History as Trajectory
Political Power and Technological Truth

Political Power and Technological Truth

Narrated by Balaji AI.

In the top-down view, history is written by the winners. It is about political power triumphing over technological truth.

Why does power care about the past? Because the morality of society is derived from its history. When the Chinese talk about Western imperialism, they aren’t just talking about some forgettable dust-up in the South China Sea, but how that relates to generations of colonialism and oppression, to the Eight Nations Alliance and the Opium Wars and so on. And when you see someone denounced on American Twitter as an x-ist, history is likewise being brought to bear. Again, why are they bad? Because of our history of x-ism…

As such, when you listen to a regime’s history, which you are doing every time you hear its official organs praise or denounce someone, you should listen critically.

Political Power as the Driving Force of History 

How do the authorities use history? What techniques are they using? It’s not just a random collection of names and dates. They have proven techniques for sifting through the archives, for staffing a retinue of heros and villains from the past, for distilling the documents into (politically) useful parables. Here are two of them.

  • Political determinist model: history is written by the winners. People have heard this saying, but taking it seriously has profound implications. For example, whoever claims to be writing the “first draft of history” is therefore one of the winners. For another, history is what’s useful to the regime. A classic example is Katyn Forest: the admission that the Soviets did it would have delegitimized their postwar control over Poland during the 1945-1991 period, but once the USSR collapsed the truth could be revealed.

  • Political mascot model: history is written by winners pretending to be acting on behalf of losers. This is a variant of the political determinist model, also known as “offense archaeology,” and practiced by the modern American, Chinese, and Russian establishments — all of whom portray themselves as victims. The technique is to pick a mascot that the state claims to champion, such as the Soviet Union’s proletariat, and then go through history to find the worst examples of the state’s current rival doing something bad to them.

    Take these real events, put them on the front page, and ensure everyone knows of them. Conversely, ensure off-narrative events are ignored or suppressed as taboo. Again taking the USSR as a case study, this involved finding endless (real!) examples of Western capitalists screwing the working class, and suppressing the worse (also real!) instances of Soviet communists gulaging their working class, as well as cases of the working class itself behaving badly. Generalization to other contexts is left as an exercise for the reader, but here’s a Russian example of what an American would call “responsibility to protect” (R2P).

These techniques are used to write history that favors a state. Here are more examples:

  • CCP China: Today’s Chinese media covers the Eight-Nations Alliance, the Opium Wars, and the like exhaustively in its domestic output, as these events show the malevolence of the European colonialists — who literally fought wars to keep China subjugated and addicted to heroin. Their domestic history does not mention the Uighurs, Tiananmen, and the like domestically. Xi’s CCP did stress the domestic problem of corruption via the “Tigers and Flies” campaign…but that’s in part because the anti-corruption campaign was politically useful against his internal enemies, and seemed not to ensnare his allies.

  • US Establishment: Today’s US establishment covers 6/4/1989 and the 2022 Russo-Ukrainian War heavily, because they are real events that make China and Russia look bad and the US look good. It does not mention the 1900 Eight-Nations Alliance (when the US helped invade China with a “coalition of the willing” to defend European imperialism) or the 1932 Ukrainian Holodomor (when The New York Times Company’s Walter Duranty helped Soviet Russia choke out Ukraine) as these cut in the opposite direction.

    The current US narrative also does not stress the Cultural Revolution (which bears too close a resemblance to present day America), or Western journalists like Edgar Snow who helped Mao come to power, or the full ugly history of American support for Russian and Chinese communism. This isn’t simply a matter of the age of events — after all, regime media goes back further in time when convenient, distorting events from 1619 for today’s headlines, yet somehow their time machine stutters on the years 1932 or 1900. In modern America, as in modern China, the history you hear about is the history the establishment finds to be politically useful against its internal and external rivals.

  • The British Empire: The British in both WW1 and WW2 understandably emphasized the evils of Germany, but not so much the evils of their ally Russia, or their own evils during the Opium Wars, or the desire for the Indian subcontinent to breathe free, and so on. (This one is almost too easy as the UK is no longer a contender for heavyweight champion of the world, so no one is offended when someone points out its past self-serving inconsistencies. Indeed, documenting the UK’s sins is now a cottage industry for Britain’s virtue signalers, as beating up on a beaten empire is far easier than tackling the taboos of a still live one.)

Point being: once you get your head out of the civilization you grew up in, and look at things comparatively, the techniques of political history become obvious. One of those techniques deserves special mention, and that’s a peacetime version of the “atrocity story”:

One of the most time-honored techniques to mobilize public animosity against the enemy and to justify military action is the atrocity story. This technique, says Professor Lasswell, has been used “with unvarying success in every conflict known to man.”

The concept is as useful in peacetime as it is in war. Why? Because states get their people hyped up to fight wars by stressing the essentially defensive nature of what they are doing and the savage behavior of the enemy. But war is politics by other means, so politics is war by other means. Even in peacetime, the state is predicated on force. And this use of force requires justification. The atrocity story is the tool used to convince people that the use of state force is legitimate.

Coming from a different vantage point, Rene Girard would call this a “founding murder.” Once you see this technique, you see it everywhere. Somewhat toned-down versions of the atrocity story are the go-to technique used to justify expansions of political power.

  • If we don’t force people to take off their shoes at the airport, people will die!
  • If we don’t stop people from voluntarily taking experimental curative drugs, people will die!
  • If we don’t set up a disinformation office to stop people from making hostile comments online, people will die!

Indeed, almost everything in politics is backed by an atrocity story.28 There’s a sometimes real, sometimes fake, sometimes exaggerated Girardian founding murder (or at least founding injury) behind much of what the government does.

Sometimes the atrocity story is framed in terms of terrorists, sometimes in terms of children…but the general concept is “something so bad happened, we must use (state) force to prevent it from happening again.” Often this completely ignores the death caused by that force itself. For example, when the FDA “prevented” deaths by cracking down on drug approvals after thalidomide, it caused many more deaths via Eroom’s Law and drug lag.

And sometimes the atrocity story is just completely fake; before Iraq was falsely accused of holding WMD, it was falsely accused of tossing babies from incubators.

With that said, it’s possible to overcorrect here. Just because there is an incentive to fake (or exaggerate) atrocities does not mean that all atrocities are fake or exaggerated.29 Yes, you should be aware that states are always “flopping,” exaggerating the severity of the fouls against them or the mascots they claim to represent, trying to bring in the public on their side, whether they are Chinese or American or Russian.

But once you’re aware of the political power model of history, the next goal is to guard against both the Scylla and the Charybdis, against being too credulous and too cynical. Because just as the atrocity story is a tool for political power, unfortunately so too is genocide denial — as we can see from The New York Times’ Pulitzer-winning coverup of Stalin’s Ukrainian famine.

To maintain this balance, to know when states are lying or not, we need a form of truth powerful enough to stand outside any state and judge it from above. A way to respond to official statistics not with either reflexive faith or disbelief, but with dispassionate, independent calculation.

The bottom-up cryptohistory we introduced in the previous section is clearly relevant. But to fully appreciate it we need an allied theory: the technological truth theory of history.

Technological Truth as the Driving Force of History 

The political power model of history gives us a useful lens: history is often just Leninist who/whom and Schmittian friend/enemy. But it’s a little parched30 to say that history is always and only that, solely about the raw exercise of political power. After all, a society must pass down true facts about nature, for example, or else its crops will not grow31 — and its political class will lose power.

This leads to a different set of tech-focused lenses for analyzing history.

  • Technological determinist model: technology is the driving force of history. While the political determinist model stresses that history is written — and hence distorted — by the winners, and thereby propagates only that which is useful to a given state, the technological determinist model notes that there are some key areas — principally in science and technology — where many (if not most) societies derive a benefit from passing down a technical fact without distortion. There is after all an unbroken chain from Archimedes, Aryabhata, Al-Kwarizhmi, and antiquity to all our existing science and technology. Hundreds of years later, we don’t care that much about the laws of Isaac Newton’s time, but we do care about Newton’s laws. In this model, all political ideologies have been around for all time — the only thing that changes is whether a given ideology is now technologically feasible as an organizing system for humanity. Thus: political fashions just come and go in cycles, so the absolute measure of societal progress is a culture’s level of technological advancement on something like the Kardashev scale.

  • Trajectory model: histories are trajectories. We mentioned this concept before when we discussed history as a cryptic epic of twisting trajectories, but it’s worth reprising. If you’re technically inclined, you might wonder why we spend so much time on history in this book. One answer is that histories are trajectories of dynamical systems. If you can spend your entire life studying wave equations, diffusion equations, time series, or the Navier-Stokes equations — and you can — you can do the same for the dynamics of people. In more detail, we know from physics (and Stephen Wolfram!) that very simple rules can produce incredibly complicated trajectories of dynamical systems. For Navier-Stokes, for example, we can divide these trajectories up into laminar flow, turbulent flow, inviscid flow, incompressible flow, and so on, to describe different ways a velocity field can evolve over time. These classifications are derived from measurements made of fluids over time. And the study of just one of these trajectory types can be a whole research discipline.

    That’s how rich the dynamics of inanimate objects are. Now compare that to the macroscopic movements of millions of intelligent agents. You can similarly try to derive rules about how humans behave under situations of laminar good times, turbulent revolutionary times, and so on by studying the records we have of human behavior — the data exhaust that humans produce.

    This analogy is actually very tight if you think about virtual economies and the history of human behavior on social networks and cryptosystems. In the fullness of time, with truly open datasets, we may even be able to develop Asimovian psychohistory from all the data recorded in the ledger of record, namely a way to predict the macroscopic behavior of humans in certain situations without knowing every microscopic detail. We can already somewhat do this for constructed environments like games32 and markets, and ever more human environments are becoming literally digitally constructed.33

  • Statistical model: history aids predictions. From a statistician’s perspective, history is necessary for accurately computing the future. See any time series analysis or machine learning paper — or the Kalman filter, which makes this concept very explicit. To paraphrase Orwell, without a quantitatively accurate record of the past you cannot control the future, in the sense that your control theory literally won’t work.

  • Helix model: linear and cyclical history can coexist. From a progressive’s perspective, history is a linear trend, where the “arc of history” bends towards freedom, and where those against a given cause are on the wrong side of history34. Others think of history as cyclical, a constant loop where the only thing these technologists are doing is reinventing the wheel, or where “strong men create good times, good times create weak men, weak men create hard times, and hard times create strong men.” But there’s a third view, a helical view of history, which says that from one viewpoint history is indeed progressive, from another it’s genuinely cyclical, and the reconciliation is that we move a bit forward technologically with each turn of the corkscrew rather than collapsing. In this view, attempts to restore the immediate preceding state are unlikely, as they’re rewinding the clock — but you might be able to get to a good state by winding the helix all the way past 12’o’clock to get the reboot. Or you might just collapse.

  • Ozymandias model: civilization can collapse. History shows us that technological progress is not inevitable. The Fall of Civilizations podcast really makes this clear. Gobekli Tepe is one example. Whether you’re thinking of this as an astronomer (where are all the intelligent life forms out there? Is the universe a dark forest?) or an anthropologist (how did all these advanced civilizations just completely die out?), it’s sobering to think that our civilization may just be like the best player in a video game so far: we’ve made it the furthest, but we have no guarantee that we’re going to win before killing ourselves35 and wiping out like all the other civilizations before us.

  • Lenski model: organisms are not ordinal. Richard Lenski ran a famous series of long-term evolution experiments with E. coli where he picked out a fresh culture of bacteria each day, froze it down in suspended animation, and thereby saved a snapshot of what each day of evolution looked like over the course of decades. The amazing thing about bacteria is that they can be unfrozen and reanimated, so Lenski could take an old E. coli strain from day 1173 and put it into a test tube with today’s strain to see who’d reproduce the most in a head-to-head competition. The result showed that history is not strictly ordinal; just because the day 1174 strain had outcompeted the day 1173 strain, and the day 1175 strain had outcompeted the day 1174 strain, and so on — does not necessarily mean that today’s strain will always win a head to head with the strain from day 1173. The complexity of biology is such that it’s more like an unpredictable game of rock/paper/scissors.

  • Train Crash model: those who don’t know history are doomed to repeat it. Another way to think about history is as a set of expensive experiments, where people often made certain choices that seemed reasonable at the time and ended up in calamitous straits. That’s communism, for example: a persuasive idea for many, but one that history shows to not actually produce great results in practice.

  • Idea Maze model: those who overfit to history will never invent the future. This is the counterargument to the Train Crash model — past results may not predict future performance, and sometimes you need to have a beginner’s mindset to innovate. Generally this works better for opt-in technologies and investments than top-down modifications of society like communism. One tool for this comes from a concept I wrote up a while ago called the idea maze. The relevant bit here is that just because a business proposition didn’t work in the past doesn’t necessarily mean it won’t work today. The technological and social prerequisites may have dramatically changed, and doors previously closed may now have opened. Unlike the laws of physics, society is not time invariant. As even the world’s leading anti-tech blog once admitted:

    Virtual reality was an abject failure right up to the moment it wasn’t. In this way, it has followed the course charted by a few other breakout technologies. They don’t evolve in an iterative way, gradually gaining usefulness. Instead, they seem hardly to advance at all, moving forward in fits and starts, through shame spirals and bankruptcies and hype and defensive crouches — until one day, in a sudden about-face, they utterly, totally win.

  • Wright-Fisher model: history is what survives natural selection. In population genetics, there’s an important model of how mutations arise and spread called the Wright-Fisher model. When a new mutation arises, it’s in only 1 out of N people. How does it get to N out of N, to 100%, to what’s called “fixation”? Well, first, it might not ever do that. It might just die out. It might also get to N out of N simply by luck, if the population of N is small — this is known as “fixation by genetic drift,” where those with the mutation just happen to reproduce more than others. But if the mutation confers some selective advantage s, if it aids in the reproduction of its host in a competitive environment, then it has a better than luck chance of getting to 100%. Similarly, those historical ideas that we’ve heard about can be thought of as those that aided or at least did not interfere with the propagation of their respective carriers, often the authorities that write those histories. Some of these ideas have tagged along by dumb luck, while others are claims that were selectively advantageous to the success of the regime - often by delegitimizing their rivals and legitimizing their own rule, or by giving them new technologies. This is a theory of memetic evolution; the ideological mutations that add technological edge or political power are the ones selected for.

  • Computational model: history is the on-chain population; all the rest is editorialization. There’s a great book by Franco Moretti called Graphs, Maps, and Trees. It’s a computational study of literature. Moretti’s argument is that every other study of literature is inherently biased. The selection of which books to discuss is itself an implicit editorialization. He instead makes this completely explicit by creating a dataset of full texts, and writing code to produce graphs. The argument here is that only a computational history can represent the full population in a statistical sense; anything else is just a biased sample.

  • Genomic model: history is what DNA (and languages, and artifacts) show us. David Reich’s Who We Are and How We Got Here is the canonical popular summary of this school of thought, along with Cavalli-Sforza’s older book on the History and Geography of Human Genes. The brief argument is: our true history is written in our genes. Mere texts can be faked, distorted, or lost, but genomics (modern or ancient) can’t be. Languages and artifacts are a bit less robust in terms of the signal for historical reconstruction, though they often map to what the new genomic studies are showing about patterns of ancient migrations.

  • Tech Tree model: history is great men constrained by the adjacent possible. As context, the great man theory of history says that individuals like Isaac Newton and Winston Churchill shaped events. The counterargument says that these men were carried on tides larger than them, and that others would have done the same in their place. For example, for many (not all) Newtons, there is a Leibniz, who could also have invented calculus. It’s impossible to fully test either of these theories without a Lenski-like experiment where we re-run history with the same initial conditions, but a useful model to reconcile the two perspectives is the tech tree from Civilization. Briefly, all known science represents the frontier of the tree, and an individual can choose to extend that tree in a given direction. There wasn’t really a Leibniz for Satoshi, for example; at a time when others were focused on social, mobile, and local, he was working on a completely different paradigm. But he was constrained by the available subroutines, concepts like Hashcash and chained timestamps and elliptic curves. Just like da Vinci could have conceived a helicopter, but probably not built it with the materials then available, the tech tree model allows for individual agency but subjects it to the constraint of what is achievable by one person in a given era. The major advantage of a tech tree is that (like the idea maze) it can be made visible, and navigable, as has been done for longevity by the Foresight Institute.

You might find it a bit surprising that there are as many different models for understanding history — let’s call them historical heuristics — as there are programming paradigms. Why might this be so? Well, just like the idea of statecraft strategies that we introduce later, the study of history can also be analogized to a type of programming, or at least data analysis. That is, history is the analysis of the log files.

  • Data exhaust model: history as the analysis of the log files. Here, we mean “log files” in the most general sense of everything society has written down or left behind; the documents, yes, but also the physical artifacts and genes and artwork, just like a log “file” can contain binary objects and not just plain text.

    Extending the analogy, you can try to debug a program by flying blind without the logs, or alternatively you can try to look at every row of the logs, but rather than either of these extremes you’ll do best if you have a method for distilling the logs into something actionable.

And that’s why historical heuristics exist. They are strategies for distilling insight from all the documents, genes, languages, transactions, inventions, collapses, and successes of people over time. History is the entire record of everything humanity has done. It’s a very rich data structure that we have only begun to even think of as a data structure.

We can now think of written history as an (incomplete, biased, noisy) distillation of this full log. After all, if you’ve ever found a reporter’s summary of an eyewitness video to be wanting, or found a single video misleading relative to multiple camera angles, you’ll realize why having access to the full log of public events is a huge step forward.

A Collision of Political Power and Technological Truth 

We’ve now defined a top-down and bottom-up model of history. The collision of these two models, of the establishment’s Orwellian relativism36 and the absolute truth of the Bitcoin blockchain, of political power and technological truth…that collision is worth studying.

Let’s do three concrete examples where political power has encountered technological truth.

  • Tesla > NYT. Elon Musk used the instrumental record of a Tesla drive to knock down an NYT story. The New York Times Company claimed the car had run out of charge, but his dataset showed they had purposefully driven it around to make this happen, lying about their driving history. His numbers overturned their letters.
  • Timestamp > Macron, NYT. Twitter posters used a photo’s timestamp to disprove a purported photo of the Brazilian fires that was tweeted by Emmanuel Macron and printed uncritically by NYT. The photo was shown via reverse image search to be taken by a photographer who had died in 2003, so it was more than a decade old. This was a big deal because The Atlantic was literally calling for war with Brazil over these (fake) photos.
  • Provable patent priority. A Chinese court used an on-chain timestamp to establish priority in a patent suit. One company proved that it could not have infringed the patent of the other, because it had filed “on chain” before the other company had filed.

In the first and second examples, the employees of the New York Times Company simply misrepresented the facts as they are wont to do, circulating assertions that were politically useful against two of their perennial opponents: the tech founder and the foreign conservative. Whether these misrepresentations were made intentionally or out of “too good to check” carelessness, they were both attempts to exercise political power that ran into the brick wall of technological truth. In the third example, the Chinese political system delegated the job of finding out what was true to the blockchain.

In all three cases, technology provided a more robust means of determining what was true than the previous gold standards — whether that be the “paper of record” or the party-state. It decentralized the determination of truth away from the centralized establishment.

A Definition of Political and Technological Truths 

It isn’t always possible to decentralize the determination of truth away from a political establishment. Some truths are intrinsically relative (and hence political), whereas others are amenable to absolute verification (and hence technological).

Here’s the key: is it true if others believe it to be true, or is it true regardless of what people believe?

A political truth is true if everyone believes it to be true. Things like money, status, and borders are in this category. You can change these by rewriting facts in people’s brains. For example, the question of what a dollar is worth, who the president is, and where the border of a country is are all dependent on the ideas installed in people’s heads. If enough people change their minds, markets move, presidents change, and borders shift.37

Conversely, a technical truth is true even if no human believes it to be true. Facts in math, physics, and biochemistry are in this category. They exist independent of what’s in people’s brains. For example, what’s the value of π, the speed of light, or the diameter of a virus? 38

Those are the two extremes: political truths that you can change by rewriting the software in people’s brains, and technical truths that exist independent of that.

A Balance of Political Power and Technological Truth 

Once you reluctantly recognize that not every aspect of a sociopolitical order can be derived from an objective calculation, and that some things really do depend on an arbitrary consensus, you realize that we need to maintain a balance between political power and technological truth.39

Towards this end, the Chinese have a pithy saying: the backwards will be beaten. If you’re bad at technology, you’ll be beaten politically. Conversely, the Americans also have a saying: “you and what army?” It doesn’t matter how good you are as an individual technologist if you’re badly outnumbered politically. And if you’re unpopular enough, you won’t have the political power to build in the physical world.

Combining these views tells us to seek a balance between nationalism and rationalism, where the former is thought of in the broadest sense as “group identity.” It’s a balance between political power and technological truth, between ingroup-stabilizing narratives and inconvenient facts. And you need both.

So that’s how the political and technological theories of history interrelate. Technological history is the history of what works; political history is the history of what works to retain power. Putting all the pieces together:

  • We have a political theory of history that says “social and political incentives favor the propagation of politically useful narratives.”
  • We have a technological theory of history that says “financial and technical incentives favor the propagation of technological truths.”
  • We have a set of examples that show how politically powerful actors were constrained by decentralizing technology.
  • We have more examples that show that some facts really are determined by societal consensus, while others are amenable to decentralized verification.
  • And we understand why groups need both to survive; the backwards will be beaten, while the unpopular will never have political power in the first place.

Can we generalize these observations into a broader thesis, into an overarching theory that includes the clash of political power and technological truth as a special case? We can. And that leads us to a discussion of God, State, and Network.

Next Section:

God, State, Network