Mike Pepi: Political Axes for the 21st Century
Mike Pepi is a New York based writer and critic in the fields of art and technology. He is the author of Against Platforms (2025). You can read his writing on Substack at Heavy Machinery.
It has become increasingly indefensible to plot ideas, people, or organizations across our existing political spectrum. The traditional left/right and authoritarian/libertarian models are outdated and misleading. Though they might have been useful in the twilight of the 20th Century, they ignore the complexities of today’s digital, networked economy. We need a new framework to grasp emerging powers and their political risks.
The formation of the MAGA/tech-right alliance is just one recent, confounding factor. Where would you plot Elon Musk and his DOGE task force? Or what about President Trump’s strategic Federal reserve of Cryptocurrency? How about AI GOV, a “revolutionary framework in public beta, turning X discussions into actionable, data-driven policies”? In recent years, many new dimensions and axes have been added to the classic “political compass.” Yet none succinctly capture the most important struggle of our time.
In the final chapter of Against Platforms: Surviving Digital Utopia, I ask that we re-imagine the axes for organizing political ideologies. I propose, instead, that the great conflict of our era exists between that of platforms and institutions. The fight for governance today isn’t related to your position in the French National Assembly in 1789, but about whether private digital platforms or democratic institutions should control the public sphere.
My new political compass is completed by a second vertical vector between humanism, which, after the death of God and the rise of “thinking” machines, can still serve as a “a system of thought in which human values, interests, and dignity are considered especially important”1 and technological determinism where “technology is understood to determine, in fairly linear simple cause-and-effect ways, the social, cultural, political, and economic aspects of people’s lives”2 Today, political actors must consider whether they are more attuned to the needs of human flourishing or a retroactive response to affordances of new technological capabilities.
Having defined humanism and technological determinism, I’ll also venture a definition of the “newer” arrivals on the scene: institutionalism and platformism.
In Part 1 of this series, I did a side-by-side comparison of the institutional form and the platform: Part 1 - The Institutional Membrane 3
Institutions are a series of habits or practices, sometimes organized into a cohesive unit, that carry out the production of objects, expressions, or political goods and services. Institutions enact a mission, a conviction, or belief in spite of external influences that might compromise it. “Institutionalness” comes most into focus when we witness them persist in spite of a prevailing wisdom or opinion, not because they are unpopular or anti-democratic, but because they stand for something. That something — an idea, a good, or a norm — is reasonably fixed for the purpose of the constituents of the institution itself.
This fixity is anathema to the platform, which must always be a passive vehicle for programmatic optimization towards a goal. To the extent that one of those signal/goals is user satisfaction, then it becomes a doormat for the fulfillment of those targets. The platform is only populist or “democratic” as an unintended, temporary mishap — its sole goal is to scale its mediation into a monopoly to earn investors a return on capital.
A platform is something that only works at scale. You can have a “small” platform, but it’s not going to last long. An institution sustains - and happily exists - even at the smallest volumes because the prime objective is that the mission is enacted without regard for expansion and competition.
Platforms use computation to optimize; yet institutions are difficult to optimize without violating their very raison d’etre. Institutions care that information is “correct”, or that its provenance is sound and useful. Platforms care only that its information and that it might host its transaction. Institutions need not quantify if they qualitatively satisfy a purpose. Their goal is that the thing is sustained, not re-engineered, monopolized, or gamified to benefit participants that want to race into the sweatshop-optimization of their craft.
A close reader will notice a gap in this picture of institutions. The term is a floating signifier. On one hand, we have public institutions (as in “democratic institutions” like the ones we find in both buildings and enshrined in legal documents and precedents. We also have cultural and social institutions (hospitals, museums, and universities): things that provide services and sustain core functions of civic and personal well-being. Then there are the harder-to-pin-down practices, where I would group things like the institution of marriage, art, or the novel. I argue that the throughline between all of these is an idea of a greater responsibility to a public, and not, as contrasted with the platform, which reserves its highest purpose for submitting actions to computation, competition, and optimization. For the platform, the public, known as users, are just another node in the network architecture, often rather low on the hierarchy of the platforms component parts.
A final necessary caveat: these are not black and white concepts, more like two ends of a spectrum. To that end, here is a breakdown of the four quadrants of the new axes of institutions and platforms.
The New Quadrants
Humanism ←→ Platformism
Web3 “decentralized” culture
What we have come to know as “decentralized”, web3, or blockchain culture exists between the poles of humanism, where artists are empowered by economic and social arrangements to exact the full expression of their will, and platformism which enlists powerful and scalable digital technology to host the infrastructure for this vision. As shown by the chart, it’s not pure platformism, since it does engage in a modicum of suspicion for the platforms’ centralization and control. Its hybrid vision is a series of federalized and robust monetization schemes by which many markets compete to financially reward creative production.
Since the rise of blockchain ledgers that enable art objects to be instantiated as digital assets, many thousands of attempts at so-called web3 projects have attempted to build a framework by which artists can be self-determined and free from the shackles of centralized power or institutional gatekeepers. The underlying conceit, however, is fatally flawed. It imagined that an endless longtail of possible consumers of niche cultural products exists. “Minting” work onto a distributed network - where every exchange is a financial investment into an underlying currency - might produce a balanced ecosystem for sustainable artistic flourishing.
This argument arrived at a time when creators were being blocked from financial remuneration by the free digital labor ethos of Silicon Valley platforms that wanted to remove any kind of economic boundary to uploading, creating, and sharing. The "free software" utopianism was sloppily grafted onto "culture" and “creators.” For a good decade there it was de riguer for most online creative work to be free. The backlash was swift, but like all radical reactions, swung too far in the opposing direction. The solution was to leverage various protocols (basically, public databases that lie underneath cryptographically mined assets) to incessantly track and monetize every single moment of cultural production. Millions of NFTs were created in editions. Many artists saw overnight success on the new tulip mania of digital currency exchanges. But absent any kind of central, institutional broker, the only way this was workable on a continuous basis was through the incentivization of wild speculation and unchecked exuberance. We have yet to see a sustained distributed protocol, platform, or cultural “community” that hasn’t crashed out under its own speculative weight.
Zora Labs launched in 2020 as a “marketplace catering to the needs of NFT enthusiasts”. As NFT values plummeted, today they rebranded to an “onchain social network revealing new opportunities to create, connect, and earn from your life online.” In a recent update, the platform announced a new feature: every post to the marketplace can be turned into a memecoin. The announcement from the @ Zora Twitter account states:
A swift rebuke came from digital artist @ blackdave. He challenged the idea that artists would view the memecoin trading as the right method to engage with their audience.
Web3 wasn't a utopia of art creation - it required artists to also be a kind of stock broker; an un-registered investment professional doling out financial advice alongside their chromatic abstractions minted on this week's hottest blockchain protocol.
You see, the problem with the “free market” is, well, that it’s free. The unchecked mechanism of overt financialization’s dark side is that, while for a moment it serves to pay artists, the longer term trend is that of re-platformitization. Entrepreneurs looks to capitalize on hosting this new marketplace, landing us back at square one of web 2.0. This is the mistake of building in this quadrant. No matter how well intended, digital platforms that eschew institutional buffers end up destroying the very market they set out to create.
Platformism ←→ Technological Determinism
Right Wing Digital Utopian Techno-Feudalism
Moving on, in the lower right hand corner you have the cursed wilderness between full platformism and technological determinism. Here there is no rest for the weariness of traditional institutions, nor is there a functional reason for the infrastructure that institutions provide. Both the little "i" institution — the unseen, social, habitual, ritual frameworks that define our consumptive patterns — and the big "I" institutions (the newspapers, museums, and legal courts) are seen as so much dead wood.
We saw the antecedents of this attitude in John Perry Barlow’s Declaration of the Independence of Cyberspace (1996) and James Dale Davidson and William Rees-Mogg’s The Sovereign Individual: Mastering the Transition to the Information Age (1997).
The technocrat recognizes that there is transactional power in cultural exchange, but looks at his material conditions and judges anything outside of computation to be a folly. Thus the means of transmission and exchange adopts the logic of computation - efficiency, scale, and networking become the new muses and virtues. This is "right wing" in that it must remove any contentions to subjectivity - an authoritarian power is seated at the head. “Liberal” cultural institutions made too much hay about cultural relativism - now the computer and the model will determine what the audience wants and what will be shown.
You can be a techno-determinist without being a platformist, of course. There is a little techno-determinist in all of us. You can believe that advancing technology is the prime engine of history yet still understand the underlying political need for institutions. (Spoiler alert: an institution can be driven by techno-determinist principles) True platformists, though, think that institutions should be abolished, and their “functions” run on digital rails, if not through AI. This way they can be optimized down to zero waste, a perfectly efficient exchange of the network actors’ mean of weighted interests. This computational equilibrium was the initial dream of cybernetics. Any crooked timber is now binary code for the model’s objective powers. The pursuit of this technical achievement is techno-determinist, clearly, but it's also a political endorsement of the platform as being endpoint of all endeavors. A platform exists to optimize for the administrative server’s goals, and little else.
Platformists fetishize growth and efficiency over "stuff" of the institution. It's why it returns things like Substack and Patreon. Those platforms use the ghosts of the media ecosystem past to mobilize exchange on their social media networks, but the point is to extract value from this decaying media ecosystem while they rebuild not so much a better way for journalism to be done than to make money on decline towards an ecosystem of social media feed slop.
The user of these platforms - both the payer/consumer and the writer/producer are entered into a zero-sum race to the bottom where they both do more work to receive less value. For the payer/consumer, they have been enticed into an arrangement where they pay more than they would have if an institutional system of newspapers and magazines was in place to productively subsidize the end product by way of forcing other stakeholders into the economic exchange.
On platforms, your work goes primarily into building the platform itself. It has only an incidental relation to the underlying subject. This is why the platform surfaces metrics related to interaction and growth at every turn. The object here is the individual user, not in the sense that it serves your needs but that your needs (your interactions and the potential revenue you represent) serve the platform.
Consider that when you “discover new music” you are also in some deeper sense re- discovering the platform. The music you discover isn’t for you, it’s for them, the host. And what they host, and why, is of less concern. They are there for you (as user), and you are there for them (a distributor). The institution doesn't care about how its production is discovered or interacted with so much as it cares about the things you will encounter.
Institutions don’t need optimization via software - that doesn’t mean they cant use software or digital media - but it means the introduction of those faculties must not usurp the basic mission of the institution. Think “a website for MoMA” rather than “Google Art Project” that hosts MoMA’s image data in a private cloud…or a hotel that exists as a node in a civic landscape rather than Airbnb’s network for optimally finding a place to sleep. Is “a place to sleep” the sole goal of a hotel? Of course, not. But in the eyes of the techno-determinist the hotel is an obstacle in the way of the perfect solution.
That’s because the platform enters the institution as a business problem to be optimized. Organizations that have distinct forms for the production and the distribution of their content are risky endeavors. Institutions happily assume that risk because they believe in the underlying purpose of the content. But investors soon discovered that managing the distribution in a manner in which the production is bundled together with the distribution are able to cut out any misallocated expenditures. Ride-sharing companies are racing toward self-driving fleets; delivery apps have “ghost kitchens” where the app itself runs the restaurant; and perhaps most freakishly, Spotify has playlists comprised entirely of AI music so no artists are paid.
The nascent world of AI cultural production takes this to its logical end. All of the data is there for you and the model’s statistical inference enables frictionless production, provided you interact with the feedback system by which you also contribute data to the training. Your very use of the AI system for cultural and intellectual production entrenches its utility while removing the demand for any new production from the original source material that made up its training data.
Gen AI is so promiscuously marketed and loosely defined as a means to convince the audience of its yet unrealized intellectual potential; while at the same time it ambles towards its goal to replace the institutional brokers that once defined what you saw and how you might see it. Gen AI attempts to be both the author, audience, judge, and institutional support. The "model" combines all four of these formerly-separate functions. The only work being done to legitimize this is the human act of looking at the outputs and conferring on them any kind of aesthetic legitimacy. If you like the slop, or you see yourself in the reflected combinatorial image, then you do the double work of endorsing the new model arrangement in which the institution is removed.
The slopstate we now see on digital platforms is the logical extension of both of these formations. Above, I noted that “institutions care that information is correct; platforms just care that it is information.” The same is true here at this new dark nexus: if computation can achieve an end result of an aesthetic object of any demonstrated value, then nothing should stand in its way.
You must keep this in mind when you look around online now - the economic activity isn’t so much “on” or “within” platforms as it is to create them anew; to find the last cognitive value scraps upon which to compete economically at the meta layer of genAI APIs. Yet what the techno-optimists fail to realize is that there is no larger point to this speedrun if humans don’t fundamentally demand this kind of cultural interaction. Silicon Valley’s attempts at cultural production will always be frustrated by the one thing that its tools cannot automate: our institutions. In order for Silicon Valley's version of creativity to triumph, it must remove these institutions. This is a slow and deliberate process and it is, recently, seems politically possible (if less likely culturally). This is why you see the present administration’s war against institutions of all stripes: they are even harder to automate away than human “intelligence”.
The first method of destroying institutions is by rhetoric about their uselessness, for which they appeal to popular utilitarian positions (an anti-establishment croon as old as time). The next way, though, is more insidious and often more effective: by flooding the zone with tools based on a false promise of empowerment, but predicated on theft of intellectual property.
Here's the catch: institutions always win. No matter how fancy and useful, no matter how new or sleek or popular a tool becomes, no matter how many creators suddenly feel that it has reconstituted their creative process and opened up new avenues for their expression, the institution will rule over all. They alone are the keepers of the important distinction between the action (or production) and its result (the object).
In the institution we have the glorious "alignment" that Silicon Valley types so thoroughly seek. For humanists, we intuitively know that institutions are the fabric of meaning. Art isn't some objective category "out there", liable to logical proofs—art is simply the agreement of many individuals, usually codified in what we give a loose name to the institutional form. This relativism drives Silicon Valley nuts. "No!" they say, "You can't just say art is what we collectively agree on!" This is devastating to them because it's not falsifiable; they can't logically prove anything superior when the entire endeavor refuses to be judged by measurable criteria.
Silicon Valley's aesthetic framework is no match for this due to the simple reason that it places the machine and tool above the symbol and the human. Along the way, they also miss the point of the broader human endeavor of creativity. You can ignore anyone who claims AI is autonomously “creative”, and you should skeptically probe at anyone who believes that computation can reproduce the work of institutions. Forget computation and models. The reason we have genAI that makes slop language is due to the invention of the platform. It was those incentives that invented the machine that reduced costs and removed intention - the final triumph of the growth-at-all costs mentality.
Humanism ←→ Institutionalism
Traditionalism, “Decel”
What this graph refers to as traditionalism, or in their enemy’s parlance "decel" is the cozy, familiar world of 20th century institutions. Here lie the “gatekeepers.” In culture and politics, from art to anesthesiologists, it is centrally important that they control the quality of what they enable, train, and “institute”. These are, one hastens to say, not all too concerned with technological obsolescence. They are models that insist that what they enable and produce - the very worlds they build - are valuable enough in and of themselves to prohibit any kind of update in response to the latest trends in digital capitalism. They must make money to sustain themselves, but they are insulated by tradition, largesse - public and private - and the basic assumptions that their existence is fully coterminous with the aesthetic and/or social value of the produce themselves.
Cultural institutions provide the easiest illustrations of this structure. However, the concept applies across a number of fields and publics. The 21st centuries rust to digital disruption has not properly accounted for the ways in which the institutional structure itself plays a critical gravitational role - the many things orbiting around an institution - the work, the social production, the ideas, and the economic value - simply implode once they lose the orbital pull of a central, agreed upon institution to define them.
For example, one simply does not imagine modern art as an idea existing without the slow, powerful work of an institution like MoMA creating the intellectual and social market for such a mode. When I say MoMA, I mean “the” MoMA in NYC, but also all the other Museums of Modern Art around the world. Similarly, academic publishing and the kind of belles lettres essay writing that decorated the 20th century literary world would sink like a brick without the lifeblood breathed into it by publications like the New York Review of Books, The New Yorker, or The Paris Review. For many, there isn't really a need to justify these institutions. The tautology comes alive here, admittedly - they are organizations which without the very constitution of the endeavor would be fundamentally altered. Recently, we’ve begun to understand just we will miss when they are gone.
It’s a productive, humanist trap. It doesn’t rely on any other incentive pressure outside of the belief in the underlying formal expression is a singular component of what it means to be in the world. The New York Review of Books thinks you should read books. They need the format; and the format, in some sense, needs them. Thus they plod on without an urge to change their appeals. Many wear this as a badge of pride. The philistines of Silicon Valley look upon this curiosity first with bewilderment and then, all of the sudden, with scorn.
While I identify with these institutions on a personal and political level, I do not think this is the way forward. The task at hand to understand these as useful artifacts of an ideal model that does not have a role in the coming century. Even if we believe in the values of the terminal point on the spectrum of institutionalism, when their implementations take too safe a comfort in humanism they sacrifice their ability to expand along novel technological configurations.
Their ghosts will live on, by which I mean their intellectual and moral campaigns will endure. They have been faithful, if hegemonic, reflections of the ruling class, reproducers of those class relations, and much much more (see “Institutional Critique”). As we have seen, after platforms, their advantages outweigh their disadvantages. It’s their support, financial and epistemological, that has conditioned us to want to keep them around. Spiritually, this will find a venue to re-inhabit. Institutions are like cockroaches. They can never truly be killed off. But their physical, economic, and social components should be re-encased in a new organizational form.
Let’s Build Here! Institutionalism ←→ Technological Determinism
Intra-institutional technology
Recall the Platform Era (2010 -2020). It was boom times for “free culture”, a large-scale transfer of previously un-digitized cultural objects to the networks, and eventually centralized platforms, of the present. This surfeit of information and exchange paved the way - culturally and technically - for the AI era (2014 - present). Catapulted out of the AI winter by the advent of user friendly and functionally coherent Generative Adversarial Networks circa 2014, the focus of Silicon Valley’s intellectual and financial propaganda shifted rapidly. AI — which is through and through a marketing term for “machine learning” and mathematical models — has now moved into the territory of cultural institutions. Why? Well, because that is where political value lies most sweet.
Like every cycle of disruption, Silicon Valley must intervene and capture the social and cultural narrative from the non-digital meaning-making frameworks. It must reproduce this “meaning” through its proprietary software-as-a-service model. And it must do this all the while optimizing the outcomes for the managers of the computational platform and its owners. All of this was buttressed by the decline of the role of the cultural institution, and "institutions" writ large.
What the platformists and the techno-determinists misunderstand is that driving automation costs to zero, and thus automating most “economic tasks”, will only reaffirm the importance and role of human institutions in constituting what is consumed and served. By making the economic concerns obsolete, it will remove the market’s authority in judging what is worth producing and worth consuming. The trick, then, is to re-situation institutional membranes in the new computational arrangement, where digital tools might liberate creativity but not to the unilateral gain of large, for-profit platforms incentivized only to grow.
While it may feel like we are on a one-way track towards this all-platform dystopia, be aware of the recoiled opposing flow:
With the platform era coming to a close (partially due to AI slop’s overreach) and the AI regime supposedly driving creation costs to zero, somewhat counterintuitively, cultural institutions will soon enter a new growth phase. Once we have retreated from platform transactions, there will be A new premium put on human provenance, authenticity, and non-transactional connection. As AI slop fizzles out into its own endless machine loop, institutions are fundamentally elevated to new and essential brokering function.
Techno-optimists, VCs, e/acc, and digital utopians can't shake their association of institutions with what they are supposed to "disrupt" and "optimize." They are in for a real surprise when they realize that those formations were fueling the entire optimization in the first place. You need something underneath it all. What these coming institutions will look like and how they will sustain within digital frameworks is a question of political design.
In short, these new institutions don’t need to be afraid of digital media. As the chart indicates, their functions should be determined by the new affordances of digital technology. They cannot, like the traditional institutions (in the upper right) ignore the technological change around them. A new type of organization must emerge - one that employs “intra-institutional” technology. The ideology underneath and within these new institutions will be a kind of enlightened techno-progressivism, defined as the idea that digital tools can augment the missions of institutions without cannibalizing them whole from the outside.
The strongest of the democratic and/or intellectual qualities of institutions can still exist in their fullest expressions while being encased in, and elasticated by, software tools. This need not be a contradiction in terms.
We are in a profound and transitional period. In the next several years we will be confronted with the ever more disruptive collapse of existing institutions of various stripes. Though this time, the destruction will be directly and purposefully enacted by the agents of the platform installing AI and other extractive technologies into the administration of public life. Our answer to this assault cannot just be to make a new, different “improved” platform. We must reconfigure institutions for the age of ubiquitous computation. It’s no small task, but I think it’s the most important thing we can work on.
Mike Pepi is a technologist and author who has written widely about the intersection between culture and the Internet. An art critic and theorist, he self-identifies as part of the “tech left” – digital natives who want to reshape technology as a force for public service. His writing has been published in Spike, Frieze, e-flux, and other venues. His book Against Platforms: Surviving Digital Utopia, was published by Melville House in 2025 - “An unsparing exposé of how digital platforms stifle personal and collective efficacy”4
1 Law, Stephen, 'Introduction: what is humanism?', Humanism: A Very Short Introduction, Very Short Introductions (Oxford, 2011; online edn, Oxford Academic, 24 Sept. 2013), https://doi.org/10.1093/actrade/9780199553648.003.0001
2 Rogers, Alisdair, Noel Castree, and Rob Kitchin. "technological determinism." A Dictionary of Human Geography. : Oxford University Press, , 2013. Oxford Reference. https://www.oxfordreference.com/view/10.1093/acref/9780199599868.001.0001/acref-9780199599868-e-1852
3 Best viewed on desktop!
4 https://www.kirkusreviews.com/book-reviews/mike-pepi/against-platforms/
 | A guest post by
|
This is pretty insightful and did a lot to help me better understand the current moment. I see it as a critical supplement to the authoritarian/libertarian model which makes some of the strange bedfellows and contradictions in today's politics make more sense.
The key differentiator I see between platforms and institutions is platforms are have at least one infinite or all encompassing dimension by definition, and institutions are always limited in scope by definition. YouTube is all the long from self published video, Substack is all the writing worth paying for, Instagram is all the photos, Internet Protocol is all the network data packets, and so on. Institutions have a more specific scope, MoMa is a set of buildings and people that can only deal with so much art, the United States is a country with specific borders. Platforms end up always having to define their "everything" within the limits that the institutions they interact with allow. While a maximal position would be allow porn on YouTube, the institutions say no.
The internet, a technology designed to be decentralized, has enabled the greatest centralization ever in the form of these platforms. Part of the problem is human nature. Having all the things put into consistent structures we can learn once and use forever makes things easier. Consistent structures can happen by consensus, or by central control. Bad actors enter in the mix and try to manipulate the consensus against its goals for their gain, putting the gatekeepers back in place. (Think of how email is full of spam, or how usenet was shut down for all sorts of illegal crap).
So what is the future? I think Discord is pretty interesting. It is a platform of platforms. It replicated the decentralized IRC which fell to the problems I mentioned before, but with a light touch to governance that has allowed for a lot of various communities to thrive. Each has a consistent set of tools for building lots of different experiences around different topics. There are lots of discord servers the mainstream would have a problem with, but the point is, no server is for everyone. This replicates reality is a sense. We cannot all hang out in the same global bar, there would end up being fights. Each community is its own institutions of sorts, within a platform of institutions. We can each choose which ones we participate in, or not. It is no panacea, but space for a multiplicity of approaches provides the relief valve that can let us all get along.