Brain Food

Ideas you can chew on for days.

The Spermaceti Whale, Stewart Lizars (1843)

A quick note: On April 27th, I’m running the London Marathon to raise money for UNICEF, a charity that helps the poorest kids in the world. With massive foreign aid cuts from Trump (and the UK government), this help is more essential than ever. I’m personally matching donations up to £1,000 ($1,350) to double the impact. If you are able to give even a few dollars, consider donating by clicking here or the button below. Thank you—I’ll be thinking of your generosity when I’m utterly wrecked.

---

Fruit Fly Robots

The 21st century is shaping up to be the golden age of neuroscience. Some of the breakthroughs will be unequivocal positives, free from moral dilemmas. The possibility of ending or delaying neurodegenerative disease, such as Alzheimer’s, now seems within plausible reach. Breakthroughs are swift and astonishing.

However, there is a potentially dark side, too. For the first time in history, our understanding of brains has advanced to such a degree that we are on the cusp of truly dystopian methods of directly controlling minds. This is not science fiction, but science—as has been proven in a darkly fascinating new paper published last week in the Proceedings of the National Academy of Sciences journal.

Researchers have long used fruit flies, Drosophila melanogaster, as a model species to study complex biology in a comparatively simple organism. Last year, scientists were able to fully map the entire neural connections in the brain of fruit fly, a precursor to more complex neuroscience in more sophisticated species. (Researchers just last week announced a full, detailed map of a tiny part of a mouse brain, an astonishing feat that will eventually unlock countless advances in medical research and our understanding of how neural connections work).

While studying Drosophila, it has become clear that fruit fly brains are predictably responsive to certain stimuli, including certain visual cues and smells. For example, flies use moving visual cues to navigate. “Essentially, the fly attempts to maintain a steady course by adjusting its flight or walking direction in response to perceived motion. For instance, if a pattern moves from left to right in the fly’s field of view, the fly will turn to the right, effectively trying to follow the motion.”

The upshot is this: if you present a fly with a pinwheel image that rotates clockwise, the fly will turn right in unison with the image. If you make the pinwheel image rotate counterclockwise, the fly will turn left. (The researchers also manipulated the fly’s movement using an array of olfactory cues—smells—though it required some more complicated interventions).

With this method, the scientists were able to predictably control 94 percent of the fly’s movements. Their degree of control was so substantial that they could get the flies to spell out words on demand. Below, the figure shows the theoretical operation of how the cues work (A & B) to spell out the prompt "HELLO WORLD,” followed by actual tracking data from how the flies spelled out the words successfully.

Diagram C shows the fly movements when scientists used visual clues, compared to Diagram D when it was based on scents (which was slightly less effective).

Moreover, while flies, unlike dogs, are not intrinsically motivated by balls, the researchers were able to use visual cues to get flies to move toward a ball, engage with it, and ultimately move it around on command. Without their interventions, the flies would—as predicted—ignore the ball. But the scientists could effectively prompt the flies to “do work,” pushing the ball from one place to another.

At this point, it’s not just flies who have their antennae alert, because my ethics antenna is tingling: scientists have effectively found a way to enslave flies to do our bidding. This field of research, known as “microrobotics,” explicitly refers to these living creatures as robots, terminology that makes me uncomfortable, especially as such neural manipulations will not only be used on flies in the future.

Worse, the researchers highlight that future research will allow them to explore directing flight while forcing the flies to carry payloads at the direction of humans. This research, they note with no apparent awareness of how dystopian this sounds, remains uncertain “particularly with respect to how much weight a backpack can add while still allowing flight and effective guidance signals.”

Maybe this makes me a biological luddite, but the notion that we can turn animals into robots through deliberate mind control seems like the kind of manipulation that could lead to some rather dark futures. It also speaks to the urgent need for careful, informed regulation to keep apace with technological advances, particularly with artificial intelligence and neuroscience, as it’s already becoming scientifically possible to accurately read human thoughts (and external manipulations of human brain activity will likely soon become more feasible, too).

Prehistoric Psychopaths

Are humans naturally violent creatures who can only be tamed by the threat of punishment, or are we a naturally peaceful species, spoiled by a few violent rotten apples?

Whether human nature exists—and what it is if it does—has triggered fierce debates for thousands of years. But in the modern era, for the first time, we can bring data to the debate—and answer these burning questions with a bit more empirical evidence. Two scholars, John Halstead and Phil Thomson, argue that our violent past can, at least partly, be chalked up to a few vicious psychopaths who always lurk within any group of humans, from prehistory to the present.

This heavily researched essay, published by Works in Progress, is worth reading in its entirety; it is densely packed with dozens of astonishing insights that can quickly send you down a million anthropological and evolutionary rabbit holes—and the scholarly research it’s based on is available here. But a few key insights are worth highlighting.

Humans were likely less violent in the hunter-gatherer prehistoric past than in the era of subsistence farming and agricultural cultivation (which began around 12,000 years ago).¹ Some scholars argue that violence increased when controlling territory became valuable due to farming, whereas in the hunter-gatherer period, there was no need to control specific bits of land.

Moreover, as agricultural production rose and cities emerged, inequality soared, which created the impetus for violence. (I explore this notion, which I call the “War and Peas” hypothesis—and its relationship to the emergence of power hierarchies and greater inequality—in the second chapter of Corruptible).

However, as social cohesion intensified—both through the threat of shared religious punishment and the rise of modern states—rates of violence eventually declined. At first glance, this gives credence to the notion that human nature is depraved—and that Hobbes may have been right. Our state of nature is base.

This graphic is from Works in Progress, available here.

Worse still, our species has seemingly even more to answer for, because it’s not just that our prehistoric past was violent; it’s also that it was far more violent than other animals.

Hunter-gatherer humans killed each other far more often than other primates—and orders of magnitude more than most other mammals. A study led by José María Gómez in 2016 found that about 2 percent of prehistoric humans would have been killed by another human—and the 2022 study by Halstead and Thomson put that figure for hunter-gatherers between 3.4 and 3.7 percent. For comparison, Gomez’s data suggest that roughly 0.3 percent of all mammals are killed by a member of their own species.²

And yet, there’s perhaps reason to be hopeful—to disprove Hobbes—even with such striking differences. Halstead and Thomson point out two key features of human societies that can take that data and allow us to re-interpret it with a much rosier sheen.

First, humans are both uniquely lethal and uniquely vulnerable.

The defining characteristic that drives high rates of violent death in our species is not our proclivity for lethal violence but rather our capacity for it. Human beings are unusually vulnerable to violence. We have massive heads, thin skin, puny muscles, little to no protective fur; we can’t fly, swim, or burrow away, and we’re not even very good at running away. Our children are even more fragile, particularly as babies, and take ages to mature.

At the same time, our offensive abilities make us the most lethal species on the planet. Violent attacks in a hunter gatherer context are essentially undefendable. We have abilities to collectively organize, plan, and deceive far in advance of any other species. Even lions are afraid of us. Our stone-tipped tools, poisons, and projectile technology appear to have killed off almost all of the planet’s megafauna, like mastodons, giant kangaroos, and saber-tooth tigers.

In other words, the rates of human-on-human homicides will understandably be elevated relative to peer species not because we’re more vicious to each other, but because our intra-species conflicts lead to more deaths. Plenty of other species attack each other without anyone dying; we are better at killing and less able to survive attacks, which manifests itself in comparatively higher death rates. Perhaps our higher homicide rate isn’t a reflection of a rotten human nature, but a statistical artefact of these unique traits. (I’ve written about how our unique ability to use ranged weapons has modified social structures—in the link below).

The origins of power dynamics, or why chimps can't play baseball

Second, adding to this more optimistic interpretation of human nature is the notion that most killing is often carried out by a few violent psychopaths within any given group of humans. The researchers point to research on a modern, heavily-studied hunter-gatherer group, the !Kung, in which the majority of murders over a thirty-five year period were carried out by just two people. Modern data from Sweden suggests that one percent of the population commits 63 percent of all violent crimes.

This interpretation is a double-edged sword. On the one hand, it provides clear and convincing evidence that humans are not innately depraved. On the other hand, as I argue in Corruptible, modern systems of power often allow psychopaths to wield control over entire societies, magnifying their ability to kill on a scale that was never before possible. When sociopaths get nukes, we’re in danger.

The key, then, is to more carefully engineer society—to relegate the violent psychopaths to the fringes, far removed from power, while designing social structures that amplify our widespread and astonishing tendency for cooperation, decency, and problem-solving. This is not going to be easy, but that challenge represents, to me, the defining social goal of the 21st century.

Ping, You’ve Got Whale

This article from the ever-fascinating science magazine, Nautilus, highlights a form of artificial intelligence that I hope we can all get behind: a new tool called WhaleSpotter that helps ships avoid striking and killing whales by giving captains real-time alerts about nearby cetacean movements.

Every year, thousands of whales are struck by ships and boats and either injured or killed. Many of those deaths are caused by container ships, the backbone of global supply chains, which are so massive that they can’t turn fast enough to avoid striking a whale even if they visually spot one and try to avoid it.

The technology is rapidly improving, getting rid of false alarms that cause ships to distrust alerts, and is on the cusp of being able to reliably provide warning of nearby whales from up to four miles away. The key, as with so many unequivocally good technologies, will be to get companies to adopt it when the only benefit is not profit, but saving majestic, innocent creatures who don’t deserve to die on the altar of global consumerism.

How Silica Gel Took Over the World

Those little silica packets are everywhere these days, warning us not to eat them. This essay, from Scope of Work, decides to ask and answer a question we’ve probably all wondered but not researched: what is up with those little sachets of dryness and where do they come from?

Tear its little Tyvek wrapping, and spill a packet of glassy silica gel beads into the palm of your hand; they won’t hurt you. They are made of the same stuff as sand: “Silica” means “silicon dioxide,” which is the primary component of most drinkware, windshields, and the screen of whatever electronic device you’re reading this on.

But glass has a density of around 2500 kilograms per cubic meter, and crystalline silicon dioxide (quartz) is around 2650. Silica gel, on the other hand, is more like 700 kilograms per cubic meter. It may look fully dense, but in fact it’s shot through with countless tiny pores. If your windowpane is like a thin sheet of solid ice, then a silica gel bead is like a tiny snowball.

The author, Spencer Wright, notes that these properties mean that a single tiny packet of silica gel could theoretically produce the “internal surface area of eight hundred square meters—the size of almost two basketball courts.” As a result, they can “absorb up to 40% of its own weight in water vapor,” which makes them excellent at keeping products dry from potentially ruinous moisture and humidity.

Silica gels have been around for centuries and were used in gas masks during World War I. But a synthetic process for producing silica gels was only patented in 1919 by a researcher named Walter Patrick at Johns Hopkins University.

The reason for the recent surge of those little packets isn’t because of a recent technological breakthrough, or a sudden realization that it’s useful to keep stuff dry. The longer that something needs to travel to get to the consumer, the greater the risk of spoilage from humidity. Locally made products that are to be opened upon arrival rarely need silica gels. The rise of those little packets is simply an indicator of globalization.

Quick Links

“The Crab” by

Sam Kriss

A profoundly moving essay about recently losing his mother to cancer. Read it.

“The Bluetooth Test and Other Keyholes to the Soul” by

Adam Mastroianni

A lovely rumination on why humans engage in pointless status games, blatant selfishness, rigid stubbornness—and how we can be better with some quick, useful, and amusing tests to examine our own behavior.

Kleptocracy, Inc by

Anne Applebaum

On the inevitable marriage between authoritarian populism and corruption in the Trump White House.

Film: Nightcrawler (2014)

I’m 11 years late to watching this 2014 Jake Gyllenhaal film, but I accidentally timed my viewing perfectly. It’s a story about an overly ambitious sociopath who wants to make it, no matter the costs, and stumbles upon the freelancing career of filming gruesome crime video for local TV news in Los Angeles. Watching it in 2025, I realize how prescient Nightcrawler was at warning us about the perils of the attention economy—and the social devastation that people will eagerly churn up to get ahead, earn money, and capture eyeballs and clicks.

Film: Black Bag (2025)

If you like artistic spy thrillers, then this is the movie for you. The taut dialogue is over-the-top (nobody talks like that), the impeccable fashion is over-the-top (spies don’t dress like that), but it all just adds to the ambiance. And the ending is great.

Book: Marble Hall Murders (2025)

Longtime readers will know that I love well-crafted murder mysteries and Anthony Horowitz simply doesn’t miss. This new book, which is the third in a trilogy involving the sleuthing book editor Susan Ryeland, repeats the plot device used in Magpie Murders and Moonflower Murders. The clues to the modern day mystery are hidden within a cosy detective novel set in the 1950s with Atticus Pünd, a German version of Hercule Poirot. I have read—and loved—both of his Sherlock Holmes novels, all five of his Hawthorne & Horowitz series, and now all three of the Susan Ryeland mysteries. If you like murder mysteries, I’d recommend starting with the Hawthorne & Horowitz series or Magpie Murders.

Book: Metazoa (2020)

This isn’t a quick or easy read, but it is fascinating. The SCUBA diving philosopher Peter Godfrey-Smith, author of Other Minds, moves on from the sexy topic of octopus intelligence to the perhaps less obviously exciting world of ancient animals. However, in the process, he convincingly explores a mind-bending question: how did intelligence and consciousness first emerge?

Thank you for reading The Garden of Forking Paths. This edition was free for everyone, but I rely exclusively on paid subscribers to keep researching and writing, so please consider upgrading to support my work—and fully unlock nearly 200 essays in the archive. Or, check out my book: FLUKE, twice named a “best book of 2024.”

1. The 2022 study that this essay is based on finds rates of violence in this prehistoric period that are four times lower than the estimates used by Steven Pinker in his book The Better Angels of Our Nature. But there’s a lot of statistical uncertainty here.

2. Thankfully, we are not close to the worst species on this metric. Nearly 20 percent of meerkats are killed by another meerkat; other elevated levels of intra-species homicide include some species of lemur (17 percent); lions (13 percent); the California ground squirrel (12 percent); and grizzly bears (10 percent).