Decentralization
The Era of Decentralization is coming.
Welcome to the 2,810 newly Not Boring people who have joined us since last Monday! If you haven’t subscribed, join 170,734 smart, curious folks by subscribing here:
🎧 For the audio version, listen on Spotify or Apple Podcasts 🤖 AI voice via play.ht
Today’s Not Boring is brought to you by… Masterworks
You may have missed a big financial story earlier this month (Not FTX). On the same day that the Dow fell over 600 points, and Bitcoin plunged to its lowest level since 2020 Microsoft Co-founder Paul Allen's art collection fetched a record $1.6 billion worth of art in less than 24 hours. The super-wealthy see art as a safe investment amid a tumultuous and uncertain global economy, per BBC.
Three things to know about the asset class: 1) fine art can help protect purchasing power when inflation is high (Goldman), (2) fine art has a -0.34 correlation to equities (Citi), and 3) contemporary art has appreciated 13.8% annually on average for the last 26 years.
So why am I telling you this if only the .01% wealthiest people can capitalize on this asset class? Well, now you can too, with Masterworks.com, the fractional art investment platform.
All offerings are SEC qualified (see offering circulars on the site), and their track record is solid so far…their last three offerings handed members 17.8%, 21.5%, and 33.1% annual net returns, respectively.1
Now there are 12 paintings in the Not Boring collection. Want to see what's currently available? You can get priority access right here.
Hi friends 👋,
Happy Monday! This is an essay I’ve been thinking about and working on for a little while – I quote myself a lot because I’ve been unintentionally writing about the theme since the beginning of Not Boring – and I’m really excited to share it with you. It’s as close to a Theory of Everything as I can come up with to understand the world in 2022.
Once I started thinking about the world this way, I’ve seen it pop up everywhere, from politics to energy to AI. Maybe it’s Baader-Meinhof, maybe it’s real.
And as I’ve tried to explain it to people, I’ve realized that I just need to write down as coherent a thesis with as much evidence as I can, and open it up to get feedback and pushback. If nothing else, I hope it can be an interesting conversation starter at Thanksgiving dinner on Thursday.
The idea is this: the world oscillates between centralization and decentralization, with progress sloping upward through the turns. We’re approaching an era of decentralization.
An important note on how you should set expectations going into this piece:
You should not read this as a bull case for crypto. While the word “decentralization” is tied to crypto, this piece is broader than that. I actually think one of the things I’ve gotten wrong about crypto is that it’s a symptom of a larger trend towards decentralization, not the cause.
You should read it like I’m Charlie Day in that conspiracy theory meme:
I’m pinning things on the board and trying to connect them, although unlike Charlie, I will try to poke holes where the connections don’t hold.
Let’s get to it.
Decentralization
The first thing you need to understand is that all of our systems are imperfect.
The second, which you need to believe as much as understand, is that they’re improving.
Churchill’s quip, that democracy is the worst form of government except for all of the others that have been tried, can be applied to the modern best practice for each and every thing we do.
When our descendants look back on our era from a thousand years in the future, they’ll view the way that we do everything as incredibly, laughably primitive. Our most advanced technology, best-laid plans, and most resilient forms of organization will be viewed, if they’re discussed at all, as historical artifacts of a time before we knew what the hell we were doing.
“Wait wait wait,” they’ll think-laugh to each other, “They used to literally type things into ‘computers’? And get this: they found their partners completely randomly based on whoever lived nearby. And when they needed a thing, they found a picture of it on the ‘internet’ and then another person physically brought it to their house. No no, also, they burned dinosaur bones to get energy! Wait, just one more: thousands of them worked for the same ‘company’ for years because of efficiency or something? And then they just… died.”
Think of any example of a Thing That Just Is, and play it back to yourself in those future kids’ voices. It all crumbles under their sneers.
Zooming that far out provides a useful perspective. It’s a little easier to accept that the things we hold true will change on that timescale. On a thousand year timescale, it’s much harder to imagine what won’t change dramatically. And if history is a guide, it’s hard to imagine that things won’t change for the better. The inconsequential things will fade away. The important things will improve. New will replace old, as it always does.
What the better versions of everything will look like is a complete unknown. But the process by which they’ll improve, at least in the near future, is knowable, at least if we can extrapolate from how things have changed and improved in the past.
One way to look at the history of the world is as a sine wave oscillating between centralization and decentralization, with an upward slope on the quality axis.
To make this idea more tangible, think of what we do for work. Two centuries ago, 84% of Americans worked on (mainly) our own farms in order to feed ourselves (decentralized, inefficient). By a century ago, we worked for a boss in big factories (centralized, more efficient), and today, while many people still work inside of big companies, more and more of us are working for ourselves or in creative jobs while automation replaces evermore of what humans used to do (decentralized, even more efficient). Technological innovation enabled that progress – first, the Industrial Revolution’s steam engines, then the Internet.
There’s this excellent 2013 Slate Star Codex post, We Wrestle Not With Flesh and Blood, But Against Powers and Principalities, in which SSC describes Puppets and Vast Formless Things (terms he borrows from Edgar Allan Poe’s Ligaea). Vast Formless Things are the drivers of change in history. Puppets are the interchangeable people or institutions that carry out the bidding of the Vast Formless Things.
“I think,” writes SSC, “the vastest and most formless Vast Formless Thing of all is technological progress.”
As an example, he argues that the Printing Press precipitated the Protestant Reformation, the Newspaper, the Renaissance, the Scientific Revolution, and the Rise of Nationalism. Martin Luther, Leonardo da Vinci, and Sir Isaac Newton were Puppets. The Printing Press was the Vast Formless Thing. After its invention, all of those other things were destined to happen, no matter who ended up bringing them to life. “So the biggest changes in history have been predetermined reactions to different technological conditions.”
In this view, technological progress sets the stage, and everything else plays its scripted part.
I’ve written a lot about technological progress, and my belief that things are only going to get crazier from here. I’ve written about progress as an exponential but smooth thing. But it might be more stark than that; technological progress may draw a more clear boundary between eras, between the way things were and the way things are going to be.
Right now, as you read this, we’re only really less than thirty years into one of the biggest technological changes in history – the Internet – and we’re standing on the precipice of a bunch of other big potential driving shifts: AI and renewable energy are two with the potential to really shake things up, among others. If things feel unstable, that’s because they probably are.
The Straussians would say that we’re about to enter our Fourth Turning. Historically, according to Strauss, Fourth Turnings have been marked by “bone-jarring Crises so monumental that, by their end, American society emerged in a wholly new form.” The last Fourth Turning was kicked off by the stock market crash of 1929 and ended with the end of World War II.
If history repeats itself, we’re on the verge of a twenty year winter set off by a major crisis like a financial crash and global war. Given everything going on in the world, that possibility doesn’t seem so far-fetched. On the other side, in the next First Turning, a new order awaits.
In a prescient 2015 blog post, A Big Sweeping Theory of Modern History, Noah Smith laid out the cycle slightly differently, but ended up in the same place: war, and then a new order.
More softly and hopefully, we’re in a Liminal Space, “neither here nor there.” My mom recently read How to Lead When You Don’t Know Where You’re Going; Leading in a Liminal Season by Rev. Susan Beaumont, and emailed me her notes. Baader-Meinhof strikes again. According to Beaumont, there are three phases of the change process:
Separation: A period in which a person, group or social order is stripped of the identity and status that previously defined it.
Liminal Period: A disorienting period of non-structure or anti-structure that opens new possibilities no longer based on old status or power hierarchies. New identities are explored and new possibilities considered.
Reorientation: A reforming period in which the person, group, or social order adopts a new identity, is granted new status and designs new structures more appropriately suited for the new identity.
Whether we’re entering a Fourth Turning, a major war, living in a Liminal Space, or all three, it’s increasingly clear that we’re heading towards something new on the other side.
Here’s my working thesis:
We’re heading towards an era of greater decentralization on all fronts – geopolitics, finance, education, journalism, and energy are just a few examples – driven by technologies including, but not limited to, the internet. This newly decentralized era will require new infrastructure and organizing principles that can adapt to the chaos and complexity inherent in decentralized systems.
My argument isn’t clean or without gaping openings for counter-arguments, and I’ll try to make some of those, too. Again, think Charlie Day at that conspiracy board. But everywhere I look, I see the seedlings of a shift from centralization to decentralization:
Energy: Fossil Fuels vs. Renewables
Manufacturing: Globalization vs. Reshoring
Manufacturing: Making vs. Growing
Science: Government Funded vs. Decentralized
Hard Tech: Government Agencies & Incumbents vs. Startups
AI: Closed vs. Open
Talent: Big Tech vs. Startups
Apps: Big vs. Small
Media: Substack vs. Journalism
Education: Factory vs. Personalized
Finance: Big Banks vs. Fintech
The push towards decentralization isn’t complete. It may just be starting. Even at maturity, there will still certainly be centralized counter-examples (and the seeds of re-centralization).
In the middle of this storm, in this Liminal Space, it feels like things are fragmenting and splintering, directionlessly and pointlessly. But I’m an optimist, and I think there’s an upward slope to all of the ups and downs. New technologies shatter old systems, but they help us piece new ones together in previously-impossible and superior ways.
Before we get to those examples, let’s take a journey through the history of that upward sloping sine wave to set the stage for the idea that not only are a bunch of things decentralizing, but that that decentralization sets the stage for a new era defined by decentralization.
The Upward Sloping Sine Wave
In the beginning, there was nothing. Then, in an instant, total centralization. All of the energy and matter in the universe existed in one single superdense, superhot point. Then… bang.
The Big Bang kicked off an ongoing, 13.8 billion year process of universal decentralization. The universe continues to expand at a rate of 74 kilometers per second per megaparsec. That’s really fast. From a single point to a 93-billion-light-year-wide universe filled with dark matter and planets and black holes and asteroids and stars and galaxies and supernovae, at least on our tiny little rock, living beings.
The emergence of those living beings, including you, me, and all of the people and animals you know and love, could be viewed as a process of centralization. As Richard Dawkins writes in The Selfish Gene:
Darwin’s theory of evolution by natural selection is satisfying because it shows us a way in which simplicity could change into complexity, how unordered atoms could group themselves into ever more complex patterns until they ended up manufacturing people.
The Big Bang sent a bunch of atoms flying, and some of those atoms found each other and teamed up, forming molecules, some of which formed genes, which mutated and replicated over billions of years to create you and me. On that level, we’re walking embodiments of centralization. On a societal level, though, each of us is an atom.
Each of us represents one node in the larger human system, what Tim Urban calls the “Human Colossus.”
The better we could communicate on a mass scale, the more our species began to function like a single organism, with humanity’s collective knowledge tower as its brain and each individual human brain like a nerve or a muscle fiber in its body. With the era of mass communication upon us, the collective human organism—the Human Colossus—rose into existence.
Human history can be viewed as a history of the ways that we’ve organized ourselves to meet our basic needs and attempted to progress beyond them.
The common story of human history goes something like this. A long time ago, we were hunter-gatherers, living equally among close kin in small bands until agriculture came along and necessitated new, hierarchical social structures. Since then, we’ve been building larger and more complex societal structures.
Compelling, but as Harri Thomas pointed out to me on Twitter, incorrect. Harri suggested that I read The Dawn of Everything: A New History of Humanity by Davids Graeber and Wengrow, and I did. The book, relying on the latest anthropological and archaeological evidence, paints a more optimistic view of humanity.
Some hunter-gatherer societies were hierarchical, just as some agrarian societies were democratic. Some early societies obeyed rulers when they gathered in one place during the winter, and broke up into egalitarian tribes in the summer; others did the same in reverse. Some cities began hierarchical and evolved to become more equal over time. In many cases, rulers were rulers in name only. No one faced repercussions for not following their rules.
All to say, the way that humans organized during prehistory were varied, creative, and flexible compared to the way that we organize ourselves today. Even then, we oscillated between centralization and decentralization, often season-by-season.
The authors set out to answer the question, “Where did inequality come from?” and ended up asking instead, “How did we get stuck?”
Graeber and Wengrow argue that humans lost three freedoms enjoyed by our ancient ancestors:
The freedom to move away or relocate from one’s surroundings
The freedom to disobey commands issued by others
The freedom to shape entirely new social realities, or shift back and forth between different ones
Importantly, the authors point out, the first two freedoms “often acted as a kind of scaffolding for the third, more creative one.” The right to relocate and disobey commands gave people the freedom to experiment with social realities.
In light of Graeber and Wengrow’s observations, that chart I drew earlier isn’t quite right. At any given point in history, there were experiments in centralized and decentralized societies being run, and drawing a line through them shows how the dominant themes have changed over time. If anything, there are fewer different structures of society now than there were then, fewer experiments being run.
They conclude the book optimistically. If the current system isn’t the historically inevitable result of larger and more productive societies, then it’s possible for us to experiment with new orders, like our ancestors did.
In that spirit, I don’t want to paint too clean a narrative here. Graeber and Wengrow disagree with SSC’s take on technological progress as the Vast Formless thing, writing:
Choosing to describe history the other way round, as a series of abrupt technological revolutions, each followed by long periods when we were prisoners of our own creations, has consequences. Ultimately it is a way of representing our species as decidedly less thoughtful, less creative, less free than we actually turn out to have been. It means not describing history as a continual series of new ideas and innovations, technical or otherwise, during which different communities made collective decisions about which technologies they saw fit to apply to everyday purposes, and which to keep confined to the domain of experimentation or ritual play.
What is true of technological creativity is, of course, even more true of social creativity… time and again in human history, that zone of ritual play has also acted as a site of social experimentation – even, in some ways, as an encyclopaedia of social possibilities.
I think that chart I just shared, however badly drawn, can resolve the two arguments. People should be able to experiment and play in sandboxes to discover new technological and social tools with low stakes (all of the little dots), but certainly, there’s an overarching shape as some technologies and social structures dominate as Vast Formless Things (the sine wave). We’re not Puppets, then, but Experimentalists.
Whatever facts of prehistory led us to modern history, the fact remains that modern societies are built in hierarchical structures, at oscillating levels of centralization, to achieve increasingly complex things.
We gave a brief history of human governance in Go Fork Yourself, which you can read for a more detailed timeline if this piece’s 13k+ words aren’t enough, but to recap, the dominant form of governance worldwide went from:
Experimental: decentralized forms of governance across many small groups
Autocracy: centralized power, simple
Democracy (small experiment vs. main global form): more decentralized power, simple
Empire: more centralized power, larger and more complex
Representative Democracy: more decentralized, larger and more complex
Charting that out looks like our upward sloping sine wave.
Of course, any narrative that clean glosses over a ton of detail, nuance, and counterpoints. For most of the period I cleanly labeled “Representative Democracy,” there have been mini-oscillating sine waves running along the smooth surface of the larger one, little “Communism” dots that break the trend.
In his book, Slouching Towards Utopia: An Economic History of the Twentieth Century, economist Brad DeLong describes that century’s battle between Communism’s centralized planning (centralized) and western capitalism’s market forces (decentralized). That battle, too, has played out on different scales. The Cold War is a stark example, with the fall of the USSR marking a win for decentralized market forces.
Within those free markets, though, the march of centralization defined industrial progress. In his excellent 2012 essay, Centralization vs. Decentralization: Two Centuries of Authority in Design, Smári McCarthy writes: “The history of the world over the last two hundred years has been a history of centralization.”
McCarthy cites the Industrial Revolution as the driver of that centralization. Prior to the Industrial Revolution, McCarthy observes, people were craftspeople, and “production was inherently local; goods were manufactured by independent craftsmen or in small factories within a small community of people.”
While “people were reduced to automata,” progress exploded. Here, too, is our oscillating, upward-sloping sine wave. Centralization took over from decentralization and improved global living standards. With the Industrial Revolution, humans were able to produce more things, and more complex things, more cheaply and efficiently. Without the Industrial Revolution, we would all enjoy a much lower standard of living today. For a few dollars, we can purchase things unavailable to even the wealthiest emperor two centuries ago.
To be clear, I don’t think that either centralization or decentralization is inherently good or bad. As we’ve discussed, a centralized system is put in place to address the shortcomings of the decentralized one before it, and a decentralized one is put in place to address the shortcomings of the centralized one. It’s a progression. Each has its time and place.
In fact, sufficiently advanced technology should make it possible to have either the perfect centralized system or the perfect decentralized system. Utopian Communism might work with perfect data and coordination. Utopian Free Markets could solve Coase’s Theory of the Firm with perfect information and no transaction costs.
What I would argue is that we’re nearing the local maximum for this current era of centralization, and that we’re headed towards an era of decentralization at an even higher level of quality and complexity than the system we inhabit today.
In the background of all of the examples of centralization and decentralization we’ve discussed, there are Vast Formless Things lurking behind the scenes in the form of new technologies. Weapons, ships, machines. It’s hard to imagine the rise of either global empires or large-scale representative democracy without the Printing Press, to use the most obvious example. And so much of how things work today – from the way we organize our companies to the food we eat to the way we educate our kids – is an outgrowth of the Industrial Revolution.
Big societal changes are driven by big technological changes.
Viewed through that lens, in fact, at the scale of empires and societal structure, what’s most surprising about the internet isn’t how much it’s changed the world, but how little.
I think that’s about to change.
The Decentralization Transition
Now it’s time to go through each of the examples I listed up top one-by-one. Some involve things that are happening right now, some involve things that are a little more speculative and future-looking. I’m just pulling threads here.
After we go through them, I’ll try to piece together a quasi-coherent rambling on what I think it means going forward, and what the new operating system for a more decentralized world will look like.
Energy: Fossil Fuels vs. Renewable Energy
I’m putting this one first, because I think it’s the one that will have the biggest impact on the way the world works, and because it’s the one that struck me as the least obvious.
The shift from fossil fuels to renewable energy like solar, wind, geothermal, and nuclear will be decentralizing on a bunch of levels, from the largest to the smallest.
Countries that control their energy control their own destiny; those that don’t, don’t. The Russian invasion of Ukraine provided a stark reminder of this fact. The need to import cheap fossil fuels forces countries to do business with bad actors, and leaves them at the mercy of those bad actors’ actions.
A big challenge with fossil fuels is that they’re not evenly distributed.
Even more challenging, the largest reserves often happen to sit under the countries with some of the world’s worst human rights records (and there’s probably a causal relationship here since the world lets countries get away with worse behavior when we need their energy). Entire economies and national currencies are hamstrung by a dependence on foreign energy. (This Ezra Klein Show episode with Adam Tooze does a good job explaining the dependence.)
Sources of renewable energy, on the other hand, are fairly evenly distributed. Here’s a map of photovoltaic power potential, for example:
In an excellent essay, We’re going to need a lot of solar panels, Terraform Industries founder Casey Handmer observed, “Some places will win the solar power lottery, much as other places have historically ‘won’ the oil lottery.” Unlike the oil lottery, though, every country will be able to pull energy from the sun, especially as the cost of solar declines and the performance of solar panels improves.
While nuclear power is more expensive and technically challenging to build – there are only 32 countries with nuclear power currently – there’s no geological reason that most countries shouldn’t be able to install nuclear capacity as long as they have bodies of water next to which they can place reactors, particularly as advanced small modular reactor technology improves. Nuclear is multiples more efficient than other forms of energy: one large 3,000 ton barge can carry enough Uranium to power the world for a year.
Of course, electricity is just part of the picture. We’ll need fuels too. Terraform Industries is one of the companies working on that. Essentially, it’s betting that as solar costs come down, it will be cheap enough to use a ton of solar power to pull carbon from the air and turn it into synthetic fuel that’s “backwards compatible with existing infrastructure and usage modalities.” If Terraform or a competitor is successful, then any country that has sunlight and CO2 in the air can produce its own diesel, jet fuel, and gas.
We could do similar overviews of wind, geothermal, and other renewable solutions, but the takeaway is this: a world powered by renewable energy is a world with more decentralized geopolitical power. It’s hard to imagine a more decentralizing force at the global scale.
Locally, renewable energy, particularly solar, is a decentralizing force as well. In the US, roughly 3% of single-family detached homes have solar panels installed to power their own homes and even sell electricity back into the grid, a 6x increase since 2014. This will accelerate with the incentives in the Inflation Reduction Act. Producing your own electricity, and even selling some back to the grid, is a more decentralized form of power generation than buying it from large utilities.
Beyond the US, the decentralizing nature of solar power is even more striking. In sunlight-rich Africa, for example, solar might help 600 million people gain electricity without ever having to connect to the grid.
Of course, as more companies come in to accelerate and capitalize on the energy transition, centralization will pop up in the form of new, powerful and profitable companies, but on the whole, the transition to renewable energy will be dramatically decentralizing, putting people and entire nations back in control of the most fundamental resource.
Plus, as renewable brings the cost of energy closer to zero, there will be a host of second- and third-order impacts that are hard to predict. One example that Matt Yglesias gives in this Huge If True episode is vertical farms to feed everyone locally while using drastically less water, and produce more crops at greater efficiency, which would decentralize food production. In fact, most of the examples of the benefits of energy superabundance that Eli Dourado and Austin Vernon lay out in this paper would have a decentralizing effect.
The transition to renewable energy demonstrates that the oscillating sine wave slopes upward: the centralized global energy system has improved humanity’s standards of living tremendously, but it’s imperfect: bad for the environment, geopolitically challenging, and unevenly distributed. Clean energy has the potential to be both more decentralized and superior to that system, bringing more, cheaper energy to more people, cleanly. If we look back in 50 years from a new, more decentralized world order, I would bet that abundant, clean energy was one of the biggest drivers of that decentralization.
Manufacturing: Globalization vs. Reshoring
One way to look at the shift to local renewables is as a reshoring of energy production after two centuries of energy globalization. That same shift is happening in manufacturing.
As I wrote in Formic: Automating Abundance, “After decades of globalization, America is bringing manufacturing back home.” While the benefits of globalization have been clear – more specialization, cheaper prices, more choice – and contributed to low inflation and a higher standard of living, COVID showed just how fragile a position a country is in when it relies on a global supply chain. The threat of conflict with China – America’s factory – exacerbates that fragility, in two ways:
We might lose access to stuff. If China were to stop exporting to the US, so many of the things we take for granted will no longer be available.
We’ve forgotten how to make stuff ourselves. As Balaji Srinivasan highlighted in his 8-hour conversation with Lex Fridman (which is full of gems on centralization vs. decentralization), while the US has a bigger Naval fleet by tonnage, China makes multiples more ships than the US does and could move much more quickly than we could to grow their fleet in wartime. Plus, they’re able to do things like build subways tunnels in weeks instead of decades. By giving China manufacturing in a globalized world, the US allowed them to strengthen while we atrophied.
As a result, a key bipartisan focus in the US has been on bringing manufacturing back home and strengthening our capabilities. From a centralized, global supply chain, the world is moving back to a decentralized system in which each nation develops more of its own manufacturing capacity at home.
But that oscillating sine wave slopes upwards, and technology will play a key role in the transition. Thanks to companies like Formic and Hadrian, and to modern automation capabilities more broadly, I’m optimistic that we can build a more antifragile supply chain, create good jobs, and benefit from abundant, cheap, high-quality goods manufactured at home.
Reshoring won’t be as decentralizing as the energy transition, but a world in which countries can use technology to create more, cheaper, better things at home is more decentralized than the current globalized system.
Manufacturing: Making vs. Growing
Taking it a step further, what if people could grow everything they need?
That’s the question Not Boring’s Elliot Hershberg set out to answer in Atoms Are Local.
Elliot’s vision is that one day, we should be able to connect a personal biomaker to the internet to create the Bionet: a device people could use to download designs and grow anything they need from the atoms in the air around them.
If the clean energy transition continues apace, then by the time the Bionet is possible, anyone with sunlight and atoms around them could produce much of what they need locally. The implications would be enormous. The marginal costs of actual material goods in the physical world could approach the marginal costs of distributing software on the internet, which is to say, we could grow anything practically for free. As Elliot asks, “What if we could grow what we wanted locally? What if our supply chain was just biology?”
While this idea is much more speculative than the previous two examples – Elliot called it “more of a long-term research vision” – it seems feasible that within a century, we’ll have something like a Bionet. That would be massively decentralizing and superior to the current supply chain on a number of important fronts like price, speed, and selection. In fact, the presentation from which Elliot drew much of the inspiration for the piece, by Stanford Professor Drew Endy, is titled Synthetic Biology for Democracy.
Science: Government Funded vs. Decentralized
The way scientific research and translation gets done in the first place is decentralizing, too. For this section, we’ll turn to another piece that Elliot wrote for Not Boring along with Dr. Jocelynn Pearl, Gassing the Miracle Machine.
One of the many areas in which the US government centralized power during and post-WWII was in scientific research. During and after the War, Vannevar Bush structured the way science is funded, and therefore the way it’s done, in the United States. Eric Lander, scientist and former director of the Office of Science and Technology Policy, calls this system the Miracle Machine.
As Elliot and Jocelynn are quick to point out, this centralized Miracle Machine has created wonders: “the Internet, artificial intelligence, cancer immunotherapies, and gene editing technologies like CRISPR.” It’s also become, to steal Elliot’s favorite word, sclerotic.
In the piece, they highlight new mechanisms for funding and doing science that all fit into a bucket called Decentralized Science, or DeSci.
Among the innovations they discuss are new web3 funding mechanisms (either direct contributions from crypto-rich folks or community funding models like those that Vibe Bio and VitaDAO are working on), Fast Grants (like those the Collisons, Tyler Cowen, and Patrick Hsu gave out in under 48 hours each during COVID), and new scientific institutions that are experimenting with every part of how research is done, like Arcadia Science, Arc Institute, and New Science.
These metaexperiments are still early, and it’s too early to tell just how much scientific research will decentralize, but it’s interesting to see signs that a space that’s relied primarily on government funding is pushing towards decentralized models, and it’s a space to watch for signs of just how decentralized the next era is going to be.
As Elliot and Jocelynn concluded, “One of the themes that we’re thinking about right now is the tension between centralization and decentralization. The story for science is no different.”
Hard Tech: Government Agencies & Incumbents vs. Startups
Speaking of categories long-dominated by big, centralized government institutions… hard tech startups are now achieving things that would until recently only have been possible inside a government organization or through government-funded research.
SpaceX is the canonical example here, outperforming NASA to drive down launch costs by an order of magnitude, with another 30x reduction expected with Starship.
Just the fact that a private company could come into the government’s territory and outcompete it points in the direction of decentralization – while SpaceX is a behemoth in its own right, I’d argue a private company’s success is a decentralizing force compared to the US government. But the bigger impact that SpaceX has had is that it’s shown that other startups could do things previously dominated by the government or large, incumbent Defense Primes.
In Making Moonshots, Not Boring’s Rahul Rana wrote that, “Scientific progress is now distributed amongst thousands of venture-backed entrepreneurial scientists that are collectively iterating and advancing their fields in a financially sustainable way.” He points to examples of societal progress being driven by private companies like Relativity Space, Enveda Biosciences, and Tesla.
Anduril is another example here. Despite the fact that it’s been around for five years, it’s already winning DoD contracts against incumbent defense primes like Lockheed Martin, Boeing, and Raytheon. Anduril, in turn, has inspired a new wave of companies to compete for government contracts. One level deeper, Anduril is helping to decentralize the way that wars are fought by putting more capabilities, powered by AI and its operating system, LatticeOS, to put greater capabilities into the hands of each soldier. As war becomes more autonomous, sheer troop count should become less and less of a deciding factor.
I think that we’re just in the earliest innings of distributing the work done by governments and massive companies out to many smaller startups, resulting in faster innovation cycles, superior capabilities, and of course, more decentralization of power.
AI: Closed vs. Open
In a 2018 debate with Reid Hoffman, Peter Thiel argued that “Crypto is libertarian, AI is communist.” In other words, crypto decentralizes power, AI centralizes power. That seemed to be the prevailing view of AI. As Ben Thompson wrote in The AI Unbundling:
If not just data but clean data was presumed to be a prerequisite, then it seemed obvious that massively centralized platforms with the resources to both harvest and clean data — Google, Facebook, etc. — would have a big advantage. This, I would admit, was also a conclusion I was particularly susceptible to, given my focus on Aggregation Theory and its description of how the Internet, contrary to initial assumptions, leads to centralization.
Thompson changed his mind, though: “the widespread assumption — including by yours truly — that AI is fundamentally centralizing may be mistaken.”
In the early innings of large language model (LLM) roll-out, things looked to be centralizing. OpenAI (backed by Microsoft) and Google put out the most impressive LLMs. Then, over the summer, Stability AI dropped Stable Diffusion, a free and open source diffusion model. Stability plans to roll out open source models that can do more than generate images – like fold proteins, write text, and make music – in the near future.
Stability AI’s entrance marked the beginning of a battle between centralization and decentralization in AI that many didn’t see coming, powered by the fact that these models learn from the entire internet, not just clean data. As Thompson wrote:
To the extent that large language models … are dependent not on carefully curated data, but rather on the Internet itself, is the extent to which AI will be democratized, for better or worse.
I think that the outcome of this battle will be as consequential, and potentially even more consequential, than the outcome of the clean energy transition, and there are a ton of open questions, including:
Should the underlying models be centrally controlled with access granted via APIs or open sourced?
If open source, how can we minimize bad actors’ ability to use them maliciously?
How will people be compensated when their work contributes to the output of a model, and how can we even track whose inputs have what impact on the outputs (Anton and I discussed this a little last week)?
As each of us builds up underlying models of ourselves – our images, our voice, our words, our work – in order to have our AI selves work for us, should we trust a company like OpenAI, Google, or Meta to host those models for us, or should we hold them ourselves and provision access when others want to use them?
The centralized vs. decentralized AI battle is still to come, and the winner is unclear, even the fact that it’s up for debate now is a shift towards decentralization from the conversation’s starting point.
Talent: Big Tech vs. Startups
The battle in AI is but one key front in the larger battle between big tech companies and startups. This battle is not new, and it will oscillate itself as plucky startups grow and become big tech companies themselves, clearing room for new plucky startups to attack them.
This thread is the shortest-term-focused of all of the ones I’ve written, but I think that the recent layoffs at big tech companies are going to be a decentralizing force.
The logic here isn’t novel and has been discussed elsewhere, including way back in one of the earliest Not Boring posts, Schumpeter’s Gale. Many people who have been laid off recently were working relatively comfortable, well-paying jobs that they believed to be secure. Some of the people who’ve been laid off will look for new well-paying, comfortable, secure-seeming jobs, but many will get closer to the metal and either start new companies or go work for smaller startups.
Twitter is in the midst of running a wild experiment in just how few employees these big tech companies need in order to thrive. It’s been a longstanding open joke that companies like Google could generate just as much revenue with ~5% of the headcount. If Twitter succeeds, shareholders are going to push even the biggest, most profitable companies to skinny down even more, pushing more people to either even safer jobs, if those exist, or to take bigger risks to make a bigger impact.
As we’ve covered throughout this piece, no industry is safe from the Great Decentralization underfoot, which means that there are going to be opportunities to make big impacts on whichever industry each person believes to be the most important. I wouldn’t be surprised if some of the biggest companies in the next decade are in clean energy, manufacturing, defense, AI, and biotech, speaking to the diversity of opportunities to live that protestant ethic of doing well by doing good.
Plus, the huge companies that seemed untouchable at the start of the year – Meta, Alphabet, Amazon, Netflix, and the like – seem more mortal now. When everything is going well, it’s hard to imagine that today’s largest companies won’t be the largest companies in a decade, but the list of top companies by market cap turns over more frequently than you’d expect.
In the short-term, pushing more talent to the edges will be a decentralizing force. Whether it will be a long-term decentralizing force will be a function of what those people choose to work on instead of strengthening large platforms, and just how many people it will turn out even fast-growing, profitable companies need to scale.
Apps: Big vs. Small
One reason that tech companies won’t need as many people to grow as they used to is that many startups and big companies are building tools that make it easier for other startups to build more with fewer people. Stripe gives you the power of the Collisons in a few lines of code, Unit makes it easy to add banking to your product. For many of the things that a company does, there’s an API for that.
It’s more than just the APIs, though. Some of my favorite companies and protocols are building infrastructure that abstracts away the complexity of the backend almost entirely, making it easy for entrepreneurs to focus on building specific, magical front-end experiences.
As I wrote in Indistinguishable from Magic:
I have a thesis I’m workshopping that there’s a wave of “Small Apps” coming as a result of tools like Replit, Urbit, Ceramic, and web3 more broadly.
Different people find different things magical, and Small Apps allow Magicians to focus on creating magical products tailored to small niches.
I used Farcaster, a web3 social protocol that also makes a client by the same name, as an example, writing, “Small Social Apps, taken together, might steal time and attention from Twitter, a once-magical product that seems so entrenched that it would be hard to defeat alone.”
Again, we’re early in this evolution, but if I had to bet, I would put money on the internet growing from here via an ecosystem of small apps, maybe even personalized apps fed by our AI models, over the emergence of new huge centralized platforms. Of course, that creates the opportunity for centralization to emerge at the discovery layer.
It’s never black or white, centralization or decentralization, but I think that the apps we use will trend towards becoming smaller, more decentralized, more personalized, and even more powerful. The upward sloping sine wave strikes again.
Media: Substack vs. Journalism
Just a few decades ago, Americans had a handful of TV channels to watch and a handful of newspapers to read. They subscribed to the local paper, and maybe The New York Times, The Washington Post, or The Wall Street Journal, depending on what they were interested in.
The internet has blown that model up, putting local newspapers out of business and giving anyone with an internet connection the ability to write to an audience of people interested in consuming a particular flavor of news or analysis.
I’m writing to over 170k of you on Substack, a lot of people for me, but a tiny and targeted audience compared to a media empire like The New York Times. I spend more time reading other Substacks and blogs than I do reading traditional newspapers or watching network news. If Elon has his way, citizen journalism on Twitter will continue to eat away at mainstream media.
That shift has its dangers. As media has fragmented, the spread of misinformation has grown. Truth is more slippery, harder to agree upon, than it used to be.
That said, it seems that the shift of media from centralized to decentralized is inevitably underway. More writers seem to leave big publications to start Substacks every week. It may fragment even further as AI personalizes information based on each and every person’s taste, like TikTok’s For You feed for everything.
It will be important to create systems that help us find our way back to agreement on at least the basic facts, or an understanding of each source’s biases, as this shift continues.
Education: Factory vs. Personalized
In education, too, I sense the breeze of a shift from centralization to decentralization. Recently, I wrote an essay with my thoughts on how Puja and I should educate my two kids, Dev and Maya, called How Do I Teach These Kids?
In it, I discussed that the factory model (centralized), which sprung up during the Industrial Revolution to prepare kids to work in the factories and persists today, no longer seems to be working. Given everything that we’ve covered in this piece about how the world is decentralizing, a model that prepares kids for a more centralized world seems like a tough fit.
Education is a tough one because people have known that our education system isn’t ideal for a while, but to date, there’s been little progress made on fixing it. The startup graveyard is filled with ambitious EdTech companies. YouTube and the internet have been great for leveling up the select few with the motivation to train themselves, but if we’re honest, having all of the world’s information at our fingertips hasn’t been nearly as transformational for education as I would have bet a decade ago.
That said, there does seem to be agreement that in a perfect world, personalized learning would be better than the factory model. In my favorite three-part essay series I’ve come across recently, Erik Hoel writes about the benefits of aristocratic tutoring (Part I, Part II, Part III), essentially pairing kids with knowledgeable adults, parents or otherwise, who are responsible for their education and enrichment.
Aristocratic tutoring is the most decentralized form of education imaginable – each kid gets his or her own tailor-made academic plan – but it’s largely impractical today, and out of reach for all but the wealthiest parents.
That said, there are companies like Synthesis and Primer (Not Boring portfolio company) that offer radically new models of early childhood education, focused on small-group learning and curiosity exploration. Odyssey (Not Boring portfolio company) that lets parents customize their childrens’ education and pay using an Education Savings Account (ESA), giving families at all socioeconomic levels the ability to personalize their childrens’ education more than they would via the public school system. I expect we’ll see attempts at using AI to scale aristocratic tutoring in response to Hoel’s essays, a la Neil Stephenson’s Young Ladies Illustrated Primer.
It’s still early in this shift, but again, over the course of the next century, it’s hard to imagine education becoming more centralized instead of more decentralized, and worse instead of better.
Big Banks vs. Fintech
When Satoshi Nakamoto mined Bitcoin’s genesis block in early January 2009, “he” included the text of that day’s headline from The London Times: “The Times 03/Jan/2009 Chancellor on brink of second bailout for banks.”
In the intervening thirteen years, as the Bitcoin legend has grown, the story that Bitcoin was a response to the GFC has taken hold. The centralized institutions failed us, so Satoshi created bitcoin so that we may decentralize. That story is untrue. Satoshi would have had to have been working on Bitcoin long before the crash in order to drop the white paper on Halloween 2008. The arrow of causality pointed in the wrong direction.
Bitcoin is a symptom of a much larger move back towards decentralization, but there are many symptoms besides Bitcoin, which we’ve been discussing throughout.
Take the GFC again. Finance has become more decentralized in direct response to the GFC in a way that has nothing to do with crypto.
In 2010, Congress passed the Dodd-Frank Act, and with it, the Durbin Amendment. The Durbin Amendment capped the interchange fees banks could charge at 21 cents plus 0.05% of the transaction plus 1 cent for fraud protection costs, versus the 44 cents they were making before. There was an important carve out, though: the Durbin Amendment only applied to banks with more than $10 billion in assets. And that had a big unforeseen consequence: fintech.
In 2010, there was practically no fintech industry to speak of. PayPal existed, and had already been acquired by eBay eight years earlier. Square was founded in 2009 to let merchants take credit card payments by plugging a dongle into their phone. Dwolla launched in 2010 to make mobile payments easier, and reached $1 million in daily transaction volume in seven months. There were a few others, but there weren't many, and none of them tried to compete with the big banks.
Then, the Durbin Amendment hit, giving a structural advantage to small players who could make twice as much off of interchange as the big banks, and could therefore spend more to acquire customers and more to develop better technology and user experiences.
Fintech exploded. From 2013 to 2021, venture funding for fintechs grew 30x from $4.1 billion to $122.6 billion. Exits – via M&A or IPO – grew 11x, from $17.2 billion to $190.2 billion.
Today, there are neobanks serving every customer niche imaginable. There’s even a growing category called embedded fintech – credit, banking, and payroll features inside of verticalized SaaS apps.
From banking in brick and mortar retail locations, to banking online with big banks, to banking in neobank apps, to banking in all of the other apps you use. Finance is decentralizing.
Of course, there are levels to this. One counter-narrative point is that huge banks have actually gained market share since the GFC. Decentralization takes time. Another is that fintechs, while in aggregate more decentralized than a few Too Big Too Fail banks, are still centralized themselves. They have CEOs and teams and the ability to approve or reject customers, to open or close accounts.
That said, the fact remains that individuals and businesses have more choice when it comes to what to do with their money than they did just a decade ago. While rising interest rates would seem to favor big, profitable incumbent banks, with large deposits and no need for venture funding, over fintechs, which won’t benefit as much from increased net interest income and do need to rely on harder-to-come-by venture funding, I expect that in the coming decades, that trend towards more choice and personalization will persist.
These are just a handful of what I view as some of the most significant technology-driven examples of our shift towards decentralization. There are others, and I’d love to hear your thoughts on what I missed or got wrong, but you have things to do so we can move on for now.
While I didn’t list “the internet” as an example of decentralization above, the internet clearly underpins and enables many of the things we’ve discussed. No internet, no AI. No internet, no remotely-controlled robots. No internet, no Substack.
The internet is our best candidate for the Vast Formless Thing driving this coming era of decentralization, but there’s just one problem with that argument:
If anything, to date, the internet has been a centralizing force.
The Internet as a Centralizing Force
While it’s inconvenient for my argument, the fact that the internet has been a centralizing force shouldn’t surprise you if you’re the kind of person who’s read this far into a blog post like this one. Ben Thompson, the GOAT, called it in my favorite series of essays about the internet on the internet: Aggregation Theory.
If you haven’t read the whole series, you should, but for your convenience, I also asked GPT-3 to summarize what Aggregation Theory predicts, and it gave two pretty good answers:
Ben Thompson's Aggregation Theory predicts that online content will become increasingly centralized over time.
Ben Thompson's Aggregation Theory predicts that online platforms will become increasingly powerful as they grow in size and amass more user data.
On the internet, owning distribution and relationships with users kicks off a virtuous cycle: a better user experience attracts users, who attract suppliers, which makes the experience better, which attracts more users, and so on. The dynamics of the internet lead to centralized power, even as it gives more capabilities and connectivity to each person.
My friend Dror Poleg put it in a way that ties even more closely to the theme of this essay in Crypto and the Conservation of Centralization: “When one part of the web gets decentralized, another part necessarily gets more centralized.”
A step further, the fact that the internet has been a centralizing force was inevitable. Gordon Brander explains in an essay titled… Centralization is inevitable.
Centralization is a spectrum, from a network with a single hub (left) to a fully flat network in which each node is equal in power. The type of centralization Brander’s referring to is the middle kind, a scale-free network: power-law distributed with a long tail.
Scale-free networks show up everywhere, from social networks (most people have a few followers, some people have millions) to airports (a few hubs are connected with a lot of airports, most airports are only connected to a few airports) to many natural systems, like ecological food webs.
The internet is a scale-free network, as this mesmerizing video shows, and has become more centralized over time:
The centralizing tendencies of scale-free networks like the internet are hard to break from the outside; they follow a natural order of things, according to Sanjay Jain and Sandeep Krishna (via Brander):
(Phase 1) Random: The system is unstructured. Random events occur without particularly changing the structure.
(Phase 2) Growth: An innovation causes a major phase transition within the structure of the system. The innovation catalyzes other innovations in a positive feedback loop.
(Phase 3) Consolidation: Growth rates saturate. The ecosystem consolidates into a highly organized network, optimized for efficiency, as each agent seeks to eke out as much as it can from its position in the value chain. Hubs (keystone species) appear at critical points.
(Phase 4) Collapse: A random shock, or new innovation demolishes one of the keystone species, causing cascade failure within the highly structured network. The ecosystem collapses into a random structure.
(Repeat): The system begins a slow crawl back up the evolutionary ladder of complexity.
It feels like we’re in the third phase: Consolidation. Like the Fourth Turning or Noah Smith’s big theory, Jain and Krishna’s study of evolution in biology and social systems predict that what comes next is Collapse.
In the case of the internet, maybe generative AI will be the innovation that will “cause cascading failure within the highly structured network, or the internet might re-de-centralize due to a random shock that shakes out of the chaos that comes when everything decentralizes. Maybe it’s my Charlie Day speaking, but after writing all of this down, I don’t find it hard to believe that the internet in a decade or two will look nothing like the internet as we know it today.
That’s a good thing. Rebirth and new experiments make the world slope upwards. And we can do a lot better than this.
The first run of the internet has created and made available a tremendous amount of information, change, wealth, and convenience, but I remember listening to Ben and David talk about whether the gains from the internet had been largely captured on the Acquired episode on Amazon and thinking, “That’s it?”
If one-day shipping, on-demand taxis, and remote work are all the benefits we’ll get from connecting all of humanity, that would be a little disappointing.
Good news about the internet, though: we’re still so early.
The Biggest Impacts of Decentralizing Technology Take Time
Here’s my counter to the “internet as a centralizing force” argument: technology moves really fast, but society takes a long time to change, so technology’s major impacts on society take time.
We’ve talked about the impact of the Printing Press earlier, but those impacts played out over hundreds of years. In 1455, sixty-two years before Martin Luther hit the Catholic Church with his 95 Theses in 1517, Johannes Guttenberg put his invention to work for the first time (~16 years after he began work on it) to print… the Bible. Until then, monks had painstakingly copied the Bible by hand, limiting who could access the Divine Word. The Printing Press helped the Church spread the Bible to the masses – in fact, the Printing Press first came to the American colonies to meet demand for the Bible and other religious texts – strengthening its control. Luther turned the Church’s efforts to spread the good word against it, arguing that the Bible, and not the Church, was the central religious authority in a way that’s hard to imagine he could have without the Printing Press-powered spread of the Bible.
For eighty-one years, when it came to religion, the Printing Press was a centralizing force; Martin Luther, by protesting the Church and inspiring a new Protestant branch of Christianity, turned it into a tool of decentralization. Luther’s efforts tracked Graeber and Wengrow’s three characteristics of freedom: exit, disobey, and create.
The other Puppets of the Printing Press’ Vast Formless Thing described by SSC also took time:
Newspaper: 1609, 154 years After Printing Press (APP). The connection is obvious.
Renaissance: ~1450, right around 0 APP. The Printing Press didn’t lead to the Renaissance, but it accelerated it, making Ancient Greek and Roman texts available to the masses when Venice became the printing capital of Europe in the late 15th century.
Scientific Revolution: 1543, 88 years APP. SSC essentially makes the case that the Printing Press allowed for composability in science for the first time, letting scientists build off the work of other scientists’ discoveries and accelerating progress.
Rise of Nationalism: 1648, 193 years APP. The Thirty Years War between the Catholics and Protestants between 1618-1648, itself a direct result of the Protestant Reformation, established the legal groundwork for the nation-state, and historians view the nation-state as a 19th century European phenomenon.
American Representative Democracy: 1789, 334 years APP. SSC didn’t mention this one, but I’m adding it because it feels like a biggie and pamphlets were an important piece.
Point being, even the most powerful Vast Formless Things can’t reshape society overnight, and in the short-term, they might even have the opposite of what ends up being their long-term effect. Nuclear weapons, as another example, were incredibly destructive in the short term, but have contributed to relative peace among great powers ever since.
The fact that the internet has been disappointing on the decentralization front, in other words, doesn’t mean that the internet won’t be a decentralizing force in the long run.
In What might an end to the Great Stagnation consist of?, economist Tyler Cowen points out that he actually called that it would take twenty years from the 2011 publishing of his book, The Great Stagnation, for the full impact of the internet on progress to be felt:
My The Great Stagnation, published in 2011, offered some pointed predictions. It argued that the “next big thing” was already with us, namely the internet, but we simply hadn’t learned to use it effectively yet. Once we put the internet at the center of many more of our institutions, rather than treating it as an add-on, the great stagnation would end. Numerous times (using roughly a 2011 start date) I predicted that the great stagnation would be over within twenty years time, though not in the next few years…
The gains from truly mobilizing the internet may in fact right now be swamping all of the accumulated obstacles we have put in the way of progress.
I think that this is what’s happening here. The things that we go to the internet to do – search for internet things, talk to internet friends, shop from unlimited internet inventories, etc… – were both, in retrospect, clearly going to be the first things that we did with the internet and clearly prone to the forces of centralization.
But the real, era-defining, world-changing, Vast Formless Thing impact of the internet is coming now, as it infiltrates other industries deeply enough to drive big shifts and “swamp all of the accumulated obstacles we have put in the way of progress.” That takes longer than just moving bits and bytes around the internet, but it’s more important, too.
Which brings me back, of course and inevitably, to crypto.
Crypto & Decentralization
I’ve made it nearly 10,000 words into an essay on decentralization with barely a mention of crypto or web3. This Great Decentralization is happening, with or without crypto.
Just as I don’t think the gains from truly mobilizing the internet have been felt yet because the internet spent its first thirty (sixty?) years mainly doing internet things, I don’t think crypto has even scratched the surface yet because it’s spent its first thirteen years mainly doing crypto things. I think it will have a bigger role in the Great Decentralization to come as we learn to use it more effectively and weave the more useful pieces into the fabric of other decentralizing industries.
Without ever spelling it out in as much depth as I have here, this big societal shift towards decentralization is the core reason I’ve been excited about crypto and web3, and why, despite all of the turmoil recently, I’m still optimistic. It’s also why some of my explanations for why I’m excited about web3 have come off as a little hand-wavy. I can feel that there’s a big shift coming, and whenever there are big shifts, whenever the sine wave oscillates from centralized to decentralized and back, new systems are needed to help orchestrate and govern the new world.
In his 2018 essay Why Decentralization Matters, Chris Dixon makes a now-famous case that Web 1.0 was decentralized but messy and hard to use, and Web 2.0 centralized things and made them easier to use. That came with a catch.
As platforms grew stronger, and they became more densely connected nodes in the internet ecosystem, they moved from “attract” – taking actions to acquire users and supply – to “extract” – squeezing as much out of those groups as they could get away with. This sounds like Jain and Krishna’s Consolidation phase.
The answer, according to Dixon, are cryptonetworks, which among other potential advantages, are inhibited from switching from attract to extract “through mechanisms for voice and exit.” Participants can participate in community governance (“voice”) and leave and sell their tokens or even fork the protocol (“exit”). Those sound an awful lot like the first two freedoms identified by Graeber and Wengrow – the freedom to get up and leave (exit) and the freedom to disobey (voice), which “act as a kind of scaffolding” for the third.
As I wrote in The Laboratory for Complex Problems, in slightly different words because I hadn’t yet read The Dawn of Everything, I think that web3 can be a playground for the third freedom: the freedom to shape entirely new social realities, or shift back and forth between different ones.
The world is decentralizing as the internet alters more of what we do. Power and resources are pushing towards the edge, however slowly, inconsistently, and messily. If I’m right in all of this, new models of governance and ownership not intermediated by powerful third parties will be needed, not for ideological reasons or because we want the price of a token to go up but because those powerful third parties will lose power, and a more chaotic, decentralized world will replace it. If nothing else, it’s worth continuing to experiment and evolve our models, to play in order to discover new social realities, while the stakes are still relatively low.
🚨 Now there’s a huge caveat here. 🚨
Crypto will have an important role to play in the decentralized era if people in crypto don’t fuck it up before it has a chance to be really useful.
By that I mean that I genuinely believe that if crypto, a decentralized, permissionless system that lets people own their digital money and things, is to play the role I think it can in many of the transitions we talked about today, it’s going to need to fix its toxic image by quickly moving to fix out the toxic parts of the system. It would be hard for me to recommend that an entrepreneur in another industry consider incorporating crypto into their product with a straight face after the FTX blow up, even if I think it could be genuinely useful.
The gambling and scams will always be there; gambling and scams still riddle the internet, but their share of internet activity has declined as they’ve been drowned out by more valuable use cases. The same thing has to happen in crypto for it to play the role that it might.
There are areas we’ve discussed in which I really believe web3 primitives will be valuable.
AI holds the most obvious applications, partially because it’s the most purely digital of the shifts we covered.
If open source AI models win out, for example, I think it will be crucially important to create systems for users and developers to govern the use of the models. The lower-stakes experimentation done on DAO governance should provide a useful toolkit.
Going forward, the more people who sign their work – writing, art, music, code, etc… – with their wallets, the easier it will be to find a solution to the attribution and payment problem that’s going to get much worse before it gets better.
And I find it uncomfortable to imagine little Packy Models stored in someone else’s servers. I’d prefer to custody my Packy Models even more than I’d prefer to custody my money. To that end, Vana is building a data vault that lets people own and control both their data and the AI models trained on it. I think it has the chance to be one of the most important companies in our portfolio.
There are clear roles for web3 primitives in the proliferation of Small Apps, Finance, DeSci, and the Bionet too, and probably less clear ones in the other categories we discussed, as well.
Interestingly, as McCarthy pointed out, the original Luddites, the ones who burned down factories, weren’t opposed to technology itself, to the automated machines and looms, but to the fact that they didn’t own or govern any piece of the machines themselves. As we develop more powerful technologies and automate more of what we do, there’s probably a lesson in there.
As it stands today, on November 21, 2022, things look bleak for crypto. Prices are way down. Belief in the space is even lower. Skeptics are rightly pointing out all of the things that went wrong and all of the things that don’t work.
Web3 technology will improve – as one example, zero-knowledge proofs will make blockchains more scalable and private – and governance models will evolve. I think the technology and models will be genuinely useful across so many of the industries we’ve discussed; the question is whether there will be too much stigma for anyone to adopt it when the time comes.
To be clear, though, decentralization is happening with or without crypto. What I got wrong is believing that crypto could be the force that brought about decentralization; now, I don’t think there’s a single force, just a whole lot of things subtly moving in the same direction. The centralization phase will end when the centralization phase ends.
At the end of his essay, Brander reaches a similar conclusion: “It seems unlikely that the cycle of network samsara can be broken or reversed.”
Centralization will reign until the Collapse, and only then can a more decentralized network thrive, a la the Fourth Turning’s “bone-jarring Crises” before a wholly new form of American society or Noah Smith’s War before New Order. Brander concludes:
But perhaps we can lean into this cycle of rebirth and build new decentralized things, throw new parties. Perhaps we can build many decentralized things, and construct a rich ecology of interacting systems at different stages of growth, where any one extinction does not take everything down with it.
I agree with that take, and think it provides a useful framework through which to view products in the decentralization era. If the thesis is correct, it will mean more, smaller companies as opposed to a re-emergence of “new multi-trillion dollar platforms, but crypto.”
To the extent there are enormous entities created in the next wave, the idea that they’ll be user-owned, community-governed protocols that support an ecosystem of small apps certainly fits the thesis of the piece better than the idea that there will be new, huge centrally-controlled platforms. Had I written this all down a couple of months ago, maybe I would have seen FTX coming.
It’s worth noting that after thinking through the powerful forces driving the centralization of the web, Brander recently launched Noosphere, a decentralized protocol for thought that aims to address a key problems with humans: “Civilizations scale until they are overwhelmed by the information environment they create.”
Think of your information environment. Does it feel like we’re reaching a point of overwhelm?
It does to me. I think it’s another sign that we’re entering a new phase.
Towards a More Decentralized World
It’s become clear to me that we’re heading towards one of those big moments in history during which an old system gives way to a new one. I think that the vector of change will be a move from centralization to decentralization, for the reasons I’ve laid out throughout this piece.
This is an idea I’ve been wrestling with for a long time – Power to the Person, The Great Online Game, and The Cooperation Economy are three in which I’ve wrestled directly with similar ideas – and one that, if any of this proves to be remotely true, I’ll be wrestling with for years to come. This piece was my first attempt to put it all in one place and start thinking through how it all fits together. If you’ve made it this far, thanks for sticking with me.
Throughout the essay, I highlighted eleven areas that seem to be moving from more centralized to more decentralized, and did a very quick spin through some relevant history, but let me be clear: I have no idea what this new world order will look like, when it will come, or what it will take to get there.
The ultra-optimistic case is that these shifts we discussed – from fossil fuels to renewable energies, from a globalized economy to a more local one, from pre-AI to AI – will play out naturally, that the people involved will make smart choices every step of the way, and that unlike previous Turnings or Noah Smith cycles, we’ll avoid a Crisis or a War. Maybe COVID counts; maybe the Ukraine War counts and ends without escalation; maybe we can just move on to the new order.
The less optimistic case is that things are going to get worse before they get better, as things have gotten worse before they’ve gotten better in the past. Certainly, there are a number of candidates for Crisis visibly stalking in plain sight, and undoubtedly many more lurk in the shadows. I hope that we can avoid them, and I do think that the emerging world order in which power is more evenly distributed will be more resilient to global-scale crises in the future. Think of all the wars that have been fought for oil, and imagine what happens in a century when everyone has all the energy they need.
However turbulent things feel right now, and however rough they get in the next few years, I’m excited for what awaits on the other side. The wave trends upwards. The next decentralization era might be defined by both more freedom and more abundance. It’s hard to study what’s happening in energy, biotech, manufacturing, space, AI, and yes, even web3, and not feel optimistic that a bunch of things are going to change for the better in rapid succession, and often in concert.
The Vast Formless Thing – the technological breakthrough driving the next era, whether the internet, clean energy, AI, or something else – will push the world in all sorts of directions that are hard to predict from where we sit today, and that will take time to play out. Looking back on today from a century, or even a millennium, out, I wonder what things to come will look obvious to our descendents, with the benefit of hindsight.
Decentralization is messy. People are imperfect. Pushing power out from the center to the edges introduces more randomness into the system. But I hope that the messiness lets us evolve new systems, structures, technologies, and ways of living that aren’t possible in a more controlled, centralized system.
We just scratched the surface today. There’s so much depth and nuance that more knowledgeable people could add to each one of the industry shifts we discussed, so many industry shifts that we didn’t touch on, and so many second and third and fourth-order effects of the intersections of all of them that I’m not nearly smart enough to predict. We didn’t even really talk about space! There’s nothing more decentralizing than humans colonizing new worlds and choosing to govern them in new ways… but that’s another post for another day.
I’m sure I got a lot wrong, too. My goal was just to put this idea out there so we can evolve it together. When everything is more decentralized, the only way we’ll be able to make sense of it all is by working together.
Thanks to Dan for editing this beast — I’m sorry for making you read so many words.
That’s all for today. Hope you have a great Thanksgiving if you celebrate, and hope you enjoy all the Americans being offline for a little bit if you don’t! We’ll be back in your inbox next Monday morning.
Thanks for reading,
Packy
Net returns refers to the annualized internal rate of return net of all fees and costs, calculated from the offering closing date to the date the sale is consummated. IRR may not be indicative of Masterworks paintings not yet sold and past performance is not indicative of future results. See important Reg A disclosures: Masterworks.com/cd
Nice essay.
It is not clear to me that "decentralization" holds the same meaning throughout your essay (e.g., when comparing energy production to education), or more generally if "decentralization" is well-defined as you use it (if, say, there is one runaway market winner in the a la carte market for education, as there usually will be given the power laws that dominate modern marketing/business success, does that entail more or less centralization than the status quo?). Even so, it's a stimulating theme.
In general, it seems to me centralization & hierarchy are at an advantage when what is needed is univocity of command and faithful, speedy execution, while decentralization will tend to predominate when what is necessary is a consideration of a wide variety of data sources & speedy execution is not at a premium. Agents jointly create the conditions for their next game -- advantages are spent as resources are exploited to the hilt, say; when there's no more map to discover, the race to the fastest ship, not the most knowledgeable navigator. One can thus see why centralization and decentralization trade off over time across different endeavors -- and, in my view, they are unlikely to do so in fields that've not reached a final equilibrium.
Wow, I was not expecting such a vast study in one article. Thank you so much for taking the time to prepare. I'm very interested in decentralization myself and will be surely returning to this piece many times in the future. Subscribing.