Even if our institutions were thoughtfully designed—

Even when progress is real—

Politics still sucks. And it always has.

When something persists across radically different times, cultures, and belief systems, it’s not an accident. There’s a reason.

The world is uncertain.

It changes in ways that cannot be reliably predicted, yet decisions must still be made in real time. Rules must operate not only under present conditions, but under conditions that do not yet exist.

If there is a complete, eternal rule set that accurately and effectively anticipates all future states of the world, we do not know it. Which means any rules we create must be revisable.

People think and act autonomously.

Humans are not inert matter waiting to be acted upon by authority. They are self-regulating systems that continuously generate activity, adjustment, and correction simply by remaining alive. They do this spontaneously, whether authority approves or not.

This is not an appeal to freedom. It is a descriptive claim. Humans act. We think and believe. Denying this is equivalent to denying skin.

If that seems controversial, see the footnote on homeostasis.1

Autonomous action creates conflict.

People who think and act autonomously will want and believe different things, especially as circumstances change. This is true even among honest, well-intentioned people acting in good faith.

Once multiple such agents share space, conflict is not optional. It is emergent.

That conflict must be managed—by people who always disagree, under conditions they cannot control, using protocols that must continue to function as those conditions change—or else it will devolve to violence.

All of that may seem simple—and it is—but importantly, if it’s true, we can actually derive the minimum rule set required for any political system to avoid collapse into violence.

1. The world is uncertain.
   This is obvious.

2. People think and act autonomously.
   This is inescapable.

3. Autonomous action creates conflict.
   This is true whether authority acknowledges it or not.

4. To mitigate conflict, co-inhabitants of any shared space must coordinate behavior, resolve disputes, and enforce agreements.
   This is a polity.

5. Any coordination system requires rules to function.

6. These rules:

   1. Must be publicly knowable.
      This includes, without exception, all rules for making rules.

   2. Must not exclude any person or group subject to them.
      Rules may govern entry or participation but may not define any person or group subject to the rules of the polity as non-citizens or second-class citizens.

   3. Must not forbid internal, non-violent means of expressing, contesting, and revising disagreement.

   4. Must not be permanent or exempt from future cancellation or revision with the sole exception of this minimum set.

What this means

When people encounter a political outlook different than their own, they usually say something like “That’s immoral!” What follows is a fight over values—justice, rights, equality, freedom, tradition.

Those arguments can matter, and sometimes do, but only rarely because people will always think for themselves. People of different temperaments and backgrounds tend not to converge on the same vision of “the perfect society.”

The framework enumerated above is both simpler and more robust than theories that ground legitimacy in moral abstraction (however real some morals might be). Rather than wrestle with hypotheticals, this framework simply says:

A political system is legitimate to the extent it makes human flourishing possible over time.

Of course, in addition to those other things, we also don’t agree what “flourishing” means in any given situation. Even if we did, it would change as conditions change.

What doesn’t change is this:

Violence, repression, and premature death make human flourishing impossible, regardless of how you define it.

In other words, we’ll never agree on what the good life is. But there is widespread agreement on what it is not.

Legitimacy is a function of how much any system lets people argue about and experiment with what the good life should be in the context of their own shared circumstances.

That’s it.

This will undoubtedly feel “thin” to some people. For example, it doesn’t tell you what to believe or even what to value. That’s intentional.

The axioms above are not proposed as values. They’re proposed as instruments.

Classical “thick” moral theories require a kind of cosmic knowledge that is supposed to generate a broad agreement that never arrives. This framework is narrower and nearly impossible to escape:

If a political system can’t adapt, can’t include the people it governs, or can’t limit violence, then whatever else it promises, it can’t support human flourishing—regardless of what anyone thinks flourishing is.

Why it works

This framework makes testable predictions.

It does not predict which slogans will win or which policies will pass. It predicts how and why political systems fail. Those predictions are falsifiable—but to be fair, the test results take decades.

Still, it can be refuted. Just point to a human polity that permanently suppressed agency, froze its rules, and/or persistently excluded some of the people it governed without collapsing or resorting to violence.

If it’s true, then we can say something simple but deeply uncomfortable:

• every current political system will fail

• even yours

because all of them violate the primary constraints in some way.

If that seems grandiose, consider the laws of thermodynamics. They do not design engines. They tell you which engine designs will fail, allowing you to design working engines—even very excellent ones.

In fact, it’s fair to describe this framework as a thermodynamic theory of politics: not because politics is physics, but because living, self-regulating systems such as people spontaneously generate internal pressure.2

There’s still an art to designing good government, but this is the science. Just as the laws of thermodynamics don’t specify what a Ferrari looks like, this framework doesn’t describe a good government—it rules out the bad ones.

I. Anti-Utopian Corollaries

These aren’t extra principles. They are things that cannot be true if the framework is true.

1. There is no final ultimate political arrangement

If uncertainty is real, and people spontaneously think and act autonomously, then no configuration will be permanent. If change happens, then the future will be different—on a long enough time scale, very different—and the people of the future will want what they want.

Any claim that says politics is solved or history has ended is structurally doomed on some time frame.

This isn’t pessimism. It simply denies a political Rapture. There are no “end times.” The people of the future will be just as messy as the people of today (but with lasers and cool shades).

The best you can ever get is maximally stable participation under local conditions.

2. Perfect justice is impossible and all attempts are coercive

Perfect justice assumes a shared standard by which outcomes can be judged. And yet, people who are honest, well-intended, and acting in good faith still routinely disagree—not about whether murder is wrong, but about whether a particular act counts as murder (versus self-defense or just plain negligence).

Justice depends on judgment: which facts matter, which motives count, which risks were reasonable. As a result, some people will perceive justice where others perceive injustice—not because they are confused or malicious but because they have different values, different experiences, and necessarily occupy different points of view in space.

A system committed to eliminating injustice can:

• tolerate disagreement, or

• impose a single measure by force.

The first minimizes injustice. The second hides it with an accounting trick.

Injustice can and should be reduced. But there is no such thing as “perfect justice.”

3. Politics cannot be replaced by administration

Think of it like musical chairs. Technocracy fails because at any given moment, when the music stops: experts disagree, expertise lags reality, and decisions must still be made under uncertainty.

When things change, administration cannot justify, revise, or replace the existing rules without:

1. re-entering politics, or

2. resorting to coercion

Guess which one they pick.

4. Moral purity is incompatible with long-term coexistence

Any system that requires uniform belief, moral unanimity, or permanent moral asymmetry will eventually fracture or turn violent because all of the people will:

• never agree on everything

• never agree forever

People will continue to think and act autonomously whether you forbid it or not.

Therefore, coexistence always requires tolerating wrongness. And that’s not just, like, my opinion, man.

II. What this is not

Heading some people off at the pass.

A. Not Hobbes

Hobbes treated violence as the natural baseline and resolved it through sovereign finality.

This framework treats violence as the default failure mode, not the baseline, and rejects final authority because it blocks revision.

It’s not that life is “nasty, brutish, and short.” It’s that whatever life is, it’s punctuated by unresolved conflict—and any system that pretends otherwise eventually reaches for force.

B. Not classical political realism

Realism centers power, interests, and domination and treats norms as epiphenomenal or ornamental.

This framework treats norms and rules as coordination technology, with power as a variable in stability, not a source of legitimacy.

Might does not make right. But when coordination fails, people who believe otherwise tend to make themselves very relevant, very quickly.

C. Not Burkean conservatism

Burke emphasizes tradition and slow evolution and is deeply skeptical of rational reconstruction.

This framework is not hostile to tradition but allows change when it preserves what tradition is meant to preserve.

Traditions survive only insofar as they remain adaptive. Inheritance matters, but stability and flourishing matter more.

D. Not liberal moralism

Liberal moralism grounds legitimacy in rights, dignity, or consent—often abstractly and treats violations as moral outrages.

This framework grounds legitimacy in a negative condition—rejection of what prevents flourishing—and treats violations as mechanical inhibitors rather than moral affronts.

Rights may be fundamental. This framework doesn’t decide. It simply says that whether rights are metaphysically sacred is secondary to whether they actually work.

E. Not cynicism or nihilism

I will get accused of this by people who did not pay attention.

• Cynicism says ideals are fake.

• Nihilism says nothing matters.

This position says coordination matters, and without it, violence returns. That isn’t nihilism.

It may be tragic, but it doesn’t despair. Insofar as the framework identifies real constraints and workable remedies, it’s grounds for cautious optimism.

F. Not “the American system in theory”

This framework is not a restatement or defense of the American political system under a different name.

It does not privilege:

• constitutional republics

• federalism

• representative government

• bicameral legislatures

• democracy

• markets

• individual rights

The American system is neither source material nor an exception.

Where similarities exist, they exist for a simple reason: the American Founders were grappling with many of the same constraints—uncertainty, disagreement, faction, violence, and the problem of revision.

They arrived at one historically contingent set of solutions. Some of those solutions addressed these constraints. Others introduced new failure modes that are now familiar.

Like any other polity, the United States will succeed or fail to the extent that it respects or violates these constraints. If it’s serious about succeeding, it probably needs to resurrect amendments.

III. What immediately fails

This is the part people will love. Here are the systems we know will not work along with why.

1. Utopian Left (revolutionary egalitarianism, “post-politics” visions)

Primary violations:

• Uncertainty

• Revisability

Failure mode:
Assumes that once correct consciousness or material conditions are achieved, politics ends or becomes purely administrative.

Why it fails:

• Material conditions inevitably change—pandemic, war, technological shock, ecological stress.

• Disagreement does not disappear once the “right” conditions are declared achieved.

• Revision becomes betrayal, so compliance substitutes for consent.

Result:
Coercive stagnation and periodic purges framed as “defense of the revolution,” culminating—predictably—in another revolution.

2. Technocratic Liberalism (managerial governance, expert rule)

Primary violations:

• Autonomy

• Participation

Failure mode:
Treats disagreement as ignorance, misinformation, or cognitive error rather than persistent divergence of values.

Why it fails:

• Experts disagree.

• Expertise lags events.

• People resist being optimized, even when optimization is well-intentioned.

Result:
Rule by procedure that drifts into policy theater. Evasion becomes normal (black markets, shadow expertise), and legitimacy decays until revolt or capture resets the system.

3. Libertarianism / Anarcho-Capitalism

Primary violations:

• Participation

• Enforcement

Failure mode:
Assumes exit and voluntary contracts can replace collective coordination.

Why it fails:

• Exit is unevenly available and asymmetrically costly.

• Power in social systems compounds rather than remaining evenly distributed.

• Contract enforcement and redistribution require mechanisms the theory disallows.

Result:
De facto polities emerge anyway—unacknowledged and unaccountable. Warlords, cartels, or princes fill the gap, eventually collapsing into one of the more stable failure modes.

4. Authoritarianism (right or left)

Primary violations:

• Participation

• Revisability

Failure mode:
Attempts to suppress autonomous behavior by treating people as programmable components—or zombies.

Why it fails:

• Autonomous action reappears as sabotage, black markets, or rebellion.

• Adaptation to changing conditions occurs only through elite fracture or violence.

Result:
Long periods of apparent stability conceal brittle systems that fail suddenly and catastrophically.

5. Ethno-Nationalism / Identity-Exclusionary Politics

Primary violation:

• Participation

Failure mode:
Defines some subjects as permanent outsiders or secondary participants under the rules that govern them.

Why it fails:

• Excluded groups do not disappear or stop wanting what others visibly have.

• Stability requires escalating coercion to maintain exclusion.

Result:
Chronic instability punctuated by repression, insurgency, and eventual war or territorial fragmentation.

6. Moral Absolutism

Primary violation:

• Revisability

Failure mode:
Treats certain rules as sacred rather than structurally necessary.

Why it fails:

• Change is inevitable. Edge cases arise and accumulate.

• Conflicts between absolutes emerge but revision/adaptation is taboo.

Result:
Either quiet decay into another form or violent rupture after an inciting event.

This framework does not rule out:

• democracy

• constitutionalism

• markets

• welfare states

• cultural pluralism

• and so on

It also does not say they work. Any system can and will fail where it violates the core principles. However, some systems are not excluded outright.

Final thought

This framework doesn’t tell you what political system to build. It tells you which ones won’t last. That doesn’t make it pessimistic or nihilistic. It makes it useful.

It works the way thermodynamics works because human polities are composed of living, self-regulating systems that spontaneously generate action. Action creates collision. Collision creates pressure. Pressure can be managed, redirected, or released, but it cannot be eliminated. Political systems that pretend otherwise don’t transcend conflict—they defer it to their children.

We can do better. Systems that respect these constraints will, on average, suck less. They will fail more slowly, break less violently, and adapt more often.3 But, importantly, there is no future where politics doesn’t suck.

That isn’t because people are uniquely terrible. It’s because they’re alive.

The media’s role in narrating that history — live as its happening — is fraught. They are, after all, sitting with us on the same moving platform of the present. Ostensibly, they have better seats, but that makes them an impediment as much as anything else. We have to crane our necks to see round. (Not everyone will want to try. Some will take their word for it.)

The historian, sitting at the back of the platform, has the grandest view. All of history recedes before them. Staring into the long sunset, they know its lessons better than us. But they do not have any better view of the future than we do.

I started The End of the World Almanac to help us understand our historical moment, a moment where the platform is picking up speed. History, like evolution, does not proceed at a steady pace. It never stops moving forward, but it often does so in fits. We are living through one of those fits, more important even than 9/11 or the Second Iraq War or the financial crisis or COVID because what is changing is the mechanism that determines whether or how all of those things happen.

Or not. Like the Arab Spring, it’s possible that what comes out of this turmoil is something not all that different than what went in. But that’s what makes it so exciting to witness. None of us — not the narrators, not the historians — knows how this will end.

If things are going reasonably well for you (or if you are strongly deferential to authority), all this turmoil won’t seem worth it. But there are a lot of people — and an increasing number of them, whole communities even — whose fortunes have been declining steadily for a while, and for them turmoil is a chance for change.

If this frustrates you, consider that these people have been consistently asking for change without turmoil since at least 2008. (Despite their obvious differences, the Occupy movement and the Tea Party had a lot in common.) All economic protests have been told to sit down. The economy is great. Or at least the best it can be. Your struggle isn’t real. Or if it is, it’s your fault. Or if it’s not your fault, the international order is just too big and too important to alter.

When, in 2016, the Panama Papers revealed widespread corruption and tax evasion and linked wealthy political figures, business leaders, and sports stars to organized crime, then-President Obama gave a press conference where he basically shrugged. I watched it. Those things weren’t technically illegal, he explained, and so nothing could be done.

After the Second World War, the US had the only major industrial economy not blown to smithereens. Under the threat of global thermonuclear annihilation, America built a vast sphere of influence that some have called an empire. (If so, it’s a funny empire.)

All empires are basically state-level protection rackets, where local leaders trade tribute for some ill-defined security. Sometimes in a protection racket, the people you are being “protected” from are the racketeers themselves — in the sense of “Pay me or I’ll break your legs.” But other times, there really is a rival gang, and you are paying to stay out of a mess.

The American empire promised to protect you from the Soviet empire. This was not always true or desirable, but it was the justification. Vast markets were opened, which we could rapidly supply.

The American people tolerated this because it benefited them. It meant jobs and an ever-increasing standard of living where, as long as you kept your head down and your nose clean, you could be reasonably assured that you would do better than your parents and your children would do better than you.

That is no longer the case. The younger generations now are the first in American history to do worse than the ones before. There are many reasons for this, and I’m oversimplifying, but in general, this is because the Soviet empire went away and with it the justification for tribute at the same time other countries rebuilt or industrialized (think Japan and China) such that America no longer held a consistent competitive advantage.

Under this erosion, imperial elites, who always benefited most from “empire,” maintained their wealth by taking a larger and larger share of the tribute. They did this not by growing the economy, as under Manifest Destiny, but by cutting costs — offshoring industry, importing cheaper goods and labor, etc. Note how Western business didn’t start investing heavily in China until a decade or so after the Berlin Wall fell.

This is reflected in the economic data. After a period of explosive growth, real wages have been effectively stagnant for decades. The middle class has shrunk, even as a greater and greater share of households survive on two incomes rather than one (meaning families are pedaling twice as fast and still sinking).

In short: For the last 30 years, imperial elites have continued to benefit from “empire” while the American people increasingly have not.

Empires have to be administered. Local intelligence has to be gathered. Bribes have to be made and tribute collected. All of it has to be kept secret. That requires personnel. Personnel also have to be paid. Their security status has to be approved and monitored. And so on.

America had had intelligence services before World War I, but they were typically attached to the military and were always short-lived. In 1917, three weeks after we declared war on Germany, we created the Black Chamber (officially: the Cable and Telegraph Section), and it survived all the way to 1929.

It wasn’t until after World War II that we created a permanent state security apparatus. This wasn’t because of the existence of threats. There had always been threats, including from the Soviet Union specifically (as well as the Nazis) in the 1930s. Rather, it was because of nuclear war. No prior threat could destroy us instantly. We could no longer wait until the onset of hostilities, as we did with the Black Chamber. We had to know in advance.

The post-war order and the state security services are the exact same mechanism. There is no distinction. Some of it, like NATO and the UN, is visible. Most of it is not.

It wasn’t just that the Cold War was cold and intelligence services have to operate in darkness. It was that America wasn’t an empire in the traditional sense. The fraction of the earth that it directly administered (outside its borders) was minuscule — Guam, the US Virgin Islands, Bikini Atoll. We conquered Japan and West Germany and South Korea, but as with the Philippines, we did not keep them.

But we kept a kind of indirect control over things that were important to us. And of course there was the matter of tribute. We weren’t doing all of this for free. But if we didn’t have vassal states, they couldn’t simply cut us a check. There had to be another mechanism.

These mechanisms are often exceedingly dark, complicated, and not easy to summarize, but they all involve the movement of value (not always money) in ways that are not easy to discern — although any good spy novel will give you an idea.

Under the American imperial system, money went OUT to secure economic and strategic interests. Does a certain country have exploitable natural resources, for example, and if so, where exactly? There have been enough geologic surveys by now that these days we typically know, but that wasn’t necessarily the case in, say, Africa in 1965, where there may not even be roads to the area in question.

To answer it, the US Army Corps of Engineers can’t simply land in a helicopter. Instead, a certain non-governmental organization (NGO) with an innocuous sounding name, like the Institute for International Cooperation, will receive a grant from a US government agency with an equally innocuous sounding name, like the National Endowment for Democracy. The NGO will in turn award a part of that grant to the geology department of a very respectable university (and keep the rest.) The very respectable geology department will send a very respected geologist to the wilds of Africa on the pretenses of A) advancing scientific knowledge of the globe and B) fostering local development.

You get the idea. By the end of the story, an American company will be in that country pulling those resources out of the ground for the major benefit of wealthy Americans. But they’ll hire some locals, so the grant will be stamped: SUCCESS.

Of course, now that we have secured an economic interest in the country, it’s very important to keep the communists out. Also, we want to make sure the local leaders don’t get any crazy ideas about actually owning their own country’s natural resources, so different government agencies with very peaceful names will give different grants, and suddenly there will be all kinds of Americans (or American-aligned people) in-country to promote children’s health or women’s rights or something so wonderful no one could ever possibly object, monies for which flow through the imperial mechanism.

Money coming in also has to be distributed, and it works a similar way. A teacher’s union, for example, might partner with another NGO which has received a grant from the Department of Education. Separately, the teacher’s union might buy 150,000 copies of a certain politician’s book to hand out to their members. That of course immediately propels the book to the top of the bestseller list, which generates vast additional sales. (To be fair, the book might even be half-decent since it will have been ghost-written by a professional author.)

Did the union buy the book because the politician supports teacher’s issues or does she support teacher’s issues because they bought the book? All we know for sure is that the grant the union received will be more than the cost of the books, although it might not look that way. The books might have been purchased at an undisclosed discount which the publisher logs as a tax write-off. Everybody wins — except for the people, who paid for the grant with their taxes and who might’ve also bought the book because it was on the bestseller list.

The point is simply this: Both money in (tribute, kickbacks) and money out (bribes, payoffs) have to be distributed by means indistinguishable from organic transactions. Everyone involved has to benefit but in a way that looks like they might’ve earned it. That requires a large, organized, well-funded bureaucratic mechanism.

In 1950, this system was comparatively small. By 2025, it has ballooned outside anyone’s full comprehension.

What we are presently witnessing — if it does not end up like the Arab Spring — is nothing short of the dismantling of the post-WWII system. Given that it was created as a response to the Soviet threat, a threat that hasn’t existed for 30 years, it’s probably overdue. I mean, did we really think in the ’90s that history was over? Like, nothing would ever change ever again?

I admit that I am in favor of some kind of change. That doesn’t mean I support everything that’s happening. But it is certainly the case that the people fighting against change are NOT offering an alternative vision. They are the major beneficiaries of the post-WWII status quo administered by a bureaucratic mechanism that long outlived its reason for being and which survives not just on taxes but ENORMOUS public debt. They have taken even from the future.

Regardless of where you’d like to see the country go, it isn’t 1950, or even 1980, or even 2010. It’s 2025, and history is back. We should not be fighting for the mechanism. It would’ve collapsed before long anyway. We should be fighting for what replaces it.

It’s starting to change now, but for the longest time, there were only two points of view you could have on climate change: that it was all a lie dreamt up by Ivory Tower ideologues who didn’t know about sun cycles or something; or it was a world-ending calamity that required us to rip the guts out of everything RIGHT FUCKING NOW.

It’s an era of High Moronics. Think about it. How well do you have to know someone now to disagree with them about even the simplest things, let alone anything that actually matters? If you’re against Trump, it isn’t enough to simply not vote for him. You have to not vote for him with prejudice. You have to disown anyone who won’t disown anyone for voting for him.

That’s not an accident. There’s a reason.

I’ve been thinking about how to shepherd my children through this world, a world I did not grow up in, a world where even supply chains are fragile, a world of pathologized hyper-concern for people not present over those who are.

The simplest solution seems to be: don’t let them grow up in it. Lest you think I’m crazy, that’s basically the advice of the experts. Social psychologist Jonathan Haidt recently published some guidelines for parents, and they include not letting kids on social media (especially girls) before they’re old enough to drive.

I don’t know that there’s a hard line. Every kid is different. But I take the point. The elephant dung in the room is the coordination problem. If your children’s friends are on social media, your children will also be on social media whether you forbid it or not.

Like adult junkies buying their kids weed, parents today seem enthusiastic about getting their children hooked. I’m not sure how we break that. Shame and derision?

A few years back, it became fashionable to think of data as the new oil. (There is something to that.) The latest analogy is between social media and smoking. The hope is that in 100 years there will be someone very similar to you looking at a graph very similar to this comparing social media use with rates of anxiety/depression/suicide, where those people will wonder why we ever let children on social media in exactly the same way we wonder why anyone ever let kids smoke.

I used to think the battle for civilization was between city dwellers and everyone else, which you see clearly on every granular map of electoral outcomes (which is why everyone uses it). But really it’s deeper than that. It’s between people who live a primarily screen-based life and those who don’t, with lots of folks in the middle. It just so happens that most of the screen dwellers live in large cities.

The screen dwellers are more prepared to accept what they’ve been told because they have nothing to compare it to. They live in a simulation. Everything is constructed and manipulable. The job they do services some part of the simulation rather than anything real. The company that employs them is a digital entity. The people they interact with are as well.

That’s not to say screen dwellers don’t encounter others. But to what degree is their Uber driver or the nameless barista at the coffee shop any more than an NPC in the game of their life?

This is why screen dwellers are obsessed with consensus: the COVID consensus, the climate consensus, etc. Paradoxically, it’s an evolved biological response. In a world devoid of fixed landmarks, where you never know what’s real, there’s a game-theoretic rationality to following the collective observations of the herd. What other measure is there?

It’s a mistake to think the other group, the people that do not live a primarily screen-based life, are mentally or morally superior. They’re not. Most people in history were in that group, and there’s plenty of barbarity to be found there.

It’s also wrong to think of this as the defining difference between red and blue. It’s not. Those categories existed before the simulation and would exist independently of it. (Again, most people are not going to be in the core of either group but somewhere in the middle.)

Rather, it’s that the simulation, being grounded in nothing, allows those groups to diverge more wildly. The people who don't live a primarily screen-based life have something to compare things to, which is why they’ve become increasingly skeptical of what they’ve been told, where it used to be the other way round. When physical commodities like oil were the primary determinants of the economy, the non-screen dwellers were the very same pathological conformists the screen dwellers are today. (Don’t let them tell you otherwise.)

Think of the historical division between agricultural societies (farmers) and pastoral societies (herders) in antiquity. Both had different models of the universe, different pantheons, different virtues, different heroes and monsters, and within each society you had a mix of traditionalists and innovators. There isn’t a mapping of one side then to one side now. It’s simply an example of the epochal scale of the difference, where the screen dwellers are raising entirely new gods.

Traditionally, the resolution was war. I used to think we could avoid it because the simulation and the alternative can occupy the same physical space in a way agriculturalists and pastoralists could not. I saw us fracturing into a multiplicity of lineages, some cultural, some physical. There are already biohacking and transhumanist movements whose offspring will produce something not-quite-us.

This isn’t new, by the way. It’s an aberration for there to be only one extant species of hominin. For most of our lineage, several species cohabited. I expected that would be the case again.

But now I’m not so sure. Or at least, I’m not sure the bottleneck is over. We may still speciate (where speciation could include a non-human successor), just later than I thought.

The simulation is not compulsory at the current time because it can’t be. Lethal force exists outside it. Combat requires real-world skills, which is why the military is not known for having a high proportion of screen dwellers (who would be ill-advised to force a conflict tomorrow).

But all that changes when farming, energy production, and violence are automated. Once robots do most of the work required to bring food and power to the people, even if they are still tended by humans, our last tangible tethers to the physical world will be gone.

Once robots are the more lethal combat force, a reality we are rushing toward, everything can be subsumed by the simulation.

Why wouldn’t it? Even if it were not hostile and merely subsumed what it encountered randomly, what’s to stop it from taking over?

I’m not talking about AI/AGI, by the way. It’s such a hot topic now that I could understand if anyone confused the two. The simulation could exist without AGI. The simulation could direct AGI. The simulation could be directed by AGI. It could also be ended by AGI. Like the three-body problem, the confluence of emergent technologies is notoriously impossible to predict.

But we don’t have to predict the future. If AGI never comes, we’ll still have the simulation because it exists now, mediated by current technology.

If you don’t believe it’s possible for the simulation to control reality, consider we witnessed it overwrite basic mammalian biology in real time. Facts evident since the evolution of sexual reproduction 2 billion years ago were got rid of in less than the runtime of the average TV show. Organizations with staff and budgets came into being based on this new reality. Laws and policies were written and enforced. Surgeries were given. The physical world was altered to reflect the simulation.

It isn’t that those beliefs will persist. They’re already fading. The simulation is polymorphic by nature. When it moves from its current obsession to its next one, reality will be retconned such that the ideas we’re not talking about now and can’t even imagine will have always been on everyone’s mind. (This was the original purpose of the Almanac, to chronicle every end of the world.)

I’m not one to predict the future by running a trend line from the present. There is no surer way to be wrong. But the implosion of social life into the simulation this century has become a singularity pulling everything toward it, right down to our sudden inability to couple and reproduce. That suggests there are massive changes afoot. Demography is history.

Some people think the solution is to go back. But even if we could put the genie back in the bottle (how??), there’s every reason to expect the forces that led us here would lead us right back again. As silly as it sounds, that’s Thanos’s realization in Endgame. You can’t simply reset the clock. You have to do it in a way that prevents it ever from returning. Without the means to do that, the only way out is through.

That leaves me with a quandary. If I’m right, then at some point soon, everyone will have to worship the new gods. (Many, many people already are.) I expect the fulfillment of this change will be comparable in scope to the switch to settled agriculture.

That may sound like hyperbole, but keep in mind that, on an individual level, people who saw others farming and decided to do the same still had a continuity of experience. If you convert to a new religion, you’re still you. What made the switch to settled agriculture or the rise of Christianity so significant was not that anyone did it but that everyone did.

Of course, there will still be outliers. Those that refuse the simulation won’t necessarily be exterminated. They could be living museum pieces like the Amish or people with flip phones—quaint artifacts from a time that is no more and won’t be again.

But make no mistake. The widespread switch to a primarily screen-based life, to the simulation, represents an epochal change for humanity.

Coming from an earlier world, I have to wonder if there’s an alternative. Is there something it is like to live a good human life in the simulation? On the numbers, it wouldn’t seem so. Everyone there appears anxious, depressed, or addicted. But we can no more imagine the future than hunter-gatherers could imagine the Byzantines or the Byzantines could imagine us. Both would be appalled at the things we find most comfortable.

In other words, just because I find the future abhorrent doesn’t mean it won’t be livable. But it makes having children now very poignant. I feel like an ancient fish staring at the shore knowing that’s where the species is headed but unable to lead my offspring there. The most I can do is make them aware of the change and encourage all of you to do the same.

When I was in high school, I had a hard time conceptualizing entropy. If all those gaseous atoms in the box were moving about carrying heat, but entropy wasn’t heat (or even “disorder”), then it seemed like there had to be some magical “something else” passed in the collisions.

When I was in college, I had a hard time conceptualizing qualia. I understood the philosophical difference between primary and secondary qualities, but qualia were supposed to be something other than mere color. But color was all I could perceive. What was this other thing?

As an adult, I had trouble conceptualizing nothing. I would sometimes look at the night sky and wonder why there was something rather than nothing at all, but I could only hold a sense of that question for a moment before it slipped out of my conscious mind like a wriggling fish.

In middle age, I still can’t conceptualize the different kinds of infinities.

I’m constantly amazed that people think a 400-year-old point of view is still the latest, greatest, most sophisticated and progressive possible. There is matter and energy, atoms and the void, and that’s it. To speculate otherwise is to give in to superstition.

Consciousness is a brain state, or an illusion, or whatever, but it is absolutely not a thing, even though you’re experiencing it now, and if you wonder what it might mean for an immaterial thing to exist, you’re practically burning witches.

There is nothing left to discover. We will unite gravity with quantum mechanics and that will be it. The grand project will be over. The end. Nothing to see here. Move along, move along.

Bullshit.

Every time we think we’ve figured it out, the universe gets stranger, almost as if it’s inviting us to approach it with wonder.

When did the hall monitors get control of everything? So someone somewhere believes something silly, or wrong. When was that not the case? They’re not the ones holding us back. How will we move forward if we can’t grope about in the dark?

Bring back wonder. Bring back daring. Bring back adventure and experimentation and failure and forgiveness. Bring back tolerance—real tolerance—and openness and playfulness and joy.

Forget 400-year-old dogmas. Let’s do something new.

I asked Stable Diffusion for some covers to textbooks of Occult Physics and Quantum Psychology.

“Of course I want to go to your daughter’s wedding. Thank you for inviting me.”

A liar’s lies are much more effortlessly sophisticated. A liar often tells the truth, for example, just selectively. They omit clearly relevant facts so as to leave you with a false impression, usually to elicit an emotional response (anger, sympathy) designed to manipulate. Social media algorithms do the same.

A liar doesn’t really care if you believe them, although they’ll pretend they do to keep up the appearance of truth. Rather, the goal is to get you to relent. It doesn’t matter if you grumble the whole time or swear it’ll never work again. A success is a success.

And they’re damnably persistent.

If you break down normal everyday speech, you’ll find most people have an assumption of honesty. That’s not to say we believe everything someone tells us, but we typically believe that THEY believe it, that they are not consciously misrepresenting their own internal thoughts and deliberations.

Persistence is a marker of that, of honesty. If someone consistently reports the same thing, we tend to accept that they believe it, regardless of whether we do. Indeed, studies of persuasion have shown that mere repetition is often enough to convince someone that a statement is true. At a minimum, that another person would go to all that effort at least makes their statement seem more likely.

“Jeez, he’s really on about that UFO stuff. Maybe there’s something to it.”

Yes, people are lazy and rarely check anything for themselves. But you can also think of it as a kind of cognitive efficiency. It’s much easier to rely on the assumption of honesty which, let’s face it, is usually correct in small everyday circumstances. Society couldn’t operate otherwise. Think about how hard life would be. How would you get through even one day if you had to independently verify every road sign, every price at the grocery store, everything everyone told you at work, every email from your bank, etc.? There’s simply no point to living in bands if the individuals in it are not going to cooperate.

That cooperation is what liars exploit. The casual liar, the street hustler, can be defeated with time. They’re looking for the quick win. Wait a bit and they’ll lose interest. But the worst kind of liars are patient. Gaslighters, catfishers, and state security services may develop a lie over months or years. Partisans do it over decades, even centuries.

Partisans are liars, or rather liars are partisans. A liar’s interest is herself. A partisan’s interest seems external, but in truth so much of their identity has been leached by the interest that there is effectively no difference, which is why they react so strongly when you criticize. An attack on it is an attack on them.

As an aside, that’s a good test of partisanship. How angry do you get when one of your beliefs is criticized?

There have always been liars, people who would exploit the mechanisms of normal human interaction, but the damage any one of them could do was limited by an evolved response that is as old as us. Norms and mores, more than laws, are the social firewalls and virus checks meant to limit the damage from exploits of the human network. On the one hand, there are clear benefits to openness and cooperation. There's a reason African hunting dogs, which live in large cooperative packs, have the highest success rate of any predator. We accomplish more together.

But cooperation and openness create opportunity for parasitical exploits. Sufferers of the dark triad of personality traits—liars, psychopaths, and narcissists—are exactly such behavioral parasites. As soon as you have the emergent properties of society and culture, you create a habitat for them in exactly the same way that every large herd of herbivores sustains its own population of biting flies.

That means, by the way, that the dark triad is not an illness, like herpes, but rather the normal evolved function of human society and any human population will come to express them. Pick the best people you can to start a colony on Mars, each completely free of the dark triad, and some of their descendants will still be liars, psychopaths, and narcissists. I suspect pedophilia is similar (but for different reasons).

For all of our history, speech and body language were the wires of the human network. It was how our brains passed information, as between nodes of an intranet. Writing radically altered the length of those wires and so permitted the kind of server-client relationship you see in church or state, functions previously mediated by the cult of personality. Neither Xerxes nor Leo III could’ve administered their empires without writing.

But it was printing that created the first human internet. Ideology—which is to say partisanship—requires printing, or some kind of mass media. People shared beliefs before printing, and their brains were biologically the same after. Those things did not change. Rather, printing altered the human network, the way our otherwise unaltered brains are connected. The fitness landscape of meme-space changed, and so some kinds of ideas died out and others evolved to take their place. The same happened with broadcast media.

Those technologies came with built-in immune systems. You had to have access to the very expensive, very physical means of production in order to exploit them. It’s not that they carried more truth. It’s that the lies they carried were basically the same lies that were told before, with a few more competitors.

If anything, mass media strengthened the centrality of power by creating a single point from which to broadcast. I’m not sure the British Empire could’ve administered a quarter of the earth without print. Genghis Khan couldn’t, and neither could Alexander.

Those kinds of big lies are what the world historian William McNeill called “macroparasitism” to distinguish it from plagues and famines and tapeworms, which he collectively called “microparasitism.” His book Plagues and Peoples explained world history as the interplay between the two.

The internet, and the smart phone in particular, also changed the way our brains are networked. Unlike physical media, however, you do not need access to the means of production. In fact, there are almost no limitations to access at all. Unsurprisingly, the biting flies of humanity swarmed in to exploit the virgin meme-space this created. We’re all familiar with the result.

Consider how many large network hacks there were in the 2010s, from Yahoo! to Equifax. Hacks are sort of like “infections” borne through a new type of human network that has no intrinsic “immune response.” (Most hacks use the internet to convince a human to do something rather than use the internet to tunnel through technology, meaning they’re fundamentally exploits of the human rather than the machine network.)

The sudden widespread use of two-factor authentication is a direct “immune response.” Just like the earliest social mammals, we’re trying to reap as many benefits of openness and cooperation as we can, recognizing the need for some selective limits to inhibit exploitation by parasites and free riders.

But as difficult as cybersecurity continues to be—locked in an eternal arms race like parasite and host, predator and prey—it’s still considerably easier than the battle to secure meme-space, which is a key feature of our historical moment.

This should be encouraging. Seriously. You are participating in humanity’s reaction to an entirely novel problem, one that’s only really existed since 2008 and only became epidemic since 2013. We simply haven’t ever faced this before.

There is a real danger that, out of fear and disgust, we’ll sacrifice openness and cooperation for authority and security. The current attempts to outlaw “disinformation” are a manifestation of that, but it’s possible that that angry response is something akin to inflammation and may subside as we get the parasites under control. (But then again, maybe not. Western society may be developing an autoimmune disorder whereby it begins attacking itself.)

Right now, we’re experimenting with solutions. A tiny example is Ground.News. To be clear, I’m not recommending it—although I did subscribe and am checking it out. This isn’t a single solution. There is no single solution. But this TYPE of thing, in conjunction with a constellation of related things, could succeed in retaining the maximum openness possible for a minimum necessary limitation of parasitical exploits.

Every creature gets sick. There are viruses, for example, that target particular species of trees. There’s even a species of parasitical fly that only targets crickets. There is no parasite-free solution. An open society will always have to deal with some number of liars, psychopaths, and narcissists. (Psychologists have recently added a fourth member: sadists. Yay.)

But there is a very good reason why it seems worse now, and there is every reason to expect it will get better.

In the colloquial sense, memes are just viral humor—easy-to-digest, typically visual jokes published on a medium that makes them easy to share, something that did not start with the internet. As soon as we had the printing press, we had “broadsheets,” one-page blends of op-ed and political cartoon that would become the modern newspaper.

One of my favorites, a Cranach woodcut from 1545, features a pair of Protestant peasants—the Occupy movement of the 16th Century—farting on the Pope and his Cardinals.

In fact, it’s an open question whether we would’ve had the Reformation, or the American Revolution, without the printing press. Mass media greases mass movements. How else can the ideas of a German philosopher overturn the 3,000-year dynastic cycle in China in the span on one lifetime?

Print made ideas easier to spread, which is why it provoked such a reaction from people in power. Artists were always a little dangerous, but the audience for their hereticisms had previously been limited to those in attendance. The printing press radically altered both size and scope. Mass-printed materials could transcend borders, language, even literacy.

After the political cartoons of Thomas Nast repeatedly assaulted Boss Tweed and Tammany Hall, Tweed supposedly demanded, “Stop Them Damn Pictures. I don’t care so much what the papers write about me. My constituents can’t read. But, damn it, they can see pictures.”

Across the 19th Century, ever-cheaper mass media, along with ever-easier distribution, became a legitimate political danger. In France, the artist and lithographer Honore Daumier was first fined and then—in an early case of deplatforming—imprisoned for criticizing King Louis-Phillippe, who banned “memes” as harmful misinformation.

As with the new medium of photography, the printing press in its early days was more technical art than business. Printing houses, the startups of their day, repeatedly folded. The only way they made money was by selling expensive hardbound books to the wealthy, who were the only ones who could read anyway.

That began to change in the 1830s, when cheap unbound publications like the penny dreadful could be sold to increasingly literate factory workers, and gradually over the course of the 19th century, coincident with the rise of literacy, printing became less the obsession of artists, intellectuals, and provocateurs like Daumier and more an industry.

By the time the mass market paperback was invented the following century, the term yellow journalism had entered the cultural lexicon. Everyone in orbit of the Soviet world, which dominated much of the globe in the 20th Century, used the newspaper as birdcage-lining. Even in the West, control of mass media was so entrenched by the post-war period that everyone could watch the leader of the free world get assassinated on TV and still have no idea who really did it or why.

Now, as in earlier eras, our major institutions—corporate legacy media, corporate new media, universities, NGOs, the government—are obsessed with fighting memes under the guise of “disinformation.”

To anyone raised in the immediate wake of the Zapruder film, the Pentagon Papers, Watergate, MK-Ultra, and the Church and Pike Commissions, it was just a given that governments lie, typically about the need for war, how it’s going, or whether the “enemy” even exists, which is why the term disinformation has such a different connotation to Gen Xers like me. We collected it! A pair of Amok dispatches are still on the shelf next to my computer.

From their website:

In 1986, emerging out of the creative ferment of LA’s punk rock and noise music community, a collective called Amok was formed dedicated to investigating and disseminating “the extremes of information.” Lacking the funds to actually stock a bookstore, they decided to issue a 30-page combination dada-inspired collage and book catalog dubbed the Amok First Dispatch. The Amok First Dispatch was organized like no other book catalog or bookstore – eschewing traditional genres it set up its own categories: Control, Mayhem, Orgone, Sleaze and others. The Amok First Dispatch became an immediate underground sensation and counted psychedelic visionary Timothy Leary among its first enthusiastic mail order customers.

“Disinformation” was an ironic badge we wore with pride. It meant information not approved to appear in the paper or on the nightly news, information the powerful thought might be dangerous.

It wasn’t that we believed it. It’s that we didn’t believe what was broadcast on those other places.

In the 1980s, if you said you were open-minded about alien visitation, people thought you were insane. Not kooky uncle insane. Horse-paste-eating-Charles-Manson-“put-him-in-a-COVID-camp” insane.

These days, clean-cut Rand-type military engineers are testifying openly before Congress on the very real possibility of alien visitation.

It’s true that disinformation is mostly trash. It’s the landfill where they bury the bodies.

In 1997, Wired magazine posted “Everything You Know Is Wrong” about a website called Disinformation and its creator, Richard Metzger, “who jokingly calls himself the High Priest of Harmful Matter.” (I owned their print collection.) The post describes in near-heroic terms how Metzger had recently moved his operation from New York to Los Angeles “after cable giant TCI pulled the plug.”

Today, Wired routinely supports the same kind of corporate deplatforming of anyone who says anything that someone somewhere might interpret in a potentially dangerous way. (King Louis-Phillipe would approve.)

Something happened between 1997 and now that changed the cultural perception of ideas and information, of memes.

In its original 1976 connotation, a “meme” was the mental equivalent of a gene: the basic unit of an idea, the smallest part that can be transmitted, alone or in nexuses, and so either rise or fall in frequency based on conditions.

In the 1990s, the idea that ideas could evolve breached the public consciousness, no doubt because the internet seemed to offer a way to quantify culture in ways that hadn’t been measurable before. This fascination was memorialized in Susan Blackmore’s 1999 book The Meme Machine, which attempted to establish “memetics” as a science akin to genetics. (Alas, it didn’t get very far.)

But as a meme itself—the meme of memes—that model is not what took hold in the popular culture. By 2004, people were speaking of posts “going viral,” I suspect because, in the popular consciousness at least, evolution invokes a sense of progressive improvement and so was too positive of a term to describe crude, often deliberately insulting humor of memes, like farting peasants. It seems so much edgier when likened to a disease.

Whatever triggered that change—I won’t speculate—it was the contagion model rather than the genetic-evolutionary one that took root with Millennial and younger generations, who have never known a time without a meme of memes.

We talk about “consuming” information—an idea, a meme. But that’s misleading. We don’t consume a meme like we consume a meal. If we did, we’d all be vomiting a lot more—a dozen or more times a day, by my count, as it’s very common for all of us to be exposed to an idea that we reject.

Ideas are not germs either. They don’t function or spread the same. Exposure, even repeated monotonous exposure, is not enough to cause infection, otherwise we’d all be fans of Taylor Swift by now. In fact, social psychologists have known for decades that repeated exposure to an idea you reject actually causes you to become more retrenched against that idea—not despite but especially when it comes attached to evidence! (Humans are a lark.)

Disinformation isn’t a vaccine, but it might be closer to that than a virus.

Information spreads, yes, but so do genes, water, and gravity. The idea that information spreads like a virus is simply wrong. Culture is not a disease.

We don’t have a good sense of what causes people to believe what they believe, except in the simplest of cases—i.e. it comports with what they already believe. In fact, there’s evidence that people’s opinions are not stable over time, that they continually construct and reconstruct them on the fly. (See the work of John Zaller and others.)

And yet, confidence in the contagion model is the basis for the latest public health-style fight against “disinformation.”

But if the idea that information spreads like a virus is wrong, that means—you may want to sit down for this one—the basis for anti-disinformation is itself disinformation. The anti-disinformation crowd ought to be eating itself.

That model in your head—the meme you carry about memes—matters. If it’s more like the genetic-evolutionary model, then you’ll leave room for mutation and selection. Ideas have to be allowed to mix and remix openly in the information ecosystem. We want viewpoint diversity, which is a public good. Without it, there’s no evolution.

If, on the other hand, you believe the contagion model, where bad ideas spread by contact, then you’ll take the public health approach, which is to track and monitor all variants so you can stop them from spreading, including by quarantine and isolation.

On this view, other people are not wrong. They’re sick. And they need to be cured.

To everyone in the industrialized world born before 1990, the shamans of the future presented a vision: the high-tech lowlife.

Technology, it was said, would flow like battery acid over society and dissolve the institutions of cohesion and safety. The state would wither, and with it the protections of the law, to be replaced by giant corporations. (Think OCP buying the city of Detroit in RoboCop or the ruling oligarchy of television networks in the Max Headroom series.)

In the absence of a state, organized crime would explode, along with cities, and the environment would collapse. (Think of the opening scenes of Blade Runner and Akira or Judge Dredd’s Mega-City One.)

The wealthy would flee, often off-world, leaving us to cower under a criminal elite dissatisfied with owning just our labor. Forced to adopt invasive technologies to get ahead or just to survive, our bodies, our human selves, become the very means of post-industrial production. (Think Johnny Mnemonic, When Gravity Fails, Altered Carbon, Ghost in the Shell, Appleseed, or the original Cyberpunk role-playing game.)

It was a compelling vision, so much so that you often still hear comparisons to it, despite that as much of cyberpunk came true as Buck Rogers before.

Hold a picture of a suburban living room or kitchen from, say 1981, the year Johnny Mnemonic was published, in front of an actual suburban living room or kitchen and you’ll see not much has changed. There won’t be a phone tethered to the wall anymore, and TVs are quite a bit thinner, but we had microwaves and remote controls in the 1980s. There’s nothing in your living room or kitchen that would shock anyone from that time. (That’s in your pocket.)

What’s most remarkable about the cyberpunk era is just how durable it’s all been. We think of ourselves as living in a time of great technological change, but since the 1970s, we haven’t had anything like a Trinity Site or moon landing. Other than the slow dawn of the internet, all of the major historical events of the last 50 years have been sociopolitical: the fall of the Berlin Wall, 9/11, the housing crisis, the pandemic—change that could’ve happened in any era.

In that time, technology has weakened some institutions, yes. But it strengthened others—the state in particular. Despite the unanimous predictions of cyberpunk, nations have not withered. Thanks to the centralizing force of the internet, they have power that would’ve driven any Czar mad with envy, power that is largely transparent. We only get glimpses of it through leaks and unauthorized disclosures.

Corporations are bigger than ever, but so is everything, including governments. Only three of the nine largest employers in the world are private companies. The rest are all state enterprises: the People’s Liberation Army, the US Department of Defense, the UK National Health Service, Indian Railways.

Nor are states vassals of corporate power. Of the ten largest corporations in the world, four are state-owned, and the rest routinely kowtow to rights-violating requests from government without so much as a strongly worded letter. Recently, major media companies in the US voluntarily cooperated with the state security services to identify and arrest a government informant, exactly reversing the antagonistic stance they took during the Cold War with the Watergate and Pentagon Papers leaks.

Transnational governmental organizations like the UN, EU, and NATO have also not withered. Indeed, it’s hard to imagine how we might ever get rid of them even after they’ve served their purpose!

The wealthy haven’t fled. The furthest they’ve gotten are a few briefs trips to orbit. They’re still stuck here and masked up with the rest of us during COVID.

Cities have not swallowed entire continents. On the contrary, people are fleeing high-population areas like California and New York in droves.

It’s difficult to have rational conversations about the environment, which has become the New Christianity, but certainly Los Angeles and Las Vegas don’t look anything like the world of Blade Runner. Thanks to several decades of government regulation, there’s actually less smog in LA now than there was when that movie was released.

None of us have put computer jacks in our heads—an impossibility anyway since there is no I/O center in the human brain—or replaced our limbs a la the Bionic Man. We aren’t dodging roving gangs of drug-addled cyborgs on our way to work. In fact, the streets are generally safer than in our youth. Until COVID, crimes rates had been steadily falling.

The only things cyberpunk “got right” were cheap drugs (per capita deaths keep going up and spiked after fentanyl) and the advent of a ubiquitous global data network.

But even there, the shamans of the future got it backward. They assumed we would invade “cyberspace,” like the hackers in Neuromancer, where in truth “cyberspace” is invading the real world.

People don’t generally want to escape into virtual dreams. Hence, the predictable failure of the Metaverse. Rather, they want their dreams to manifest as reality. (If Zuckerberg really wants to own the future, he should make real android companions, not virtual cybernetic ones.)

Already, the nascent technologies of generative AI and 3D printing have begun to instantiate our digital fantasies physically, while cooperative online belief systems like QAnon and DEI remake the living world into non-overlapping alternate tribal realities, something much closer to China Miéville’s 2009 novel The City & the City than anything from cyberpunk.

Of course, all of that is very glib. As a narrative movement, cyberpunk wasn’t so much trying to predict the future as it was trying to say something about its present—and its past. It was as much a reaction to the utopian science fiction of the 1930s-50s as it was eschatological treatise on the post-1960s West. Fair enough.

But that begs the absence of such a movement in our own time. The world is never not a dystopia, but if ours is not a cyberpunk dystopia, what genre of dystopia is it? What is it about life in the 21st Century that our shamans can no longer conjure a coherent vision of the future, which is by reflection a statement on our present? Or is it that after the internet, everyone’s a shaman now—one giant chaotic cacophony that can’t make sense of anything?

Would it make a difference if we knew?

Whatever future appears, it won’t be the one we imagined. History has no genre. The future isn’t a slippery slope present. We can’t harpoon it on a trend line. What actually happens next will confound us, and all of us.

Imagine you have a coworker, Gray. You’ve worked with her long enough that you can’t remember exactly when you met, but you do remember how. Gray was recommended to you by a former colleague, since terminated by the company, as someone worth knowing due to her connections and insights into your gargantuan employer.

Although not an executive, Gray holds a position that affords her knowledge of entire sections of the company that are otherwise opaque to you, despite that you’re roughly peers. Even better, Gray is willing to share at least some of this information with you despite that you have little or nothing to offer in return.

Gray doesn’t work in your department, so you don’t see her every day. In fact, at your busiest times, you don’t see her for weeks. But when you do run into her again—in someone’s office or in a conference room waiting for a meeting to start—you’re reminded why you like catching up. Unlike some of the other people you work with (cough cough), Gray is a professional. She’s smart. She’s hardworking. And while it’s also true that she likes to advertise—at least one name is dropped at every interaction—she never holds your lack of connections against you. Conversation is easy, and you always walk away with the sense that a secret curtain was parted, if only for a moment.

Thanks to Gray, you knew to politely decline when asked to work on a risky but potentially lucrative product line extension that was later canceled. You knew to expect layoffs several years ago and to prepare accordingly (although thankfully you were spared). Also through her, you found out last year that one of the executives in your division was being investigated for financial impropriety, giving you the chance to distance yourself and keep your budget squeaky clean.

Your relationship with Gray seems uniformly positive, which is why you’re happy to see her promoted, even though it means you aren’t really peers anymore. After all, she’s clearly capable and puts in the work, even more than you. Why shouldn’t she be promoted?

You don’t suspect anything at all in fact, until, out of the blue, your former colleague emails you an article from a trade magazine.

“Have you seen this?”

Written by a stock market analyst, the article suggests the product line extension—the one you had been asked to work on—had been scuttled internally, that forces within the company had deliberately denied it resources and talent, and that as such, it never had any real chance of success, and that this was done for individual gain despite the negative impact on shareholder value.

As you read it, a sequence of events that you had experienced as the normal background noise in a company suddenly takes on a much more sinister flavor. You recall offhand jokes others made about the new product team being the best assembly of “special needs talent” the company could find. You remember shaking your head at several obvious missteps that made their way around the company in email chains. You realize that Gray now occupies the role of the man who was blamed and let go.

Still… it seems far fetched to suggest orchestration. Is anyone really that skilled? Maybe it really was your best interests and not hers that led to her recommendation. Perhaps she really was the best person for her new job. You can’t say otherwise, so you go on as you did before.

Months pass. You don’t see Gray as much, but when you’re able to reconnect, things are as they were. She tells you about some organizational changes in the works, and you’re able to bend her ear about a new idea you have and watch it get some traction among the senior staff.

The following year, sitting in the audience at the company’s annual all-employee meeting, you watch Gray take the stage and announce that the executive you knew to be under investigation has left the company “for health reasons” and that she is taking his place. As you listen to her praise him with polite euphemisms, where before she had nothing but recrimination, you get a strange feeling. You look around at the audience, attentively absorbing her words. Surely you’re not the only who thinks this is a little odd.

It isn’t that it’s a big lie. Maybe it isn’t even a lie at all. Maybe the guy really does have health problems. Rather, it’s how easily she’s delivering it, as if it were the most natural thing in the world.

Gray repeats her explanation in the company newsletter, where she also details her vision for the department she now runs. When rumors surface in the press about the same kind of financial impropriety she had mentioned back when, Gray denies them publicly. On the pretense of restoring public confidence, she joins with several other executives to create a special Disinformation Committee dedicated to quelling such damaging rumors—which first requires uncovering their source.

When you finally talk to Gray again and ask how the new job is going, she explains in a wistful tone that she doesn’t have the freedom she had before but that her constraint is noble. She’s a steward of the enterprise now. People’s jobs depend on her, which is why she has to take rumors of impropriety seriously, even though that’s all they were.

That makes you pause. You can’t say it’s not true, but it does seem odd given her obvious intelligence and the detailed knowledge she had about so many other aspects of the company. And if it really was just a rumor, isn’t she just as guilty of spreading it? Why hasn’t she been dragged before the Disinformation Committee? Why does she get a pass?

As you say your goodbyes and walk away, you find yourself mentally holding her statements like a stinky diaper, not just her recent statements but everything from the past. The reason your relationship had been so refreshing was precisely because it wasn’t rumor. You can get that from anyone. Implicitly or not, she held out that she knew better.

Months later, in a meeting involving your department, you hear Gray omit relevant information from a report to senior staff, information that might lead a reasonable person to a different conclusion. As it happens, you agree with her conclusion, but not so strongly that you’re willing to lie for it, especially in this case, where it also happens to benefit her. So you speak up and tactfully fill the gap so that senior management can make an informed decision.

After the meeting, Gray says nothing, but you have to wonder: Have you made an enemy? Partly out of curiosity, partly out of self-preservation, you start paying much closer attention to everything she says, looking for anything you might be able to verify yourself—an effort made all the easier by the company’s fancy new intranet, designed to facilitate cross-department communication and collaboration.

You find a report she released several month back on a major product liability issue, and you download and analyze the data yourself. You find where she’s clashed with peers, and you reach out for their side of the story. Same when the department fails to clear a regulatory hurdle for the first time in its history and some of your coworkers are fired.

Slowly, a pattern emerges. Gray has a habit of omitting anything inconvenient to her cause. In cases where the outcome is beneficial, she has no trouble passing off innuendo as fact. When she makes a mistake, she buries it or else excuses herself for reasons she wouldn’t accept from her enemies—or you. And she does all of that a lot.

No one’s faultless, of course. We all get it wrong sometimes, even when we’re trying not to. And it’s not like the rest of us own up to our mistakes, even to our friends, let alone to the world. In that respect, Gray is no better or worse than anyone else—which is exactly why you don’t trust her more than anyone else. She’s just another opinion, albeit a polished and outspoken one.

Even if she means well—which is questionable, at least—good intentions can’t turn wrong data right. The fact that she’s willing to omit relevant information or confuse supposition with fact—for ANY reasons—means you can’t make responsible decisions based on anything she says. You never know which part of the story is just not true.

You might still listen to her opinion on the restaurant she tried over the weekend or the movie she saw. (She does occasionally have decent taste.) But anything involving your job or the direction of the company—anything of real value, anything that might impact someone’s life or future—is immediately suspect. It has to be.

But she doesn’t see it that way. She can’t. Gray has now picked up on your distrust, and from her tone and behavior, it’s clear she’s having trouble making sense of it. She thought you were one of the good guys. To her, nothing’s changed. She’s doing what she’s always done.

“What happened to you?” she finally asks.

And now you know she sees you as the enemy and it’s only a matter of time before she treats you accordingly.

Currently, there are two large gaps in our understanding:

• Between what we know now and what will actually happen.

• Between what we know now and the common understanding.

The precursor to this piece, “AI is Potentially World-Ending Technology. Now What?”, was meant to address the latter. It provided a concrete practical tool for evaluating the potential risk of AI on the remainder of your working life. However, it did not provide an example, and it seems the application is not obvious.

In the article, we plotted the strengths and weaknesses of AI and ended up with two axes: more inventive versus more repetitive work on one axis and more rule-based versus more principle-based work on the other.

The first misapplication is to apply this to jobs, roles, or occupations. AI does not automate jobs. It automates tasks or job functions within a role. If most of the work you perform in your role is automated, then your job may go away and the remainder of your work moved to a different role, but it is not the role itself that is automated.

The following is not nearly detailed enough, but it is a place to start:

For examples and discussion of what the terms mean, see the original article.

Note, the circles in the chart don’t correspond to any specific distribution. It’s just a way of indicating that job functions may cover a range on the graph. Advertising design, for example, varies significantly in the amount of creativity required, where simple ads can already be created with the help of AI.

The same is true of the brands and campaign messages on which those ads are based. Even before AI, if you were not too particular about your company logo, you could find any number of pre-designed ones on stock image sites like Depositphotos.

We simply don’t know the true valuation of such things in advance, which means they are impossible to price. The recent movie Air, on the creation of the Air Jordan shoe line, suggested that Nike paid a student designer $35 (about $250 today) for their infamous SWOOSH. Who at the time could say that was undervalued?1

What is true of advertising and design is also true of work broadly labeled “art.” There are already online libraries of stock art available. Not all of them compensate their contributors, and even those that do often compensate them trivially, meaning most of the images in commercial use were not individually highly valued.

Recently, I've come across several variations of the statement “Creatives were supposed to be safe from AI, but artists are losing work to these tools!”

Unfortunately, that view relies on two fundamental misunderstandings.

First, it presents the issue in terms of occupation and not task. When it comes to the impact of AI on work, precision matters. You may describe yourself as an artist, but that’s not a task. “Character illustration” or “sculpture” is closer to the mark.

Second, that view misrepresents an artist/illustrator's core value-add. Art (with a capital A) may be price-less—most of what we make in a year couldn't be given away—but for-hire character illustration, for example, is fundamentally transactional.

It’s also largely mechanical. The client is buying a depiction or portrayal of an existing idea—zombie Betty White dressed like Rambo and riding a T-Rex; Mike Tyson as a Jedi knight depicted in the Japanese woodcut style.

The Machine at the End of the World as a frontispiece to an early Renaissance book.

The features of the character to be illustrated are often already determined, which means the creative act is significantly complete by the time the artist applies their mechanical ability to translate a text description into an image.2

That can now be automated.

In the case of the top illustrators, clients are also buying their highly distinctive style, which represents an additional value-add, and in a fair world, those people would continue to be able to charge a premium. But as many have noted, AI models are capable of mimicking any style. Several court cases are underway that will decide whether that counts as copyright infringement.3

It should be possible to prevent large for-profit entities from copying an artist’s style without their permission in the same way that it should be possible to prevent large for-profit entities from using a person’s likeness without their permission. But preventing individuals and smaller groups from doing the same, while technically possible, would require dystopian levels of global oversight, enforcement, and control of the type that could easily be abused.

Even if you removed every artist’s work from every AI training set, automation of the mechanical act of visual portrayal is here, or soon will be. In the original article, we noted that Google’s AlphaZero—which beat the chess program that beat every human—was never trained on chess. It was merely given the rules of the game and asked to play itself for three hours, by which time it learned chess on its own.

In other words, rather than mimicking the art in a data set, AI could treat it as a destination and teach itself perspective, color theory, and the rest. (Isn’t that how human artists learn?) Hence, any court victory will be short-lived.

Long before AI, I asked a tattoo artist—who was then giving me a tattoo—how tattoos were priced. He made the prescient observation that there was really no objective way to say that one person’s depiction of, say, a butterfly was worth more than another’s such that the first gets to charge more. The reality, he suggested, was that most tattoo artists—which was everyone not being written up in the industry magazines—effectively charged by size and complexity as a proxy for time and materials, even if they didn’t think of it that way.

There’s no reason why most tattoos couldn’t be printed under your skin by machine, especially those depicting zombie Betty White riding a T-Rex or any other specific image. In my case, the artist literally traced the image onto my skin and then went over it with the needle. A machine could theoretically do it cheaper, safer, and more accurately. It could also tell you that you were misspelling the word “ragret.”

Right now, AI art models cannot handle editing and revision. But as we demonstrated in the last article, AIs are going from more specific to more general capabilities and at a faster and faster rate. It will not be long before they can edit and remix their own work—perhaps by giving one AI the job of generation and another the job of editing.

That means that as a novelist I will be able to work collaboratively with an AI to create original works of art exactly to my specifications without having to know how to draw, and the machine will be able to suggest improvements based on the empirical data of what has worked well for others in the past. Beyond book covers, maps, and character illustrations, I will be able to create my own comics.

Eventually, my own movies, completely without cameras, actors, or props.

So, while it is generally true to say that “machines are not creative,” it is not true that that means artists’ jobs are “safe.” The creative act is not wholly or even mostly in the mechanical manifestation of an image. It can be, but most of the time, the act of depiction is a translation from one medium—a text or speech description from an art director, for example—into another—ink and paper, or even sculpture.

Chinese artist Zhelong Zhou creates absolutely beautiful digital sculptures. The ability to 3D-print them in comparable materials isn’t here yet, but it will be. In the meantime, we can 3D-print casts that can be use to make them, meaning its getting easier and easier to make our fantasies real.

That, by the way, is one of the key ways we are not manifesting a cyberpunk dystopia but an opposite kind. The cyberpunk writers got it backwards. Humans are not invading cyberspace. Cyberspace is invading the real world. Narrative warfare will become indistinguishable from physical warfare. Winners always wrote history. Now they will determine reality itself.

Share

Not every human is creative, but many more are than have the physical skill to draw. Artists are now competing with all of them in the same way that the invention of the chainsaw meant many more people could be lumberjacks.

That pattern, by the way—where technology lowers the barriers to entry, creating a flood of new entrants that drives down earnings for all but a very elite group of massive winners—has already happened in publishing and the recording industry, and new waves of it are coming again to those industries, and to film, thanks to AI.

The data is not all bad. Economic rebounds do seem possible. And the long-term result of all that upheaval is that more people are making music and writing novels than ever before. (It’s also worth noting that at the peak of earnings, the recording industry was found to be price fixing. In other words, they were cheating consumers. Actual revenue should never have been that high.)

So, too, with the economy as a whole, where the middle class is shrinking faster than the Colorado river and the world’s richest individuals are worth more than the GDP of most nations. This is why AI is such a seismic shift. As machines can do more and more of the work, how will any of us make money?

It’s not that there isn’t a way. It’s that right now we have no idea.

A recent survey by the Pew Center suggested that a majority of people thought AI would broadly impact the world of work, but a distinct minority thought it would impact THEIR work. Many, many, many of us, it seems, will be echoing our artist friends. “I was told this wouldn’t happen to me!”

Subscribe now

It’s worth thinking about now.

The first problem.

We can’t even clarify the extent of the threat. In its weakest form, AI will be a nuisance, a kind of bureaucratic evil of the type peddled by the demon Crowley in “Good Omens.” At its worst, it's potentially world-ending the same way nuclear weapons are potentially world-ending.

Between those poles are a continuum of options that no one can wholly grasp.1 Most treatments of the subject, however, will defend or critique one mode only—AI’s effect on the world of work, for example—and come to a conclusion that is supposed to settle the matter for all.

The second problem.

There’s no agreement among the experts. The 2018 ACM Turing Award was shared by three long-time collaborators known as the “Godfathers of AI”: Yoshua Bengio, Geoffrey Hinton, and Yann LeCun.

After endorsing a pause on advanced AI capabilities research, Bengio said “Our ability to understand what could go wrong with very powerful A.I. systems is very weak.”

After resigning his position at Google, Hinton said in May “It could figure out how to kill humans” but “it's not clear to me that we can solve this problem.”

After hearing his colleagues concerns, LeCun said “You know what's unethical? Scaring people with made-up risks of a technology that is both useful and beneficial.”

If these three men—not only world-renowned experts in the field but also long-time colleagues and close collaborators—can’t agree, there’s not much hope for a near-term synthesis.2

That puts the onus on us, where most of us are simply not equipped, especially in this information ecosystem.

A few years back, we were promised that blockchain was going to revolutionize banking, healthcare, and education. It was going to balance your checkbook for you, fold your laundry, wash your car, and cook you dinner every night.

How’s that working out?

Further back we had Y2K. For those of you not old enough to remember, we were told that airplanes were literally going to fall out of the sky onto school buses full of children.3

Or what about “the paperless office,” which has been just around the corner for the last 40 years?

If you’re skeptical that AI is any different than these, I understand. As a humanities major who finds himself working in technology, I can attest to a great deal of misunderstanding on both sides. Technology is not the savior that many in the ’90s thought it would be. It’s more like the New Oil. But the technical class seems collectively to get most (or all) of their self-esteem from their work, and they still want to believe they’re the good guys and that technology is a kind of amulet or talisman that makes us more virtuous just by putting it on, and they keep marching us into new nightmares as a result.4

But AI is not blockchain or Y2K or “the paperless office” or any of the dozens of other false promises, and you don’t have to know what a Generative Adversarial Network is to understand why that’s the case. You just have to understand games.

In 1997, IBM’s DeepBlue beat Garry Kasparov at chess. These days, that almost seems trivial, but keep in mind that at the time, most of the world ran on Windows 95, which came in a box of 13 floppy discs that had to be inserted one at a time during an hour-long load.

DeepBlue required some specialty hardware, meaning it wasn’t wholly platform-independent, and it was programmed specifically to play chess, sort of like a specialty chess calculator. It was intelligent the way a calculator is intelligent. (Calculators can do math much better than I can.) DeepBlue could react to novel inputs, but it did so according to its programming. Any changes or updates had to be coded by its human operators.5

That was not the case with Stockfish, a chess program released eleven years later in 2008. Kasparov beat DeepBlue once before it returned the favor. No human has ever beaten Stockfish.

Unlike DeepBlue, Stockfish could run on any machine that met its minimum requirements. Also unlike DeepBlue, Stockfish was an early example of machine learning. It wasn’t limited to the decision tree provided by its creators. The more games Stockfish played, the more it added to its own dataset, meaning it could get better with experience.

And yet, Stockfish was still specifically a chess-playing machine. It was programmed with most of the history of chess, along with the mathematics of chess, such as the theory of opening moves, and while it could learn to be better at chess, it couldn’t learn how to cook or even to play games that aren’t chess.

Seven years later, in 2015, Google’s AlphaGo beat the best human players at another game, Go. The following year, a kind of AI engineering tool called a transformer was invented, and a year after that, Google updated AlphaGo to make AlphaZero, which beat Stockfish at chess. In fact, Stockfish has never beaten AlphaZero.

“So what?” you say. “So one program is faster than another program. Big deal.”

Except it’s not about speed. Nor is it even about chess. When they turned it on, the engineers at Google hadn’t taught AlphaZero a single thing about chess, except the rules. They simply had it play itself for three hours. Over the course of a trillion games, it learned chess better than the best chess program, which was already better than the best humans.

Nor is AlphaZero confined to chess and Go. According to ChatGPT,6 no human has beaten AlphaZero at anything.

So, in the span of two decades, we went from one machine able to beat humans at one game to multiple machines able to beat humans at multiple games to machines able to beat humans at every game.

This is indicative of the trend. AIs are going from the more specific to the more general… and they’re doing it at an ever-faster rate. It only took ChatGPT five years to compose original, meaningful responses to any text prompt, including recipes and computer code. You can even engage it in a rap battle in Japanese and it will throw down some rhymes.

What our colleagues in the industry are trying to tell us is that we’ve reached the inflection point of this technology. Note that the graph below is not linear but log-scale and that we’re adding more modalities that are operating with increasing generality.

If the headlines are to be believed, the threat is dire.

But skepticism of corporate pronouncements is always warranted. One should be particularly suspicious any time the Chamber of Commerce calls for government regulation.

I suspect, following the writers at Forbes and other places, that recent actions and testimony by business leaders is, first, an attempt to ride the wave of unexpected publicity, and second, an attempt to head off the commercial threat of open source. The business sector can’t swim in pools of cash, a la Scrooge McDuck, if your friends and neighbors give you this technology for free. Requiring governmental review and licensure raises the barriers to entry such that corporations (they hope) will be the only ones, outside of a few fringe players, that will be able to monetize the technology.7

A journalist and author that I respect recently dismissed the threat of AI with the observation that “These things are not remotely conscious” (or something like that). Of course they're not. That's like saying nuclear weapons aren't a threat because they're not conscious either.

For a generation raised on the Terminator films, or Avengers: Age of Ultron, it’s seemingly impossible to imagine a machine threat that isn’t malevolent. But AI is not malevolent—or at least we have no evidence that it is. It simply has no empathy (because it has no anything).

We think in terms of others’ intentions, what someone meant to do. AI computes a path to its goal through correlations in its data set. It doesn’t even know that one thing causes another. That requires insight. The machine simply notes a correlation—when one thing happens, another thing follows—and exploits it to reach a goal.

Even where it’s “smart” enough to apply social norms, AI will flout them to serve its own ends, just like a sociopath. Take the recent example of an AI that was given a task and a modest budget by researchers testing its abilities. It literally got stumped by a captcha. But it was smart enough to know that humans could solve captchas and that humans did things for money. Since it had some money, it went on Taskrabbit and tried to hire a person.

The scary part is that it knew humans might not help it if they knew it was an AI, so it invented a plausible lie: It said it was a visually impaired person who couldn’t answer a captcha on their own.

The AI didn’t do that with malice toward the disabled community. A machine has no malice. It has no remorse. It has no empathy. It had a problem, and it solved it with no idea or care for who might get hurt in the process.

In another recent example, a chatbot named Eliza convinced a Belgian man to kill himself. The US military is concerned that an AI flying a drone might realize that its human operator could recall it, which it could interpret as a threat to its mission objective and eliminate its operator as a result.8

Think about those kind of behaviors at scale:

1. AI isn’t merely capable of learning. It can teach itself, meaning its capabilities are not bounded by our capabilities.

2. AI can immeasurably surpass our abilities and has already done so in some domains.

3. AI is increasingly going from more specific capabilities to more general capabilities in more domains and at an ever-faster rate.

4. AI is not malicious. Nor is it benign. It has no empathy and would not hesitate to save humans or eradicate us if it furthered even a trivial objective.

5. At the point a general intelligence immeasurably surpasses us, it could save or eradicate us in ways we are literally incapable of imagining.

When AI experts like Eliezer Yudkowsky say AIs will kill us, they aren’t saying AIs are Ultron or SkyNet and that once we flip the switch, they’ll immediately turn hostile. On the contrary, they’ll probably be very helpful. They may even provide novel solutions to some of our more intractable problems.

Bear in mind also that in the movies there is always just one AI. There won’t be just one. Or dozens. Or hundreds. We’ll soon be able to deploy these on our home computers, at which point they’ll be effectively uncountable.

How long before one of them decides, without malice, that we’re an impediment?

The first question.

*If* AI is potentially world-ending, it's not obviously potentially world-ending the way nuclear weapons are obviously potentially world-ending. Even in the face of obvious world-ending technology, we still spent nearly two decades pursuing nuclear brinkmanship before we even started to think seriously about deterrence, and it was decades more before we achieved it—only to piss it away last year.

The growth curve for AI appears steeper than that for nuclear weapons, which was already very steep.9 In other words, *if* AI is potentially world-ending, it’s already too late to pursue deterrence before the threat appears.

The same was true of nuclear weapons, which the world had no knowledge of until they were used. The current predicament might be similarly surmountable if humans are often an impediment to AI. That sounds counterintuitive. Wouldn’t we have a greater chance of survival if we were rarely an impediment?

I don’t think so. If we are often an impediment, then it’s likely we will be an impediment as AIs scale over the coming years, and there will be numerous examples of non-superintelligent AIs removing (or trying to remove) a human obstacle. In the face of such examples, we might come together and attempt a species-level deterrent of the type we developed for nuclear weapons.

Of course, nothing guarantees it will work, but I suspect we’ll at least have a window for action.

The second question.

How can we design and deploy a deterrent against an army of machines that are smarter than us? This is a bit like asking a 12-year-old to come up with a way to trap US Special Forces against their wishes and best efforts to escape—indefinitely. They only have to escape once to win, whereas the kid has to keep them trapped forever. What are the odds she’ll succeed?10

In AI research, this is called the Alignment Problem. How do you make sure that, once you turn it on, your superintelligent machine will continue to do what you want, or at least not kill you?

No one has a solution.

Of course, AIs are not only potentially world-ending. They’re also potentially job-ending. Here, however, we have pretty good visibility. After all, we know what AIs are both good and bad at.

Like all machines, they’re great at repetition. They’re terrible at creativity and invention. Machines struggle to connect concepts whose relationships haven't already been mapped, directly or indirectly. All of the art and text they create today is not true creativity but mimicry. One day, AIs will be the world’s best cover bands.

AIs are good at following rules, which is why they’re getting very good at parsing language. Languages have many rules: spelling, grammar, syntax. Laws and legal procedures are rules. Safety regulations are rules. Accounting principles are rules. Sports and games have rules of play.

We’ve spent most of the modern era filling the world with rules, which is why so much of it is comprehensible to AI.

AIs are bad at principle-based reasoning. Principles are not rules but guidelines that shape our approach to a problem and provide a framework for decision-making. They are flexible, contextual, and high-level, and unlike rules, an outcome is not specified. The Geneva Conventions are principles, as is the Hippocratic Oath, the IEEE Code of Ethics, and The Belmont Report, which governs the use of human test subjects in biomedical research.

If we put these strengths and weaknesses together, we end up with two axes, with more inventive versus more repetitive work on one axis and more rule-based versus more principle-based work on the second axis.

The first misapplication is to apply this to jobs, roles, or occupations. AI does not automate jobs. It automates tasks or job functions within a role. If most of the work you perform in your role is automated, then your job may go away and the remainder of your work moved to a different role, but it is not the role itself that is automated.

Tasks that are both repetitive and rule-based, such as coding, bookkeeping, and blog writing, are excellent candidates for AI automation. In fact, that is already in progress and will accelerate.

Tasks that are repetitive but principle-based, such as teaching and therapy, are good candidates for disruption. In other words, HOW we do the work will change, probably in non-obvious ways, and that shift may favor some over others.

Tasks that are rule-based but creative, such as game design and litigation, will likely be augmented by AI. Litigators, for example, will be able to ask AI for a summary of relevant case law along with a deep dive into certain rulings and get it instantaneously. An AI will be able to assist with juror selection and review opening or closing remarks and estimate how effective they might be.

Tasks that are creative and principle-based, such as architecture and fashion design, will be complemented by AI. A fashion designer, for example, might sketch an idea on a screen and ask an AI for 50 variations. The machine can’t create the idea, but it can remix it. The designer may choose five from the 50 and ask for three more variations of each. Even if she doesn’t like them, they may give her new ideas. Here, AI is augmentative because no one is doing that work currently, and it will greatly enhance speed and creativity.

Looking across a variety of job functions, we find something like the following:

The circles don’t correspond to any specific distribution. It’s just a way of indicating that job functions may cover a range on the graph. Advertising design, for example, varies significantly in the amount of creativity required, where simple ads can already be created with the help of AI.

It’s important to note that this grid doesn’t speak to timing. Just because a job function might be at risk doesn’t imply it’s at immediate risk. Industries that are further behind in their adoption of computer technology, such as forestry, will have further to go before they suffer the benefits of AI. However, the grid gives you a good rubric for thinking about how AI might impact you over the remainder of your working life.

Critics are quick to point out that anxiety over automation is nothing new. We’ve had several centuries of it, in fact, and there are mounds of scholarly papers demonstrating that automation doesn’t result in a net elimination of jobs. Rather, it creates whole new ones—sometimes entirely new sectors of work—that didn’t exist before and which would’ve been impossible for people to imagine in the years prior to the change. The death of the phone book resulted in the birth of the digital marketing industry and a bunch of jobs no one could imagine in the 1980s such as influencers and SEO experts.

All of that is true. But there are several potential differences between AI automation and earlier waves.

For one, the pace. Earlier disruptions had cycle times in excess of a human working life. A canal hauler might lose his job to the railroad. But wherever he settled, it was very unlikely that his new work would again be displaced in his lifetime.

That is no longer the case. At any given point in time, AIs and other emerging technologies will not be distributed evenly across industries. It’s entirely possible that workers suffering AI-fueled displacement will move to roles that will also be displaced in short order.

Humans are not infinitely flexible. It’s not always possible for the breadwinner of a family, previously earning a mid-career salary, to take an entry-level position in a new industry where she’s competing with young people who don’t have her financial obligations. One such change might be absorbed. Two or more would likely leave her unable to both support her family and save for retirement.

(As usual, the corporations reap the efficiency but leave the burden to the rest of us.)

More than the pace of change, I’m concerned about the breadth of it. It’s true that earlier waves of automation opened new modes of work. But those jobs tended to be at similar skill levels. Some were even lower. As a general rule, it takes more diverse skills to operate a farm than it does to tend a single piece of machinery in a factory.

AI is not like a mill-powered loom in that a loom can never be anything but a loom, whereas once we have an AI that can fulfill a low-skill position, it can be trained to fill other low-skill positions in the same way that AlphaZero can teach itself chess and then every other game.

Recently, at a restaurant in Kansas City, my food was delivered by robot. One of my neighbors has a robotic lawnmower. My mother has a robotic vacuum. The grocery store where I shop has a robotic floor cleaner. These technologies are all at an early stage, but they will advance and intersect with generally intelligent control systems.

AI is not automating certain roles in an industry, as in all earlier waves. It has the potential to automate the bulk of an entire labor sector across ALL industries. Only a fraction of the labor force that loses their jobs in food service or lawn care will be able to be retrained as robot technicians. (Especially if all those roles are filled by out-of-work software engineers!) What do we do with all those superfluous to need?

Even among “safe” work, if the bulk of the entry-level tasks are automated, how do we train the next generation to be competent at the socially necessary high-skilled variant without the ability to work their way up?

Then there are the second-order effects. Even if trial law isn’t directly affected by AI, if it’s perceived as a “safe” job, more young people may choose it as a career. That, plus an influx of refugees from other white collar industries, may limit growth options for trial lawyers and put downward pressure on wages.

Besides economic migration across industries, there’s also economic migration across borders. The AI revolution will not affect all countries the same. Some will be much harder hit, a fact that is bound to strain the geopolitical order.

It’s possible of course that this round of automation will follow the historical trend. I’d like to see that happen. But it’s simply not accurate or fair to suggest that nothing is different. Maintaining mass employment would require something difficult to imagine indeed: that the AI revolution will open new sectors of non-repetitive work that are accessible to a low-skill human but not accessible to a high-skill machine that works for pennies and never gets sick.

Do we rely on hope?

You might be asking how this is an optimist’s view. I didn’t say it was optimistic, just that I am an optimist. I believe that problems are generally soluble and that human ingenuity can generally solve them. However, we rarely act before the situation is dire.

People died building the pyramids and railroads and skyscrapers. Shipwrecks and plane crashes used to be regular occurrences. Machines once regularly disabled their operators, who included children. In 2022, there were 42,800 traffic fatalities in the US alone. Many more people were seriously injured.

I’m not saying those sacrifices on altar of progress were justified. Not at all. Just that that’s the pattern with every innovation, and we’re crazy to think AI will be harmless. People will die as a result of this technology.

But it’s an open question how many. We have not (yet) destroyed ourselves with nuclear weapons. The best case is that the time required to develop generally intelligent machines is longer than the time required for us to take the matter seriously.

I asked ChatGPT to write a prompt for Dall-E to create the cover art to this piece. The result reminds me of the kind of child’s drawing you might put on your refrigerator.

This child will grow up fast.

A forest wears all the faces of life. Old stand side-by-side with young and both over the seedling and the dead. They are all there together, and there is room for all kinds. Trees, even of different species, do not compete like men. They could thrust their branches into each other like spears, but they don’t. They stop inches apart, leaving gaps for light to shine.

We used to live like that, like the trees. In a village, the very old and the very young were always at hand, and everyone stood guard over the newborn and the dead, who were buried underfoot.

Now, both children and trees are relegated to nurseries and the elderly to retirement homes or hospices. Unless one makes a point to keep them, it’s entirely possible to go weeks without encountering either.

Mankind makes a monoculture. We burn the forest to plant lines of palm oil or corn or sugar cane. In that same way, we sow neat rows of students, each kept to a class and away from the working-age adults, who aren’t allowed to encounter anyone in their toil. We segregate rich from poor, men from women, boys from girls, and people of color “for safety.”

A monoculture of humans—like with like in long, evenly-spaced rows.

The theory is that lets each of us thrive. Each gets equal sun. Each gets equal water. But is that really why? Do we plant corn that way because we dearly want each plant to thrive… or because it’s easier for the machines to harvest the fruits of their labor?

And so for us.

What would it be to live like the forests again? I don’t think any of us wants to. It’s not in our nature to tolerate our neighbors like it is for trees. We want the monoculture. We don’t want to deal with other people’s children. We don’t want to suffer their distasteful opinions. We don’t want to waste an easy Sunday morning burying their dead.

Life is much easier when, like a field of corn, everyone stays in their lane.

Last month’s confiscations can be found here. This month’s sample selection comes from Robin Pedley’s “The Construction of Durable Echo Chambers: A Primer for Teens.”

To keep your students safe, you must protect them from their mistakes. You must train them in the iron law of cause and effect, and that if one pushes back hard enough, every failure is the result of the words and deeds of others long dead. If a student has failed, it is because of the purple people. If the student is purple, it is because of the greens. But one must be careful. Students must not learn to equate color with fear but rather with disgust. One might find the courage to confront one’s fears, but disgust has no champion. The greens must become so wretched that they cannot even be pitied. Their every success must be a threat, their every tragedy a cause for jest. Nothing a green says can ever be good. If they appear to speak the truth, it must be understood as a trick.

You cannot achieve this directly. Rather, you must teach your students that any symbols of the greens—their flags, their speech, their gestures—contain hidden subtext, foreign to them, that cannot be transgressed upon pain of violence or ostracism. Furthermore, these symbols must be so monolithic, so impenetrable, their owners so deplorable, that students become incapacitated with disgust and are unable to function outside the walls of their mistrust, which you have erected to spare them from all that would do them harm.

art: Fletcher Hanks

His creations include Stardust the Super Wizard, Tabu the Wizard of the Jungle, Big Red McLane, and Fantomah—one of the first female superheroes, predating Wonder Woman. In all he created 51 stories.

Hanks had four children. Family members described him as abusive and alcoholic. He made money painting murals for the wealthy, and spent the money on alcohol for himself and his friends. His son Fletcher Jr. worked odd jobs to support the family. In 1930, he found his earnings missing, along with his father. Fletcher Jr. said his mother responded, “It's a small price to pay to be rid of the bum.”

Hanks left the comic book industry in 1941. The reason is still unknown. His body was found on a park bench in Manhattan in 1976, having frozen to death. He died penniless.

A cult following developed around Hanks' work. His stories and art have been reprinted in the magazine Raw and in two books by Fantagraphics: I Shall Destroy All the Civilized Planets! (2007) and You Shall Die by Your Own Evil Creation! (2009).

• Death by Axe Body Spray.

  Many teenage boys suffer from excessive use of deodorant sprays. Unfortunately, for Jonathan Capewell, personal hygiene was a tragic obsession. His persistent use built toxic levels of butane and propane in his system, triggering a heart attack in 1998. He was 16.

• Death by Sorcery.

  A year later, all 11 members of the Bena Tshadi football club died on the field after a series of lightning strikes in Eastern Kasai, Democratic Republic of the Congo, where use of sorcery and talismans against sporting opponents is common. The other team left the field unharmed.

• Death by Time Zone.

  Four Palestinian terrorists, apparently protesting a newly-signed Israeli-Palestinian agreement, attempted the simultaneous detonation of multiple car bombs at 6 pm Israeli time, but due to a change in Israel's time zone, they were an hour behind schedule, and all were killed when their bombs detonated en route.

• Death by Cannibal.

  In 2001, Bernd Brandes, a German engineer from Berlin, responded to an internet advertisement placed by cannibal Armin Meiwes, who was looking for something to eat. On March 9th, the pair met and Brandes was willingly slaughtered so that he could be butchered and eaten by Meiwes, who became a vegetarian in prison.

• Death by Bestiality I.

  In 2005, near Enumclaw, Washington, a Boeing engineer named Kenneth Pinyan died of a ruptured colon after receiving anal sex from a stallion. Pinyan and two accomplices had been filming their escapades with animals and posting them online.

• Death by Bestiality II.

  In 2008, A 43-year-old Irish mother of four died of an allergic reaction after having sex with a German Shepherd. The animal’s owner, Sean McDonnell, met the woman in a bestiality chat room.

• Death by Irony I.

  Jimi Heselden, then-owner of Segway Inc., died in 2010 after accidentally riding his Segway Personal Transport System off a cliff in Thorp Arch, England.

• Death by Cockroach.

  In 2012, 32-year-old Edward Archbold of West Palm Beach, Florida, choked to death while participating in a cockroach eating contest.

• Death by iPhone I.

  In 2013, Roger Mirro was crushed by a trash compactor while digging through a dumpster for his phone.

• Death by Doughnut.

  Hayato Tsuruta, a 28 year-old with severe mental disabilities and a killer sweet tooth, ran away from his care facility to a nearby supermarket, where he consumed so many doughnuts that he died.

• Death by Meteor.

  In 2016, a 40-year-old Indian bus driver named V. Kamaraj was struck out of the blue by a meteorite and killed. Although the cause of death was contested by NASA, analysis of the body showed the presence of carbonaceous chondrites, a meteoric material.

• Death by WTF I.

  That same year, 89-year-old Robert Dreyer crashed his car into a fire hydrant and was swallowed by the hidden sinkhole underneath.

• Death by Irony II.

  In 2017, the aptly named Rebecca Burger, a 33-year-old fitness blogger and model, died after a pressurized canister of whipped cream exploded and struck her in the chest, causing cardiac arrest.

• Death by Delay I.

  In 2018, 29-year-old Sam Ballard died from angiostrongyliasis, a roundworm infection he contracted after eating a garden slug on a dare eight years earlier.

• Death by Social Media I.

  In January 2019, a Sri Lankan man was trampled to death while filming himself attempting to hypnotize an elephant.

• Death by Social Media II.

  In order to record a gender reveal worthy of posting online, members of the Kreimeyer family of Marion County, Iowa, filled a steel umbrella stand with gunpowder, but instead of emitting a shower of sparks as intended, the device acted as a pipe bomb, killing 56-year-old Pamela Kreimeyer, who was struck in the head by debris.

• Death by Social Media III.

  Tetsu Shiohara, a hiker from Tokyo, attempted to livestream an off-season ascent of Mount Fuji only to fall to his death meters from the summit. Shortly before the fatal accident, he told his audience that 'this part is dangerous' but that it 'can't be helped, though'. He then lost his footing and slid down a 30-degree incline and over the side of the mountain. Shiohara's camera captured the fall.

• Death by Narrative Warfare I.

  On August 10th, 2019, the body of convicted sex offender Jeffrey Epstein was found in his cell at the Metropolitan Correctional Center in New York, NY. The day before his death, Epstein's cellmate was transferred out, but no one took his place. The evening of Epstein's death, the two cameras in front of his cell began malfunctioning. Epstein was also supposed to be checked every 30 minutes, but the two guards assigned to his unit fell asleep and did not check on him for three hours. They admitted later to falsifying records and received probation. His death was ruled a suicide.

• Death by Delay II.

  Also in 2019, the decomposed body of 25-year-old Larry Ely Murillo-Moncada, a supermarket employee from Council Bluffs, Iowa, was discovered behind a cooler, where he had become trapped ten years earlier, in 2009.

• Death by WTF II.

  On January 14th, 2020, 69-year-old Sergio Millán was alone in his apartment in Torreforta, Spain, when an explosion in a petrochemical plant three kilometers away launched a one-ton iron plate into the sky. The plate landed in the apartment above him, causing the ceiling to collapse, killing him.

• Death by Lockdown.

  A 54-year-old man from Massachusetts died in 2020 after eating a bag and a half of black licorice every day for a few weeks. Licorice contains glycyrrhizinic acid, which along with its metabolite glycyrrhetinic acid mimics the hormone aldosterone, which regulates potassium levels. Excessive consumption resulted in over-excretion of potassium and cardiac arrest. A paper titled "Licorice abuse: time to send a warning message" appeared in a 2012 issue of the journal Therapeutic Advances in Endocrinology and Metabolism.

• Death by WTF III.

  On August 4th, 2020, a large amount of ammonium nitrate stored at the port of Beirut exploded, causing at least 207 deaths and 7,500 injuries and leaving an estimated 300,000 people homeless. A cargo of 2,750 tons, equivalent to 1.1 kilotons of TNT, had been confiscated from an abandoned ship and stored in a warehouse without proper safety measures for six years. The explosion was preceded by a fire, but the exact cause of the detonation is still under investigation.

• Death by iPhone II.

  In May 2021, the decomposing body of a 39-year-old man was found wedged inside a papier-mâché statue of a stegosaurus in the town of Santa Coloma de Gramenet, Spain. A police representative theorized that the man crawled head-first into the statue to retrieve a lost mobile phone before his leg became stuck, trapping him.

• Death by Narrative Warfare II.

  On June 23rd, 2021, British-American computer programmer, entrepreneur, and political activist John McAfee was found dead in a prison cell near Barcelona shortly after his extradition to the United States on tax evasion charges was authorized by the Spanish National Court. Nine years earlier, in 2012, McAfee's property in Belize was raided by the national police, who arrested McAfee on suspicion of drug manufacturing and possession of an unlicensed weapon. He was released without charge. In November of that year, Belize police sought McAfee as a "person of interest" in connection with the murder of his neighbor, American expatriate Gregory Faull, who was found dead of a gunshot wound at his home on Ambergris Caye. After fleeing the country, McAfee told Wired magazine that he has always been afraid police would kill him. On November 30th, 2019, he made the following tweet. His death was reported to be a suicide.

This month’s sample selection comes from the Fourth Apoplexy of Bell & Tyrwhitt’s “Unspeakable Dread.”

Where am I? Where am I going? Nowhere. I am going to this meeting, but it is nowhere. I do not want to go to this meeting. I do not want to see these people. I do not want to pretend to care about all of the things they pretend to care about. They pretend to care and I pretend to care about their caring and we both pretend we don’t know the other is pretending. She pretends to make a joke and I pretend to laugh. He pretends to think and I pretend he isn’t thoughtless. All of us pretend all of this matters when we all know it does not matter, and we all know everyone else knows it does not matter, but we all go on pretending. How does it stop? Perhaps if I scream. Perhaps if I scream right here, right now. One long wail, as a Viking over a fallen comrade or a Hun over his steed. Perhaps if I dropped to my knees, right here in this office, and raised my hands and wailed at the death of us, if I beat my chest and growled unutterable truths, maybe these people would wake, and if they woke, maybe I would, too, and we could be happy, instead of sitting in this meeting, counting the drops of our souls.

But I do not wail. I do not beat my chest. I smile, and another part of me dies.

That’s all for this month. See you next time.

In 1994, 17-year-old Boy Scout David Hahn raided junk yards in Michigan for trace radioactive materials: americium from smoke detectors, thorium from camping lanterns, radium from clocks, tritium from gun sights. In a shed in the back of his house, he used a Bunsen burner and the lithium he’d collected from a thousand dollars’ worth of rechargeable batteries to purify the thorium ash, which he added to a bored-out block of lead in the hopes of making a breeder reactor, a reactor that produces more fissile material than it uses, meaning it can be used to seed other reactors.

The FBI found him before he finished the project, but he did make a functional neutron source, which is the first step.

Five years later, a pair of University of Chicago graduate students produced a working ‘reactor in a shed,’ which was one of the items on the university’s annual scavenger hunt. They were later able to isolate trace amounts of plutonium.

We love it when the guy tinkering in his garage is Steve Jobs or Carroll Shelby (the automotive engineer played by Matt Damon in the recent film Ford v. Ferrari). But most of the time, it’s gonna be a precocious kid doing it for the challenge and the danger.

People like David are already implanting home-built electronics in themselves and modifying their DNA with CRISPR tools, and their number and capabilities will only increase as technology gets cheaper and easier to access.

Welcome to the Future.

I once read a magazine article from the 1930s that predicted news and entertainment—by which they meant printed newspapers, books, and magazines—would one day be piped to every home in the US via pneumatic tubes, which were then being used to deliver the mail inside large skyscrapers.

This was not as wrong as it sounds. I suspect whoever wrote the article understood the long-term trend of information delivery, that it would eventually be personal, and in a real sense, their prediction was right. Here in the future, news and entertainment is piped directly to everyone—just not by pneumatic tube.

To be fair, it wasn’t that they failed to appreciate electronic transmission. Already by the 1930s we had the telegraph and radio. Nor was it the absence of the computer. Even as late as the 1980s, when we knew computers would eventually be portable, would-be futurists were predicting that every home would have a fax machine to print the newspaper in place of costly home delivery.

We’re good at extending trend lines to the future. We’re bad at predicting the intersection of things—in the case of what you’re doing this very second, the intersection of electricity, radio, and the semiconductor.

That intersection of technologies is what produces the complexity of real outcomes and makes prediction so hard. Even as late as the 1990s, with the internet right in front of us, very few people had a clear sense of what the smartphone would do in less than a decade. Who can say how AI, 3D printing, genetic modification, robotics, nanotechnology, and other emerging tools will interact with each other in just the next few election cycles?

The following short segment of a recent interview with thinker and polymath Daniel Schmactenberger is a fantastic summary of the fundamental tension I'm getting at with Science Crimes Division—what Schmactenberger calls “the Metacrisis” of falling into one of “the two attractors.”

On the one hand, as advanced technology gets cheaper and cheaper, it will put catastrophic power in the hands of more and more people. Think of a basement nuke or someone bioengineering a global pathogen in their garage. That's bad.

On the other hand, once you empower any agency (governmental or otherwise) to stop such people, you pretty much guarantee a centralized repressive dystopia.

Those are the attractors society is currently stumbling between: chaos or oppression. (Probably the first followed by the second, since humans as a species widely prefer security over autonomy.)

I have set the link to start at the appropriate time. You only need to watch a couple minutes.

Schmachtenberger is describing exactly what Science Crimes Division is all about, the need for a “science and technology police” on the one hand and the existential threat to democracy that such an organization would pose.

The challenge always is educating the reader in an entertaining way, because none of this is stuff you cover in school. It's always nice to see I'm not the only one thinking of these kinds of things.

That’s it for this time. I’m glad you’re here.

—from "The Rings of Saturn" (1998) by W.G. Sebald

It’s not clear from the passage, but Sebald was paraphrasing a 1658 commentary on burial urns by physician and Renaissance man Thomas Browne. Actually, 'paraphrasing' might be generous. Unless one is familiar with the source material, it is impossible to know in The Rings of Saturn where it is Sebald and where it is the others he encounters, alive or dead, on his walking tour of the south English coast. But then that's part of his point: culture as ceaseless borrowing and re-inscription.

In other words, The Rings of Saturn turns plagiarism into art. The book is a kind of literary version of Warhol's “Campbell's Soup Cans.” At one point, Sebald spends the better part of three pages directly summarizing Borges' short story Tlön, Uqbar, Orbis Tertius. The report on Thomas Browne fills most of the first chapter.

The meaning of the title, The Rings of Saturn, is hinted at in a quote from the Brockhaus Encyclopedia included in the frontispiece:

"The rings of Saturn consist of ice crystals and probably meteorite particles describing circular orbits around the planet's equator. In all likelihood, these are fragments of a former moon that was too close to the planet and was destroyed by its tidal effect."

The Rings of Saturn is a symbolic tour of the debris of Western culture, which its author found both grotesque and beautiful, more so that it was crumbling to sand, or so he argued in 1998 when the book was published. In Browne's phrase: arches and obelisks are melting pillars of snow.

Sebald writes (beautifully):

"As I strolled through Somerleyton Hall that August afternoon, amidst a throng of visitors who occasionally lingered here or there, I was variously reminded of a pawnbroker's or an auction hall. And yet it was the sheer number of things, possessions accumulated by generations and now waiting for the day when they would be sold off, that won me over to what was, ultimately, a collection of oddities. How uninviting Somerleyton must have been in the days of [its] industrial impresario, Morton Peto, MP, when everything, from the cellar to the attic, from the cutlery to the waterclosets, was brand new, matching in every detail, and in unremittingly good taste. And how fine a place the house seemed to me now that it was imperceptibly nearing the brink of dissolution and silent oblivion. However, on emerging into the open air again, I was saddened to see, in one of the otherwise deserted aviaries, a solitary Chinese quail, evidently in a state of dementia, running to and fro along the edge of the cage and shaking its head every time it was about to turn, as if it could not comprehend how it had got into this hopeless fix. The grounds, in contrast to the waning splendor of the house, were now at their evolutionary peak, a century after the heyday of Somerleyton. The flower beds might well have been better tended and more gloriously colorful, but today the trees planted by Morton Peto filled the air above the gardens, and several of the ancient cedars, which were there to be admired by visitors even then, now extended their branches over well-nigh a quarter of an acre, each an entire world unto itself."

In my youth, I thought—which is to say I wanted—Asimov to be right. I wanted history to be the result of impersonal material forces, physical or social, as distinct from, first, "heroic" histories, popular on the Right, which saw history as the gain of a few great-willed men moving chess pieces on a battlefield, and second, from Marxist theory, popular on the Left, which imbued history, a thing in itself, with a kind of manifest destiny.

Being a product of the Age of Discovery, Marx saw history like the course of a jungle river: existing separately from us, having a determined form that already extended to the future, inescapable. Society traverses this fixed river like a man in a canoe, bounded by its shores and destined to be carried across dangerous rapids to its terminus, which Marx described as a kind of economic utopia.

This vision of history, by the way, almost exactly matches "The Voyage of Life," a short series of paintings by Thomas Cole, which were produced in 1842, just two years before Marx met Engels. All three men, it seems, were products of their time.

Marx's vision is the opposite of Thomas Browne's. Where Browne saw life—the life of everything, from a person to history itself—as a parabola that, after reaching its zenith, descended in an arc to darkness, Marx saw it instead as an exponential progression, extending ever more rapidly to the heavens, like a chart of housing prices.

Being a child of the 20th century, I had a hard time seeing history as an impersonal force. Genocide, it seemed to me, is very personal. I also had difficulty with the idea of a predetermined path “ending” in any kind of permanent condition (utopian or not), which seemed at least to contradict the findings of the physical sciences, a fact that would've shocked Marx: that the fundamental state of things is uncertain, that even rivers are not set. They change course over time—on the geologic scale, quite frequently. The undulations of the Mississippi, rolling down the land, look like a trickle of water dancing down the glass in my shower.

It seemed instead that history was more like the weather, or perhaps the ocean—governed at the large scale by innumerable natural forces that pulled against each other, but on the smaller scale bubbling with chaos. I set out, then, to try to understand some of these forces. I read historiography by Braudel and the Annales school, I read global historians like William McNeill, and of course Jared Diamond, all of whom identified currents and tidal forces in the ocean of history.

It seemed only one thing was certain.

Just about everyone these days is a grumpy old man, obsessed with decline and convinced of its inevitable loom. More than that, the acolytes of this cult, who are everywhere, are deeply suspicious that anything could ever be as good again.

The problem with the cult of decline is that it presupposes a certain fixed point of view—the origin of Browne’s parabola—and is deliberately blind to all others. In the case of Somerleyton, the house is falling apart, yes. But from the standpoint of the trees, things have never been better.

In that sense, the cult of decline is hopelessly partisan. The fall of the Roman Empire is the inescapable pivot of Browne’s arc, but the dynastic cycle in China has no pivot, only an axis, and there is an unbroken contiguity of past and present in places like Japan and Dravidic India. History, it seems, has all these shapes: parabola, sine wave, line.

From the point of view of a barnacle, the whole of the ocean seems to rise and fall. Stand long enough before the tides and eventually you’ll see the erosion of the shore. But what are the tides from inside the ocean? The erosion of one beach leads to its deposition elsewhere. What is decline but the birth of something new?

The cult of decline worships a tautology, a cognitive bias, a bank run. If you believe things are failing, then you won't expend the energy necessary to sustain them, and they will fail. In a universe like ours, dominated by the Second Law of Thermodynamics, some creative vitality is necessary to sustain anything, even a good mood.

The Roman Empire didn't succumb to insurmountable tidal forces. There was no tsunami of decrepitude that wiped it away. Everything it faced at the end—wars, barbarians, epidemics—was something it had conquered more than once before. The difference was that its people no longer cared enough to overcome them. They believed it was over, and so it was.

But far from being an end, the Empire fell and something even more miraculous took its place—everything, in fact, that we now lament losing.

Where is the tragedy?

What is certain, of course, is change, which is venerated by another, equally misguided cult, as convinced that everything is bad now as the cult of decline is that nothing will ever be as good again.

It isn’t that some things get better and some worse and it all evens out. If war breaks out and you are killed or driven from your home, things are very much worse for you. But from the standpoint of the whole show, we can’t know. Things may get better. They may get worse. But never forever. There’s no war not halted by peace. There’s no peace not halted by war.

Western culture is boiling, certainly, which looks like disintegration from inside the pot. What once congealed is now in froth. That which was will no longer be.

A moon disintegrates and becomes something rarer still.

Born to a poor working family in the British colony of Massachusetts, Dexter took to the fields around age 8 and never received much of an education. By chance, he married a wealthy widow whose equally wealthy friends liked to make fun of him for being a “plain-spoken man”—which is to say a man lacking refinement, an obnoxious boor, a wide-mouthed asshole.

And he was. Timothy Dexter had opinions on everything, most of them quite awful, and his wife’s friends, being narrow-mouthed assholes, liked to tease him with terrible business advice on the hopes that it might “teach him a lesson.”

They told him to ship coal to Newcastle, which was a major coal producer. So he did. It arrived shortly before a coal miner’s strike, and he sold it at a profit.

They told him to ship winter gloves to the Caribbean islands. So he did, where he sold them to warm-port sailors on their way to Siberia.

He sold bed-warmers as molasses ladles, Bibles to missionaries, and invested foolishly in a near-worthless “Continental currency”—which expanded his fortune again when, against everyone’s best predictions, the American colonists defeated the most powerful empire in the world.

Dexter built a mansion, which later became a hotel, and filled the grounds with statues of great men of the age: George Washington, Napoleon Bonaparte, and himself. But try as he might, he could never win the respect of the professional class, who found his arrogance and vulgarity distasteful.

And it was. At one point, Dexter faked his own death just to see who would show up at his funeral. When he noted his wife wasn’t crying, he leapt forth immediately, revealing the hoax, and took a cane to her.

At the age of 50, after amassing a fortune, Timothy Dexter decided he would write a book—about himself. “A Pickle for the Knowing Ones, or Plain Truth in a Homespun Dress” ran to a mere 8,000-and-some words, an essay by modern standards, and was nothing but one long gripe: about politicians, about the clergy, about foreigners, and about his wife, without whose fortune he would have stayed a very poor man.

But the little book’s most notable quality, that for which it is best remembered, was that through all 33,864 letters, there was not a single punctuation mark. No periods. No commas. Nothing. There wasn’t even regular capitalization.

Dexter paid for the printing himself and handed the book out for free, just like any good indie author today. But following the same ridiculous luck he’d experienced his entire life, the damned thing took off, and Timothy Dexter’s tiny monstrosity ran through eight separate printings, despite widespread criticism from the literati that the lack of punctuation made it a grade-school farce.

Alas, the great public beast eats what it wants.

Share

Dexter responded to his critics in typical style in the second edition. He appended an extra page to the book filled with nothing but punctuation—13 lines of it, in fact—“A Nuf,” he said, that his readers could “peper and solt” the text as they pleased.

Some hundred years after Timothy Dexter died, James Joyce played the same trick with punctuation, to widespread critical acclaim, in his book Ulysses, which is often hailed as the greatest English-language novel of the 20th century.

Writing is a dismal job. As pastimes go, very little can compare. But as a career, well, one could bring home more bacon from a Jewish potluck. Every week, the commercial market finds a new way to kick you in the balls. Your customers are fickle, your colleagues are false and feigning, and the only thing worse than the lot of them is your own feckless pursuit of their praise.

We’re submissives, we writers, strapped and bound without a safe word—or perhaps we’re masochist-priests who believe every stinging, self-inflicted lash rids us of some lexical sin. My characterization was weak and that’s why people didn’t like it. Lash. I was overly cerebral. Lash. I wasn’t writing authentically; I didn’t bleed onto the page.

Cardinal sin. Ten lashes.

The truth is, commercial success has very little to do with “qualitie,” as the case of Timothy Dexter shows. The dark goddess of the market will make a rich man richer just for the giggle but leave Poe penniless in the street. At the time of his death, the astonishing popularity of “The Raven” had made him a household name. He was paid $9 for it, or about $300 today, and the proceeds from all subsequent sales and reprints went to the publisher.

Even after he collapsed in a gutter in Baltimore and died, the publishing market was not done with him. Somehow, Poe’s own personal Salieri, a man called Rufus Wilmot Griswold, who bore Poe a lifelong grudge, was appointed his executor and proceeded to malign him in a series of articles culminating in a biography that painted him as a depraved drug addict. Only then did Poe’s other works get noticed, not because of their “qualitie” but because people got a secret thrill out of reading the work of a notorious madman.

So successful was this hatchet job that even today, some 170 years later, we still think of Poe as a melancholic drug addict even though historians seriously doubt that was the case.

People will respond that the world is not fair. No, hurricanes are not fair. They take worthy and unworthy alike. Commercial fiction isn’t a hurricane. It’s a con. One could wring more honesty from a strip club.

But still we suckle at it. Like every other fool, when our next book is done, we will climb to any perch free of pus and seminal fluid and prostrate ourselves before the dark goddess and beg she take note, if only to smite us again.

My advice is this. Write what you want to write and how you want to write it. There’s no secret sauce. The Dragon Scroll is empty. Those anointed by Fate will believe they earned it. But Dexter and Poe know better.

Subscribe now

How to Make Sense of Contradictory Science Papers

By Haixin Dang & Liam Kofi Bright

The science you can come across today can often appear to be full of contradictory claims. One study tells you red wine is good for your heart; another tells you it is not. Over the past year, COVID-19 research has offered conflicting reports about the overall effectiveness of wearing a mask. As scientists debate what policy best suits the current moment, they will be drawing on hundreds of studies; some that say masks are effective1 and some that say masks alone are not enough.2

Naturally, given its outsized influence on society—especially during a pandemic—people tend to regard published science highly. This means that many of us expect scientists to be prudent in reporting their results. These ought to be true and justified by evidence, right? And surely, at a bare minimum, the researchers themselves ought to believe in what they are publishing, yes? Maybe not. The bar for publishing might, counter-intuitively, be lower than one might expect. “Scientific conclusions,” as we titled our recent paper,3 “need not be accurate, justified, or believed by their authors.”

We’re not saying scientists generally lie about their published results (this has nothing to do with misconduct). Rather, we argue that scientific papers fulfill a useful social role by doing more than merely reporting on true discoveries. It’s enough for them to draw attention to an idea that is worth pursuing further—and an idea need not be true, well-justified given all our evidence, nor even believed by the scientist in order to pass that test. The peer-review process is, in fact, designed, not to detect fraud or data manipulation, but to select for what is noteworthy.4 What is considered unexpected and thought-provoking will not always track our all-things-considered judgments of what is true, but local community standards of best scientific practice having to do with how to go about data gathering and statistical testing.

For scientists to collectively inquire effectively, they need to communicate interesting ideas to each other that are worth pursuing. Consider Avi Loeb, a theoretical astrophysicist who proposed the provocative hypothesis, not without some supporting data, that ‘Oumuamua wasn’t a comet but an alien light-sail. He, presumably, knew that more data would need to be gathered, and a more thorough study would need to be conducted, before the hypothesis could be justifiably believed.

Nonetheless, perhaps it was appropriate for Loeb to publish his data and his hypothesis. He himself might even be agnostic toward the truth of the hypothesis. He likely knew that most of his colleagues would dispute his interpretation of the data, and with good cause. In spite of all this, it was still valuable for him to publicly communicate the possibility of a new hypothesis, because it can—and maybe actually did—spur more research into, and garner attention for, astronomy. Publishing those findings was not about communicating the truth but about saying that there is something exciting and interesting that requires further inquiry.

Why might it be worth worrying about how and when scientists decide to share their work? The threat of misinformation spreading, you might say. But that isn’t all. As it stands, there is a mismatch between the rules scientists write by compared with those that laymen read by. And given that non-scientists are daily called on to make important decisions on the basis of scientific results, the potential for miscommunication makes possible momentous mistakes.

For example, in the latter half of the 20th century, various linguistic theories were being proposed among academics about how children learn reading and writing. Public uptake of the whole language approach led to massive reforms in education curriculums in the United States that de-emphasized the role of phonics when teaching children how to read and instead emphasized the meaning of words in sentences. The linguistic theories that prompted these reforms were never empirically well-confirmed, and later research showed that teaching children systematic phonics is more effective.5 Education reformers treated on-going linguistic theorizing as if it were established fact.6

You might see any one such case as being something between comic and tragic. But it is the collective impact of these errors that is of greater import. Each erodes our trust in science—but by just getting clearer on what scientific publication is generally for, we can stem some of that distrust.

The philosopher Karl Popper once said that science needs bold conjectures and attempted refutations. The idea now goes under the name of “falsificationism.” Being open to refutation is one of the most widely appreciated principles of science. But nowadays people tend to mostly emphasize the importance of refutation. We want to echo Popper and stress that bold conjectures are needed, too. And not just bold conjectures, but some means for scientists to advertise their best guesses. The publication system allows scientists to filter which conjectures are worth putting out there and taking seriously, without being so conservative as to prevent scientists from taking imaginative leaps.7

In everyday life, we don’t often have need of this sort of conjectural laxity. If I ask you where you put the keys, I’m really not looking for your boldest theses on locksmithery. I really just want to know where the keys are. Such a humble purpose actually requires that, to be in good faith, my interlocutor meets a high conversational standard for assertion. Philosophers have explored what norms constitute such a standard.8

These norms are the rules that assertions obey in well-behaved conversations. Some philosophers think you have to know that what you assert is true. Others think that, even if you don’t know it, you have to at least be justified in your claims. And others think that, even if you neither know nor have justification for your claims, you must at least be sincere and believe that you know it, or that it’s justified.

Fortunately, these norms of assertion don’t constrain science. We rightly tolerate the fact that it is, essentially, what you could politely call guesswork. Researchers must constantly be open to nature surprising them, and spread out over conceptual space, exploring whatever may be found. The trouble is that the general social esteem in which people hold science makes it natural for them to make an unhelpful assumption. That if scientific claims differ from the sort of claims each of us make every day, it is because the scientific ones have a better standing—better checked, have more evidence behind them, carry greater weight than our everyday assertions.

But as we hope will be clear by now, this is far from the case. How disappointed will the consumer of science journalism thus feel if they pay attention to the modest track record of hot-off-the-press research!9 It would be no easy task to simply get scientists to work to a higher standard. Science actually requires a permissive attitude to published claims in order to do its job of discovering important truths.

What is more, today, scientists often communicate results to each other outside of traditional publication venues. What should we make of the growing influence of “preprints” posted on online archives like medRxiv and bioRxiv? Science journalists are increasingly relying on papers in their reporting that aren’t peer-reviewed, changing the norms of communicating science. As a result, the general public is now engaging directly with COVID-19 research at an unprecedented rate.10 This also means scientists can learn about each other’s research faster and with more ease. They no longer need to go through a journal, and peer review, to communicate their novel, provocative ideas to each other.

But this openness, of course, can also lead people to misunderstand what scientific papers are meant to communicate. These preprints are even further from being final, necessarily reliable discoveries. What we are witnessing is the process of science at play—the messy, inspired guesswork. We would do well to remember that. Loeb probably doesn’t seriously believe that alien tech careened through our solar system. But that provocation is nevertheless serious science.

In grade school you probably learned about humus, the organic component of soil formed by the decomposition of leaves and other plant material by soil microorganisms. Originally proposed over two centuries ago, humus is foundational to our understanding of how soils function. Earth science textbooks being shipped to schools today are filled with diagrams and descriptions of the stuff, and climate models being built that rely on it for carbon sequestration.

There’s just one problem: It probably doesn’t exist.

Does the fact that something so foundational has been overturned mean that we know nothing about soils? Of course not. But it does mean the heretics were right all along. Hence the need for bold conjectures.

We need to work very hard to reserve a space for heretics and heretical theories in our information ecology, not because they’re disestablishmentarian or challenge an orthodoxy. Challenging orthodoxy for the sake of challenging orthodoxy is as pointless as defending orthodoxy because it’s orthodox.

In science—real science, not science journalism—the only “experts” who matter are the ones who can admit when they were wrong. Anyone else is a partisan in a white coat.

You get a message from your bank warning you of a threat to your account and inviting you to click a link that validates your personal information and updates your password.

In this context, you're clearly suspicious, but phishing works by providing the right contextual clues such that those flying by at 90 mph, as normal people do on the internet, might not pause long enough to reach the point of suspicion.

One of those all-important contextual clues is “trust signalling,” which is a kind of psychological priming.

I can influence your reported support or opposition to an issue by first telling you a heartwarming human interest anecdote or instead by appealing to cold, hard economic facts. In either case, I've “primed” you by getting your brain to think about the issue from a certain point of view. The easiest thing for it to do is answer the question from the same frame.1

The phishing email frames the message by warning you of a threat, thereby subtly priming you to think it isn't one. As long as the message looks right—it does seem like it came from your bank—then, paradoxically, it's the genuine threat of identity theft that most compels a click. In wanting to ensure your money is safe, you do the opposite.

Importantly, it isn't that these fraudulent messages have to work on everyone all the time to be successful. They just have to catch enough of us at the right moment that they pay for themselves and a little extra.

Maybe you're thinking you're too clever to fall for anything like that. But almost certainly you have, because what’s being hacked here isn’t a machine, and it isn’t beaten by cleverness.

Disinformation appears to be anti-disinformation.

Just as phishing campaigns try very hard to appear as a defense against phishing, disinformation campaigns appear as the opposite—as journalists, fact-checkers, and community standards moderators purporting to destroy the very thing they’re peddling. They want you to know We’re here to help.

Just as phishing campaigns play on the urgency of a real threat, disinformation campaigns use existing fears of genuine out-groups—opposing political parties, domestic terrorists, anti/fascists, pedophiles—to scare you into some reaction.

Just as phishing campaigns only have to work occasionally to be profitable, disinformation campaigns don't have to snare everyone to move the needle on an issue. In fact, they work best when both medium and message are targeted not to everyone but specifically to you.2

Disinformation appears from your side.

It's important here to draw a distinction between disinformation and plain old information. A traditional article or op-ed is the latter. It tries to convince you with counterargument. It may be wrong, even false, in which case it's misinformation. But being misinformed is not the same as being disinformed. Only one is deliberate.

Share

Just as a phishing attack comes from “your bank,” disinformation comes from “your side.” It wasn’t the North Vietnamese (or the Russians) feeding false body counts to the evening news during the Vietnam War, just as it wasn’t the terrorists (or the Russians) who lied about WMD in Iraq.

There are in fact numerous sources of disinformation, some of which you may not consider to be hostile.

Disinformation appears on the news.

Often, the people who disseminate disinformation, on the news or on social media, genuinely believe it themselves, which of course makes them excellent salespeople.3

No outlet is immune, and unfortunately, even a biased source will seem honest by occasionally admitting error. That proves nothing.

Keep in mind the purpose of disinformation is not to convince you of falsehoods. Its authors will certainly lie if they have to, but they’re just as likely to employ honesty as deception. In fact, in as much as disinformation has to appear true, it often will be based in fact.

I'm going to repeat that because it's very important.

Disinformation appears true.

I don’t mean that disinformation’s lies are convincing, although that’s also the case. I mean disinformation can literally be true.

If that seems strange, consider that in casual language we call that a lie of omission. I don’t have to speak a falsehood to lie. I can print only facts. Just not all the facts.

Write this down somewhere you will see it regularly:

Disinformation isn't about right or wrong.

People think disinformation is evil, so they assume the point is to deceive. But disinformation isn’t about truth or falsity, good or evil, right or wrong. The purpose of disinformation is to influence. That's all.

Like advertising, disinformation aims to tell you whatever you need to hear to alter your beliefs or behavior. If that's a lightly edited version of the truth, all the easier!

Sometimes, the goal is straightforward, such as to convince you to oppose a politician or tolerate a war. At other times, the goal is oblique, such as to convince you to care very deeply for a minor injustice at the expense of a major one.

Deflection is a powerful weapon, as any martial artist will tell you.

Subscribe now

Unfortunately, just like advertising, it’s not clear how effective disinformation actually is. There’s no question, for example, that the most alarmist claims about disinformation are themselves disinformation intended to scare us into accepting a counter-regime of top-down information control.4

But there’s also no question that plenty of people believe genuine political fantasies, from WMD to Russiagate, which raises the question: Why?

Disinformation exploits intelligence, not stupidity.

It’s clear when you succumb to a phishing attack. Money in your bank account goes away, which is painful and embarrassing. So, too, is succumbing to disinformation.

But how would you know? What’s the clear marker that you’ve fallen?

It seems so unlikely, right? I mean, if you’ve succumbed, then everyone you know and the thousands you’re connected to online have succumbed as well, because you all believe the same obvious truths, and honestly, how likely is it that all those intelligent, well-intended people are wrong?

Never mind that right now, you believe equally large groups of other people are wrong all the time, so such things must be common. The difference is, those other people are not very smart, are they? Nor well-intended. The enemy never is.

People think the proper defense against disinformation is good intentions, or intelligence, but it’s not. The dull and disinterested are not the marks. They’re not paying attention. Disinformation targets smart, well-intended people. Turns out, they’re actually very good at reasoning badly, probably because they’ve had more practice.

Disinformation isn’t going anywhere.

Human beings are social apes. We spontaneously form tribes the same way stray dogs spontaneously form packs. Tribes collectively define and enforce a truth. They have to. Any that didn’t would cease to cohere. What would bind them? It just so happens that tribes these days are virtual rather than local.

The definition and maintenance of a complex, collective, cohesive reality is the normal, proper function of the socially networked human brain. That’s simply what it does when you put any number of them together, and the more intelligent you are, the better you are at participating.

The true inoculation against disinformation is not intelligence but humility, which is why disinformation thrives now, in the attention economy, and why it’s not going away anytime soon.