This is the first draft of an article I wrote on behalf of Disjunctions Magazine. For the fully polished piece, you can find the link on their site here. The text below is kept purely as an archive of cut content and footnotes.

What does it actually mean to take on Big Tech? When we apply this language of conflict, do we take seriously the real possibilities of winning and losing brought with it? This language cannot be reduced down to terms of discourses, solidarities, and tactics. Taken in its own terms, conflict is a fundamentally strategic exercise.

As labor historian John Womack puts it, a strategic perspective requires “calculating the most probable powers and fields of the forces in conflict for the period you intend to fight… what you can win or lose.. what you most want to win… what you have to win, and what you cannot risk losing”¹.

In other words, conflict requires its actors to take a live situation, break down its terrain, and gamble on a response. The language is concerned less with “good” versus “bad”, and more with strengths versus weaknesses, “cans” versus “cannots”. Scenarios may come without clear precedent, and analyzing them incorrectly comes with real stakes. The answers one comes to could cease to be relevant ten years from now as the situation evolves. Conflict does not stop at criticism – an actor has to apply their analysis.

It’s not hard to find criticisms of the tech industry nowadays. Headlines are full of thinkpieces, ideological profiles, and industry horse-race coverage – all zooming in on the ideologies of the billionaires and companies behind the curtain². But a lot of this is unserious, more focused on sensationalism than anything practically applicable.

Even among the more sophisticated tech-critics, for all their skill in identifying the problem, they fail to provide solutions which extend past the liberal state. Cory Doctorow can dedicate multiple chapters of his book chronicling how much the emergence of the modern tech industry dovetailed with the suppression of organized labor³, but when pressed on how to fight back, he falls back on calls for regulation. His weapons of choice to “keep the market fair” are antitrust and the goodwill of bureaucrats⁴.

Lawerence Lessig can recognize the ways in which intellectual property is designed to promote rent-seeking and stifle creativity, but then funnel grassroots energy into “getting money out of politics” so he can then pass legislation.

The conflict these critics describe is not one between lawmakers and the market. They explicitly lay out how it is one between the agents of capital and working people. Within such a context, the state’s role isn’t to aid struggle, but to mediate it. They depend on economic growth for both legitimacy and tax revenue. Law enforcement is also a basic state objective⁵ – the same EU commission that passed the GDPR is actively pushing to gut it in the name of tapping into AI-driven growth⁶. Even where the state recognizes the industry as a threat, it pushes policies that protect its own interests – age verification, chat control, domain-blocking – over that of users and hackers.

The sorts of reforms progressive-minded activists most often propose are some of the hardest to get through, because they run up against some very deep-seated realities of the modern economy. Lessig wants to majorly slash copyright terms⁷, but modern economies are extremely dependent on IP revenue⁸. Doctorow’s emphasis on union power hinges on an incredibly rosy⁹ picture of the current state of organized labor¹⁰ and overlooks genuine grassroots advances (such as worker-to-worker organizing models) in favor of overstating the role of “non-corrupt judges” and “good union leaders”¹¹¹².

Any serious response to the current situation needs to reflect an awareness of the conflict in question. It must directly focus on the issue in contention (tech), emerge out of the actions of those with the most skin in the game, and not bet its entire house on external circumstances changing.

To this end, it’s worth returning back to the FOSS (free-and-open source software) to the find seeds of a radical break. It is at it’s core a culture of commons-based peer production¹³ which has, in a very practical manner, challenged long-held notions surrounding property and production. The people involved in forming the culture didn’t start this as some kind of socialist experiment, but accidentally discovered it through the process of their work and what made sense. They saw for themselves how these things were made and used and built up their ideological frameworks out of that.

Working in this fashion opened up the space to genuinely new horizons that weren’t anchored to abstract political principles. But the hazard for these communities came as they began to evolve into a movement – lacking a clear sense of its own political grounding, FOSS activists tended to take the views around them for granted. Richard Stallman believed that solving software was just a matter of appealing to civic virtue¹⁴, whereas ESR was more inclined to put his hopes in the “invisible hand of the free market”¹⁵.

But despite the real grassroots enthusiasm behind them, all of these movements got rolled over the course of the 2010s. They underestimated the radicality of what they were proposing and were ill-prepared for both a market and a state that were actively declaring war on them. To have meaningfully fought back, there was a need to actually understand how the industry functions and what it was up to.

How Silicon Valley Works – Or Doesn’t

Silicon Valley’s mode of production is capitalism – not “techno-feudalism”¹⁶ or whatever other term tech critics try to devise for the sake of novelty. The modern incarnation of the industry emerged as the load-bearing pillar to an economy reeling from the aftermath of the Great Recession¹⁷. Where the housing crisis left a lot of the mystique and growth surrounding Wall Street to dry up, the relatively insulated tech sector¹⁸ swooped in to pick up the pieces. Tech provided a safe offramp for all the talent, capital, and political goodwill that had been burned in the ‘08 crisis¹⁹.

Historically, when faced with moments of stagnation, capitalism would often recover its growth by tapping into the periphery²⁰ – underdeveloped areas which provide access to new markets, resources, and cheaper labor. As Kohei Saito points out²¹, there’s strong parallels between this process and the commercialization of the digital space.

As a result, the market came to have a rather ambivalent relationship to the communities that inhabited the internet. On one hand, business held a genuine contempt for the culture of free and open sharing within these communities.²² On the other hand, the rapid growth of these startups depended on the wealth of free resources, labor, and access provided by them. This is where enclosure and flattening tactics (such as embrace-extend-extinguish²³ and open-washing) were actively and aggressively pursued by the market to swiftly neuter any opposition.

Given enough time, this process was refined into a formula. Capitalism depends on continuous growth, an industry can’t sustain itself off of one miracle business. There constantly needs to be a new iPhone, a new Uber, a new Amazon to keep consumer demand rising at the rates demanded by industry. Tech investors need to refine strategies that can reliably keep churning out The Next Big Thing.

This is where venture capital comes in: investors pool their money into a fund managed by industry “experts” who employ aggressive investment strategies optimized for high-risk, high-reward bets. They achieve this by exploiting power laws within the market – a disproportionate share of growth across a market empirically tends to concentrate in only a small handful of firms²⁴.

Knowing this fact tells you a lot about where you should place your bets. Focus on sectors with the highest potential growth ceilings, such as startups and cutting-edge tech. Double on triple down on any startups which show promise of becoming that golden goose – cull the rest. Break up funding into “rounds”²⁵, so you can hold founders to match an exponential curve throughout the whole cycle.

Return on venture funds tend to follow a J-curve, so revenue is an unreliable metric for success – it’s difficult to distinguish between the beginning of a linear curve and that of an exponential one.²⁶. Instead, modern VCs tend to prioritize a firm’s potential for monopoly. This is because properly competitive markets are a race to a bottom – matching your competitors thins your margins. But companies which are capable of “escaping competition” can indefinitely squeeze their customers for all they’re worth²⁷.

To this end, it’s encouraged to run a startup “fat”²⁸ – running massively in the red for years in service of capturing market share down the line. Uber famously did this²⁹, burning through its $25 billion investment so that they could sell taxi services at a rate impossible for the rest of the market to compete with.

Once the carrot has been set, it becomes very easy to begin locking in customers to your network, even as you then cut costs and hike up prices to pull yourself back in the black. For “platforms”, there’s a double lock-in at play since you’re connecting buyers and sellers³⁰. Amazon holds vendors and buyers hostage to each other, YouTube creators and audiences, the list goes on. Soon you’re dealing with companies that aren’t just monopolies, but also monopsonies – they can control the flow of an entire market.

That is the reality of how these businesses work. The VC narrative instead prefers to stress how “innovative” and “one-of-a-kind” these products are to explain the success. The fact is, however, the desired lock-in effect cannot occur in a truly free market. Two of the very core pillars Peter Thiel cites as essential to a successful business – proprietary technology and network effects³¹ – are often fiercely enforced by intellectual property and anti-circumvention law.

His airport book (a heavily watered-down version of his Yale lectures) retains the core thesis that capitalism and competition are intrinsically opposed³². But the sleight of hand comes in implicitly accepting that the process of eliminating competition is the same thing as building organic value. One look at the tech sector can say otherwise – plenty of monopoly-building tools exist (acquisitions³³, loss-leading, outright illegal collusion in some cases³⁴) which speak more to a firm’s ability to freely deploy capital than anything about the product itself.

This is why it’s not helpful to spend too much time picking apart any sort of ideological theories to come out of this space. NRx cameral systems, great founder theory, network states, or even Andreesen’s “techno-optimism”: just like that airport book The Atlantic heaps praise on³⁵, all of it is a repackaging of the same set of tropes in Silicon Valley conventional wisdom. That an optimal society is one run by venture capitalists continuing to do the things venture capitalists already do³⁶.

That’s why, then, this textbook model of Investing 101 keeps getting dredged up over and over again when VCs try to propose solutions to any number of social issues from prediction markets³⁷ to science funding³⁸. Venture capital reflects not just the strategy for accumulating short-term power but also the end goal. The rent-seeking slaughterhouse that is the industry comes to define the countours of the internet which in turn is expected to be a model for society writ large.

For any of this to be meaningful, it’s actually highly relevant whether or not the basic assumptions behind venture capital’s model actually hold up to reality. It’s worth noting, then, how the industry’s model of success has only proven itself within the narrowly artificial constraints of a market. Being one of the “greats” of the early-mid 2010s meant benefitting from low interest rates, a frothy investment market³⁹, and a wealth of technological commons ready to be exploited.

TV-over-internet, taxis-over-internet, pawnshops-over-internet, books-over-internet: none of these ever represented a true “zero to one” in the larger picture. There may have been a good deal of ingenuity behind these designs, they may have “revolutionized” a product category, all of them involved taking newly commercializable infrastructure and figuring out how to apply it to a pre-existing industry.

This is engineering, not black magic. There may be a real level of skill in developing these products, but that’s a far cry from wielding an ability to regularly conjure generational leaps out of thin air. Yet, that’s exactly what the growth curves of the industry hinge on.

VC does best in bull markets – where investors have the surplus to take on higher risk in pursuit of higher rewards. While there are real analogies that can be made to peripheries, the internet is not a perfect replacement. It can only be tapped so for so long. Where growth (both technological and economic) inevitably begins to slow down, the sector is forced to start gambling on increasingly dubious bets (metaverse, cryptocurrency, now GenAI⁴⁰) with higher and higher stakes and less and less clear practical applications.

There’s a constant deferring of risk in favor of short term gains, as a larger and larger share of the money flowing through the investment market finds itself based on speculative rather than tangible value. Those involved in this market have a vested interest⁴¹ in ensuring the speculative value of their assets remain pumped up. Those very same social and intellectual networks VCs rely on to coordinate (both financially and ideologically) become viral incubators for runaway speculation. In addition to burning credibility, these hype-cycles surrounding the “newest big thing” create a fundamentally delusional culture, one where there’s strong disincentives to raining on the parade.

The culture in Silicon Valley circles has a terrible habit of ascribing eternal significance to phenomena derived from about twenty years of business history. This is doubly exacerbated by the fact that the highly isolated intellectual and mentorship networks which populate this space (see: MIRI, YCombinator, etc.) often serve as feeders into the professional ones. Because of this, there’s often not much to stop the terms of business from bleeding in with that of the rest of reality. “Founder” comes to refer to not just entrepreneurs, but a quality found in every great genius of history. Effective governance is analogized to business management, public opinion is equated to crowd effects⁴², and playing politics often means treating democratic institutions like a competitive market.

This inability of those who represent the industry to analyze their own situation objectively is a real weakness, especially when stuff stops working. The obstacles the Silicon Valley playbook accounts for are political and financial ones – not material ones. Whether a pitch will sell, whether a message will win an election – but very little given to the technology itself.

The business model of locking in and squeezing customers produces cyclical and recurrent crises as platforms begin to sabotage their own value and stability. Pushed far enough, network effects can begin to unravel – a platform’s lowered value can end up tanking the value of the network as a whole (junk items on Etsy, SEO spam across Google, etc). Suddenly, quitting becomes a lot easier.

The market tries to correct for this by creating new platforms and product categories to serve the same degraded function (see: ChatGPT replacing Google), but in the process opens itself up to new, harder-to-manage risks⁴³ as technological complexity balloons out of control. Monopolized networks only add fuel to the fire, as they provide less redundancy in times of failure. For example, Amazon’s mistaken push towards AI-driven development has led to multiple widespread outages for many of AWS’ customers, forcing full-on rollbacks.⁴⁴

The breakdown of network effects is amplified with social media platforms, which by design act as decentralized megaphones for propaganda⁴⁵. Where a platform does begin to enshittify, users across a network will notice and begin using the network’s own channels to vocally resist it. We’ve seen this pattern with the Reddit API blackouts⁴⁶ and the various user revolts since Twitter’s acquisition⁴⁷. These efforts aren’t always successful, but they are cyclical.

This leads into quite possibly the greatest irony of all: that the very passive habits instilled by tech companies onto their users creates a sense of learned helplessness which leads to moral panics whenever the industry’s breakdowns begin to ripple into the rest of society. The opaqueness these companies created surrounding tech bleeds into the political discourse as the state finds itself under increasing pressure to pass sweeping and reckless legislation. The industry is incapable of providing a satisfactory answer to the public, as they themselves don’t understand the crisis. In turn, it finds itself diverting attention and resources trying towards electoral politics – trying to capture the state for itself. Rather than admit the intractable causes of its own slowed growth, its agents prefer to scapegoat what they see as the face of the regulatory state – “decels”, “village”, “luddites”, the list of ephitets goes on.

What was originally was a set of issues confined to a narrow industry has suddenly grown into a massive political problem.

What to Do?

The great finding of the Marxist tradition was to identify how capitalism produces recurrent, destabilizing crises which ripple in highly indirect fashions⁴⁸. These crises are a real, intractable weakness that socialist movements throughout history have seized upon to bring about change.

Crisis itself doesn’t guarantee collapse, nor can it bring about change on its own. Systems have a remarkable capacity to limp along and prop themselves back up relying on the volunteer efforts of regular members whose survival is threatened by crisis⁴⁹.

But, what these crises do provide are moments of reflection, whereby we are forced to recognize on both a political and practical level that the systems we have do not work. What we see going on in the tech sector may seem pettier, but is no exception to this rule. The recurrent platform-level and industry-level crises plaguing the industry are a direct product of its core assumptions surrounding property and value.

In order to meaningfully take advantage of these situations, a movement has to be able to issue a challenge to the present system on both a political level and on a practical level. The root of the crisis has to be vocally explained (and a line held against the state/market’s proposed reforms), but proving the reality of an alternative requires real work already be done towards experimentation and practice.

In acute moments like these, the goal shouldn’t be to force a change of policy or government. Neither is it to simply stir animosity against the “oligarchs” or promote a broadly oppositional stance to industry. It needs to be to use the turmoil as a springboard to begin concretely transforming the habits and relations of those caught up in it.

If the crises facing tech are caused by top-heavy strategies which prioritize short-term growth, then the counter is developing systems for the long-term: bottom-up, stable, and low-overhead. If the industry’s inability to manage crisis stems from an alienation from the technology – then taking advantage of that starts by recentering the tech itself.

The real subversive potential of FOSS communities comes not in criticizing Big Tech, but in challenging it on these specific grounds. The premise starts from a recognition of the artificiality of the digital space’s boundaries – it takes (often inconsistently applied) intellectual property law and DRM schemes to draw lines around what is otherwise just ones and zeroes. By doing this, it avoids the trap of essentializing any unfreedom as a property of tech itself – the underlying infrastructure hasn’t changed. Copying is free, code can be rewritten, and networks can be routed around.

Hacker culture builds itself on poking holes through these barriers. DRM can be cracked, blocks can be circumvented, and devices can be homebrewed. Where software is developed, it need not meet the design There’s a complete rejection of the producer-consumer dichotomy which underpins the commodity-market that industry tries to assimilate tech into. Users are encouraged to fork and contribute to projects, networks are strengthened by self-hosting, and users are encouraged to learn along as they fix and set up their own systems. A new definition of ownership becomes pervasive: one centered around what we use and depend on rather than abstract legal or financial status.

Remix cultures are capable of producing innovation without needing to rent-seek an intellectual property. Commons-based peer-production models have led to the rapid development of complex systems such as Linux⁵⁰ without needing to depend on large up-front capital infusions or full-time wage labor.

And all of this comes together to preserve a tradition where the primary purpose of technology is direct use as opposed to marketability. Old systems and solutions do not need to be constantly rendered obselecent or “revamped”. Classic UNIX principles stressed simplicity, reusability, focusedness, and longevity in the designs it produced⁵¹. There’s a strong sense for dissecting everything that goes into a product, from the codebase to the production and maintenance process.

Having this perspective is vital because it’s able to make sense of the under-the-hood nuances of tech that falls squarely within SV’s blindspot. By considering each input and step of the production process in its own terms (rather than flattening it into dollar valuations), you’re able to more finely delinate between strategic categories⁵².

Thinking outside of the purely financialized language allows you to reframe a lot of the “lessons” SV conventional wisdom routinely passes down. Power laws don’t just apply to startup growth trajectory: a small percentage of servers hold the majority of online traffic, a small handful of developers contribute the majority of code. Network effects apply to the GPL too – applying it to several key pieces of early software virally bolstered its growth to this day. Even non-commercial projects greatly benefit from building up strong, formalized professional networks and talent pipelines.

Open-source theorists often run the risk of overstating how much the organic advantages of hacker culture are capable of bringing about change on their own. The big difference between the days of Linux’s early development and now is that there no longer is a singular “holy grail” of free-software. A fully-free operating system is a clear, technical goal that users across the space could unite behind. Now, so much of our computing is done across mobile phones and web-apps that the landscape has severely fragmented.

The FOSS space of the modern day is littered with a litany of local groups, projects, and teams all working towards roughly common goals with relatively little coordination or overlap. This ends up having real consequences: it is incredibly difficult for projects, conferences, and community groups to find support without courting corporate sponsors. The average community-driven project often lives or dies on the passion of a handful of isolated individuals. Even the best-supported projects, such as Wikipedia, face a shrinking contributor pool⁵³ as people become less and less tech-literate.

Effectively coordinating and training up labor remains the number one issue facing FOSS communities right now. Maintaining the overall health of the space is going to depend on organizations capable of consciously filling in the gaps – directing limited resources in a targeted fashion according to a long-term strategic plan.

In theory, one would expect historic advocacy groups such as the FSF or OSI to carry out this task, but the reality is they’re ill-equipped and internally resistant to change. Being structured as advocacy non-profits⁵⁴, they often burn resources on advertising and direct-conversion (both incredibly inefficient uses of funds) and run decently large financial overheads to even carry out basic administrative functions.

Instead, just as FOSS paved the way for a new model in terms of producing software, it’s also necessary to break from the nonprofit model of organizing. Financial constraints are a reality of the basic calculus – you need to find ways to design operations that are capable of running lean. The advantage of this approach is that it can force the space to evolve in a way that selects for long-term resillience over trying to match corporate spending blitzes.

Power laws can be applied in reverse – targeting a small handful of enthusiasts, activists, and developers can have outsized effects for a fraction of the cost. The nature of adoption cycles is that they occur in a concentric fashion – the recommendations of a small core of enthusiasts funnel outwards through a series of adjacent audiences in a concentric fashion⁵⁵. The best advertisements often don’t come from billboards or endorsements, but the regularity with which the competition screws up and alienates its own customers. This is most seen when you look at the perpetual waves of exodus to Linux and Fediverse – factors out of their control produced a name recognition money couldn’t buy.

The real hurdle to adoption actually comes when people try the suggested FOSS offerings for themselves and see whether or not there’s any real friction⁵⁶. If a project is unprepared for a new wave of interest, it can quickly get overwhelmed and waste a valuable window. This is why instead of trying to directly reach audiences, it’s a more efficient use of resources to build structures that can fill in the gaps for these projects: hosting infrastructure, targeted code-contributions⁵⁷, cultivating content networks, and refining onboarding processes to ease the transition for new users. These are all relatively low-cost tasks which an ad-hoc, small-scale volunteer network could solve with some creativity.

Taken alone, these are individual tactics – anyone can help out with these tasks at an individual or project-level, but the greatest effect comes from performing it in aggregate. Forming this into a wider strategy needs to start with rebuilding the movement’s social infrastructure. Compiling rolodexes, developing public forums whereby isolated activist silos can be brought into a common discussion⁵⁸, and developing a consistent pipeline for recruiting, educating, and training up people is key. The main expense here is social – rather than financial – capital.

You don’t need a fully-funded nonprofit for this, but you do need a group of some sorts to model out the link between theory and practice. A tightly-knit network⁵⁹ of local, in-person chapters would be capable of bridging this gap between theorists and activists. It’d be capable of providing a pipeline for a small set of passionate enthusiasts to “get involved” with supporting the space at large. Via this pipeline, you can begin to instill a set of habits and mentalities which encourage members to creatively recognize and adapt to the real constraints: start with education (explaining the anatomy and dynamics underpinning both the tech industry and FOSS communities), before moving into application (bridging and shining a spotlight on various niches within the FOSS space and what potential they may have), and finally activation (getting these groups involved with actually taking action to support FOSS infrastructure).

In the end, it all comes back to building strong habits, because those habits end up trickling down to the rest of the space over the long term. It wasn’t that long ago that kids reading Wikipedia fell down rabbitholes of editing it. The game jams hosted by projects such as Godot have resulted in a flourishing of its community and ecosystem – and a respect for free-software among game developers.

The same industry that has drilled the concepts of non-fungible tokens and large language models into the public consciousness will also try to tell you that people are too technically illiterate to accept an internet of open protocols. Don’t accept their terms, but figure them out for yourself. That’s the only way you can win.