized though, most companies producing vegetable oils switched to different formulations. By the time of the ban, trans-fat-free cooking oils like Canola and Soybean had taken over, advocating their Polyunsaturated Fatty Acids (PUFAs) instead of Trans Fats.
Like Crisco, these are fats made from industrial processing of huge numbers of seeds along with other solvents and chemicals to produce something like natural oils such as olive and coconut oil. But the question remains: how accurate is the map? And is it possible that in ten, twenty, or fifty years, we will discover that these attempts at making a blueprint of something from nature will fail just as spectacularly as Crisco did?
Let’s consider another example from psychology and mental health. In the early 1900s, as we developed a deeper understanding of the brain, a Portuguese neurologist named Moniz came up with a hypothesis: since the brain operates similar to a machine, certain mental pathways are likely the cause of mental illness. People have a defect in how their brains are wired, and if we can fix that wiring, their problems will go away. It’s probably no coincidence that this theory gained popularity as electricity was spreading around the world. Our maps of consciousness often mirror the dominant technology of the day, which is why we now think of the brain as a “computer.”
So Moniz came up with a radical treatment: break the faulty pathways. Identify where the brain is mis-wired, leading to mental illness, and then sever those wires. It’s kind of compelling, isn’t it? And you can imagine how easy the marketing was. “Your brain is mis-wired, but we understand the brain now, and we can help fix the wiring!”
His solution was what became known as the lobotomy. He drilled holes in patients' skulls and then severed connections in their prefrontal cortex, trying to disrupt the “broken pathways.” Eventually, lobotomies were also performed by forgoing the skull-holes and shoving a sharp instrument up through your eye socket and poking around a bit. The “ice-pick lobotomy.”
The logic that led to this solution is completely understandable. Moniz believed he had an accurate map of the brain, and that based on that map he could make subtle changes that would only have localized effects without broader consequences. But it’s a similar error to what happened during the first experiments with intentionally damming rivers. You can never fully anticipate what the spillover effects will be when you change something in a highly complex system, especially one that you don’t actually have a deep understanding of.
Some of the results from lobotomies were promising, with a few patients showing improvements. But there were also many showing concerning side effects ranging from apathy and being emotionally disconnected from life to committing suicide.
Thankfully, today we have a more accurate map of the brain. We now know that people with mental health problems have chemical imbalances in their brains which can be solved with a lifelong prescription to certain pharmaceuticals. These pharmaceuticals are incredibly effective for some patients, though others have shown concerning side effects ranging from apathy and being emotionally disconnected from life to committing suicide.
I’d like to give you one more example, one which is going to become increasingly prominent and concerning over the next few decades: fertilizer.
Before the early 1900s, you had broadly two ways to fertilize soil and support crop growth. Manure, made from the excrement of animals, and compost, made from decayed organic matter. In both cases you were reintroducing nutrients to the soil that had been extracted from it. The cows and sheep eat the grass and poop back out some of the nutrients, and the rest gets returned to the soil when they die through their carcass and the waste of the people who consume them. The same goes for plants: a crop of corn pulls nutrients from the soil, gets consumed, and then the nutrients are returned to the soil through waste and compost.
As people moved off farms and into cities, and as agriculture became more industrialized, this cycle was broken. No one is shipping their poop from New York City to Iowa to remineralize the corn fields. And with animals moved off of pasture and into feed lots, their manure is often not making its way back into the soil either.
This problem was fine for a while, but you can only deplete the nutrition in the soil for so long before there are consequences. And, worse, a growing urban and suburban population demanded an even greater investment in industrial agriculture to support it.
This conflict led to the famous predictions of Thomas Malthus that humanity would eventually reach a crisis point where the planet's resources could no longer sustain our species’ exponential population growth. Agricultural production could only increase linearly because 1 acre of land can only produce 1 acre of food and there are a finite number of acres of land in the world. But humans are increasing exponentially; in 1800, there were about 1 billion people in the world, in 1900 it had only risen to 1.6 billion, but in 2000 there were 6 billion. As I’m writing this in January 2024, there are an estimated 8 billion people.
Malthus’s predictions were taken seriously and created quite a bit of alarm, but they never came to pass. We found ways to get more calories per acre through crop selection, mechanization, pesticides, fungicides, and fertilizers. Estimates put the amount of corn per acre produced in 1800 at about 25 bushels, but in 2000 we could produce 150 bushels. A 6-fold increase.
We clearly achieved something amazing with food production, right? Well, maybe. One of the core innovations that led to this dramatic increase in food production was fertilizer. Since the nutrients in the soil that plants need to grow were no longer being returned to it and were slowly being depleted, we started replacing them with synthetic fertilizers, first developed in the early 1900s.
The impact of these fertilizers was incredible. Crop yields on land that added fertilizers could as much as triple. And land that had previously been barren could be rejuvenated into healthy usable land through the liberal application of these fertilizers. Suddenly there was more farmable land, and that land produced more food. A double strike against Malthus.
But, was it? Or did it just delay the realization of his predictions? We now know that foods grown with synthetic fertilizers often end up with significantly lower mineral and nutrient density than foods grown in their more naturally fertilized way. Fruits, vegetables, and even grains like rice and wheat might have as little as 25% of the vitamins and minerals they had 100 years ago. You need to eat somewhere between two and four times as much food today to get the same amount of essential vitamins and nutrients, and this nutrient depletion is one contender in the medley of theories for why we’re seeing such a dramatic rise in obesity in the west. People are eating more because they’re nutrient-starved, because we thought we thought we had an accurate map of soil health. We used poor metrics like crop size and yield and ignored the more important metrics like nutrient density. And this doesn’t even touch on the disastrous consequences we’re starting to see from many of these synthetic pesticides and fungicides.
I’m sure that by this point you are understanding the general character of the problem:
First: We see a biological, natural process in the world, and we want to understand it. So we study it as intensely as possible and make our best efforts to map it to explain it to ourselves and others.
Second: We see some deficiency in the process as it exists today. Farmland is running out. Some of us have mental health struggles. Natural fats are expensive and unkosher.
Third: Using our map of the process, we create a solution to the problem. Trans fats, lobotomies, synthetic fertilizers.
Fourth: The solution will seem like a home run. It will address all the problems of the original system, with no side effects, because the location of the side effects is outside the map. The consequences of hubris will always be hiding in the parts we don’t yet understand.
Fifth: Problems will eventually emerge, and we will realize that we did not have an accurate map in the first place. Heart disease, suicides, soil depletion. So we go back to the source, study it as intensely as possible, and make a new best effort to map it and explain it to ourselves and others.
And then the cycle repeats.
I am not anti-technology or anti-science. But I am strongly against the mistake of thinking fertilizer is like an airplane. Or that understanding mental health is like understanding a circuit board. This doesn’t mean it’s not worth trying to solve these problems. But we need to approach these interventions with a hefty dose of humility.
Sometimes, the risk is worth it. The smallpox vaccine had side effects, but that risk was a hell of a lot better than the risk of smallpox. If your choice is between starving to death and eating nutrient-poor food, the latter is certainly preferable.
But other times, it's an absurd risk to expose yourself to. The latest frankenfoods like Oatly or Impossible will certainly have harmful ingredients we don’t understand yet. New “miracle” drugs like Ozempic will certainly have side effects we don’t know about yet. It is not a question of if, but when, and how bad the consequences are.
The processes they’re attempting to map developed over billions of years of evolution, and there are an infinite number of hidden coves, secret caves, and forgotten beaches we may not have explored.
Sail at your own risk.
Thanks for reading! If you enjoyed this, join 45,000 other subscribers to get the next essay in your inbox.