The internet's decline isn't due to greed alone, but because tech companies no longer face consequences from four key forces: competitors (bought out), regulators (captured), workers (laid off and not unionized), and new market entry (blocked by IP law like DMCA Section 1201)

The internet's decline isn't due to greed alone, but because tech companies no longer face consequences from four key forces: competitors (bought out), regulators (captured), workers (laid off and not unionized), and new market entry (blocked by IP law like DMCA Section 1201)

Both Facebook and Amazon followed a similar pattern: attract users/sellers with good service, lock them in through network effects or dependency, then progressively extract more value while providing less—with Amazon's take rising from under 20% to over 50% and Facebook moving to only 7% of Instagram time spent on content from people you follow

Both Facebook and Amazon followed a similar pattern: attract users/sellers with good service, lock them in through network effects or dependency, then progressively extract more value while providing less—with Amazon's take rising from under 20% to over 50% and Facebook moving to only 7% of Instagram time spent on content from people you follow

Maximally efficient price discrimination and algorithmic optimization may be economically 'efficient' but creates an oppressive, dehumanizing society where everyone constantly pays the maximum they're willing to pay and workers are surveilled and paid based on their desperation

Maximally efficient price discrimination and algorithmic optimization may be economically 'efficient' but creates an oppressive, dehumanizing society where everyone constantly pays the maximum they're willing to pay and workers are surveilled and paid based on their desperation

Solutions require moving beyond neutral rules to making value judgments about what kind of society we want: banning toxic business models (like exploiting children or extreme surveillance), mandating interoperability so users can leave platforms while staying connected, treating essential platforms as utilities with non-discrimination duties, and passing comprehensive privacy laws with private right of action

Solutions require moving beyond neutral rules to making value judgments about what kind of society we want: banning toxic business models (like exploiting children or extreme surveillance), mandating interoperability so users can leave platforms while staying connected, treating essential platforms as utilities with non-discrimination duties, and passing comprehensive privacy laws with private right of action

The key question isn't whether we want competition, but what kind of competition—healthy competition means making great content and products, while toxic competition means manipulation, addiction, and exploitation. We need the courage to distinguish between these and ban the worst practices rather than assuming all innovation is good.

The key question isn't whether we want competition, but what kind of competition—healthy competition means making great content and products, while toxic competition means manipulation, addiction, and exploitation. We need the courage to distinguish between these and ban the worst practices rather than assuming all innovation is good.

We Didn’t Ask for This Internet | The Ezra Klein Show

Name: We Didn’t Ask for This Internet | The Ezra Klein Show
Uploaded: 2026-02-07T00:26:15.893133
Duration: 1 h 28 min 31 s
Description: This episode of The Ezra Klein Show features a conversation with Cory Doctorow and Tim Wu about how the internet has deteriorated from its early promise into a landscape dominated by extractive corporate platforms. They discuss their respective frameworks—"enshittification" and "extraction"—for unde

By The Ezra Klein Show

Watch on YouTube (1:28:31)

Overview

This episode of The Ezra Klein Show features a conversation with Cory Doctorow and Tim Wu about how the internet has deteriorated from its early promise into a landscape dominated by extractive corporate platforms. They discuss their respective frameworks—"enshittification" and "extraction"—for understanding how tech companies like Facebook, Amazon, and Google have moved from serving users well to maximizing profits at the expense of users, small businesses, and workers. The conversation explores solutions including antitrust enforcement, interoperability mandates, privacy regulation, and rethinking what kind of competition and economy we want.

Key Takeaways

The internet's decline isn't due to greed alone, but because tech companies no longer face consequences from four key forces: competitors (bought out), regulators (captured), workers (laid off and not unionized), and new market entry (blocked by IP law like DMCA Section 1201)
Both Facebook and Amazon followed a similar pattern: attract users/sellers with good service, lock them in through network effects or dependency, then progressively extract more value while providing less—with Amazon's take rising from under 20% to over 50% and Facebook moving to only 7% of Instagram time spent on content from people you follow
Maximally efficient price discrimination and algorithmic optimization may be economically 'efficient' but creates an oppressive, dehumanizing society where everyone constantly pays the maximum they're willing to pay and workers are surveilled and paid based on their desperation
Solutions require moving beyond neutral rules to making value judgments about what kind of society we want: banning toxic business models (like exploiting children or extreme surveillance), mandating interoperability so users can leave platforms while staying connected, treating essential platforms as utilities with non-discrimination duties, and passing comprehensive privacy laws with private right of action
The key question isn't whether we want competition, but what kind of competition—healthy competition means making great content and products, while toxic competition means manipulation, addiction, and exploitation. We need the courage to distinguish between these and ban the worst practices rather than assuming all innovation is good.

Foundation Concepts

Market competition dynamics (econ): In healthy markets, multiple firms compete for customers by offering better products or lower prices. Competition disciplines firms because customers can switch to rivals if one company treats them poorly. When competition disappears through consolidation, the remaining firms gain pricing power and can degrade service without losing customers.
Regulatory capture (econ): Regulatory capture occurs when industries gain influence over the government agencies meant to oversee them, often through lobbying, revolving-door employment, or campaign contributions. Captured regulators tend to make rules that benefit the industry rather than the public, effectively neutering oversight.
Labor union bargaining power (econ): Unions give workers collective power to negotiate wages, conditions, and company practices by threatening coordinated action like strikes. Without unions, individual workers have little leverage against large employers, especially when jobs are scarce. Tech workers historically avoided unionization, leaving them vulnerable when their scarcity value disappeared.
Intellectual property law (law): IP laws like copyright, patents, and anti-circumvention rules give creators temporary monopolies over their inventions or works. The Digital Millennium Copyright Act Section 1201 makes it illegal to bypass technological protection measures, even for legal purposes. This prevents competitors from creating compatible products or modifications without permission.
Network effects and lock-in (econ): Network effects occur when a product becomes more valuable as more people use it, like social networks where you go where your friends are. This creates lock-in where users can't easily leave even if service degrades, because the value is in the network itself. Companies exploit this by degrading service once users are trapped.
Two-sided platform markets (econ): Platforms like Amazon and Facebook connect two groups: users and sellers (or users and advertisers). The platform's power comes from controlling access between these groups. Once both sides depend on the platform, it can extract value from both by charging sellers more or degrading the user experience.
Switching costs (econ): Switching costs are the time, money, or effort required to change from one product or service to another. High switching costs trap customers even when they're unhappy. For social media, switching costs include losing your social connections; for sellers, it's losing access to customers and having to rebuild reputation.
Algorithmic content curation (tech): Algorithms decide what content users see based on predicted engagement, profitability, or other goals set by the platform. This replaced chronological feeds where you saw what you asked to see. Algorithmic feeds let platforms insert ads, promoted content, and engagement-maximizing material while reducing organic content from people you follow.
Marketplace commission structures (econ): Online marketplaces charge sellers fees for access to customers, typically as a percentage of sales plus additional charges for services like fulfillment, advertising, and premium placement. As platforms gain power, they can increase these fees because sellers have nowhere else to reach the same customer base.
Price discrimination (econ): Price discrimination means charging different customers different prices for the same product based on their willingness to pay. Airlines do this with dynamic pricing; luxury brands do it with regional pricing. Perfect price discrimination extracts maximum value from each customer, leaving them no consumer surplus.
Consumer surplus (econ): Consumer surplus is the difference between what you're willing to pay and what you actually pay—essentially the 'deal' you get. If you'd pay $10 for something but only pay $6, you have $4 of surplus. Perfect price discrimination eliminates this surplus by charging everyone their maximum willingness to pay.
Algorithmic surveillance (tech): Modern platforms collect vast data about users' behavior, preferences, financial status, and vulnerabilities through tracking across websites, apps, and devices. Algorithms analyze this data to predict behavior and optimize outcomes for the platform. This enables personalized pricing, targeted manipulation, and discrimination at scale.
Wage discrimination (econ): Wage discrimination occurs when workers doing the same job are paid differently based on factors like their desperation, debt level, or perceived alternatives. Historically this was hard to implement, but algorithmic systems can now assess each worker's financial vulnerability and offer personalized wages that extract maximum labor for minimum pay.
Economic efficiency vs welfare (econ): Economic efficiency means maximizing total output or value in an economy, often measured by GDP. However, efficiency doesn't account for distribution, dignity, or quality of life. A maximally efficient economy where firms extract all surplus may be technically optimal but create widespread misery and inequality.
Utility regulation (econ): Utilities are essential services like electricity, water, or telecommunications that are treated as natural monopolies and heavily regulated. Utility regulation typically includes price controls, non-discrimination requirements (serving all customers equally), and quality standards. The trade-off is accepting monopoly in exchange for public-interest obligations.
Interoperability mandates (tech): Interoperability means different systems can work together and exchange information. Mandating interoperability forces platforms to let users communicate across services, like how you can call any phone from any carrier. This reduces lock-in by letting users switch platforms while maintaining connections, similar to email working across providers.
Private right of action (law): A private right of action allows individuals to sue to enforce a law, rather than relying solely on government agencies. This creates an army of potential enforcers and gives victims direct recourse. Without it, laws depend entirely on under-resourced agencies that may be captured or simply lack bandwidth.
Non-discrimination duties (law): Non-discrimination duties require that essential service providers treat all customers equally and fairly, without favoring some over others or their own services. Common carriers like phone companies have historically had these obligations. Applying this to platforms would prevent them from favoring their own products or extracting rents from business customers.
Value-neutral vs normative policy (philosophy): Value-neutral policy claims to avoid making judgments about what's good or bad, instead relying on procedural rules or market outcomes. Normative policy explicitly makes value judgments about desired social outcomes. The pretense of neutrality often masks implicit values while preventing democratic deliberation about what kind of society we want.
Competition as discipline (econ): Competition is valued in economics because it disciplines firms—they must serve customers well or lose them to rivals. However, this assumes competition happens on socially beneficial dimensions like quality and price. When firms compete on harmful dimensions like manipulation or addiction, competition makes things worse, not better.
Attention economy (econ): The attention economy treats human attention as a scarce resource to be captured and monetized. Companies compete to hold attention longer through engaging content, but also through psychological manipulation, infinite scroll, and algorithmic optimization for engagement over wellbeing. This creates a race to the bottom in exploiting human psychology.
Addictive design patterns (psych): Addictive design uses psychological principles like variable rewards, social validation, and fear of missing out to create compulsive usage. Features like infinite scroll, autoplay, and notification systems are engineered to maximize time spent rather than user satisfaction. These patterns exploit the same brain mechanisms as gambling.
Innovation bias (philosophy): Innovation bias is the assumption that new technologies and business models are inherently good and should be allowed to flourish without restriction. This ideology treats disruption as automatically beneficial and regulation as harmful, preventing democratic deliberation about whether specific innovations serve human flourishing or cause harm.
Platform business models (econ): Platforms are businesses that create value by facilitating exchanges between different user groups, like connecting buyers and sellers or content creators and audiences. Unlike traditional businesses that sell products, platforms sell access and charge rents for intermediation. Their power grows with the size of their network.
Multi-stage exploitation (econ): Multi-stage exploitation describes how platforms sequentially extract value from different stakeholders. First they subsidize users to build the network, then they exploit users to attract business customers, then they exploit business customers once they're locked in. Each stage depends on the previous one creating dependency.
Collective action problems (econ): Collective action problems occur when a group would benefit from coordinating but individuals can't organize effectively. Leaving a social network requires convincing your whole friend group to move together, which is nearly impossible. This coordination failure keeps people trapped even when everyone is unhappy.
Antitrust enforcement history (law): Antitrust laws prohibit monopolies and anti-competitive practices, but enforcement varies dramatically by era. From the 1980s through 2010s, US antitrust focused narrowly on consumer prices, allowing massive consolidation. This permitted tech companies to acquire potential competitors freely, eliminating competitive discipline.
Market power (econ): Market power is the ability of a firm to raise prices above competitive levels without losing all its customers. It comes from barriers to entry, network effects, or controlling essential infrastructure. Firms with market power can extract rents—profits beyond what competitive markets would allow.
Rent-seeking behavior (econ): Rent-seeking means pursuing profit by manipulating the economic or political environment rather than creating new value. This includes lobbying for favorable regulations, erecting barriers to competition, or exploiting monopoly positions. Rent-seeking redistributes wealth without growing the overall economy.
Productive vs extractive growth (econ): Productive growth comes from innovation, efficiency improvements, or better serving customers—creating new value. Extractive growth comes from capturing more of existing value through market power, price increases, or degrading quality while maintaining prices. Extractive growth enriches firms while making society poorer overall.
Monopoly pricing theory (econ): Monopolies maximize profit by restricting output and raising prices above competitive levels. This creates deadweight loss—transactions that would benefit both parties but don't happen because of high prices. Monopoly pricing is economically inefficient and transfers wealth from consumers to the monopolist.
Surveillance capitalism (econ): Surveillance capitalism is a business model based on collecting personal data, analyzing it to predict behavior, and selling that predictive power to advertisers or using it to manipulate users. It treats human experience as free raw material for commercial extraction. Facebook pioneered this model at massive scale.
Engagement optimization (tech): Engagement optimization means designing systems to maximize time spent, clicks, shares, or other interaction metrics. Algorithms learn what content keeps users scrolling and prioritize it, often favoring outrage, conflict, and emotional manipulation over quality or user satisfaction. Engagement correlates with ad revenue, not wellbeing.
Chronological vs algorithmic feeds (tech): Chronological feeds show content in time order from sources you chose to follow. Algorithmic feeds use machine learning to select and order content based on predicted engagement or other platform goals. The shift to algorithmic feeds gave platforms control over what users see, enabling monetization through promoted content and ads.
Social network lock-in (tech): Social networks create powerful lock-in because their value comes from your connections to other people. You can't take your social graph with you when you leave. This means even if you hate the platform, leaving means losing touch with friends, family, and communities you care about.
Marketplace platform dynamics (econ): Marketplace platforms connect independent sellers with buyers, providing infrastructure like payment processing, search, and fulfillment. The platform's power comes from controlling access to customers. As the marketplace grows, sellers become dependent on it, allowing the platform to increase fees and extract more value.
Search result manipulation (tech): Search result manipulation means ordering results based on factors other than relevance to the user's query, such as which seller pays the most in fees or advertising. This degrades search quality while extracting money from sellers, who must pay to be visible. Users get worse results while sellers' margins shrink.
Fulfillment by Amazon model (econ): Fulfillment by Amazon (FBA) lets sellers store inventory in Amazon warehouses and use Amazon's shipping network. This provides convenience but creates dependency—sellers must pay substantial fees and follow Amazon's rules. FBA also gives Amazon data on what products sell well, enabling them to compete with their own sellers.
Sponsored product advertising (econ): Sponsored products are paid placements in search results that look like organic results. Sellers bid for placement, with higher bids getting better positioning. This creates a pay-to-play system where visibility depends on advertising spend rather than product quality, degrading search while generating massive revenue for the platform.
Horizontal mergers (law): Horizontal mergers combine direct competitors in the same market, like Amazon buying Diapers.com or Facebook buying Instagram. These mergers reduce competition and can create or strengthen monopolies. Antitrust law traditionally scrutinizes horizontal mergers closely, but enforcement weakened dramatically from the 1980s onward.
Potential competition doctrine (law): Potential competition doctrine says antitrust should block mergers that eliminate future competitors, even if they don't compete directly today. A startup might grow to challenge the incumbent, so allowing acquisition eliminates that possibility. This doctrine was largely abandoned but is being revived in tech antitrust.
Predatory acquisition (econ): Predatory acquisition means buying competitors not to integrate their products but to eliminate competitive threats. Amazon's purchase of Diapers.com is an example—they bought a potential rival and shut down its independent operation. This is a form of anti-competitive conduct that maintains monopoly power.
Market concentration (econ): Market concentration measures how much of a market is controlled by the largest firms. High concentration means a few companies dominate, reducing competition. Tech markets have become extremely concentrated, with one or two firms controlling most activity in search, social media, e-commerce, and mobile operating systems.
Monopsony power (econ): Monopsony is monopoly power on the buying side—when one employer dominates a labor market. Monopsony employers can pay below-market wages because workers have few alternatives. Platform companies often create monopsony conditions by dominating gig work or using non-compete agreements and no-poaching deals.
Workplace surveillance technology (tech): Modern surveillance technology monitors workers through cameras, keystroke logging, GPS tracking, productivity metrics, and biometric sensors. This enables constant monitoring and algorithmic management, where software makes decisions about scheduling, pay, and discipline. Surveillance intensifies work and eliminates autonomy and dignity.
Algorithmic management (tech): Algorithmic management uses software to direct, evaluate, and discipline workers, often with minimal human oversight. Algorithms assign tasks, monitor performance, adjust pay, and even fire workers automatically. This creates opacity and removes human judgment, while optimizing for firm metrics rather than worker wellbeing.
Shitty technology adoption curve (sociology): The shitty technology adoption curve describes how harmful technologies are first imposed on the most vulnerable populations (prisoners, asylum seekers, low-wage workers) to work out the kinks, then gradually move up to affect more privileged groups. This pattern shows how exploitation of the powerless normalizes practices that eventually affect everyone.
Dynamic pricing algorithms (tech): Dynamic pricing uses algorithms to adjust prices in real-time based on demand, inventory, competitor prices, and individual customer characteristics. While this can reflect genuine supply and demand, it also enables personalized price discrimination where each customer is charged based on their predicted willingness to pay.
First-degree price discrimination (econ): First-degree or perfect price discrimination means charging each customer exactly their maximum willingness to pay. This extracts all consumer surplus, leaving customers with no benefit from the transaction beyond the product itself. Historically impossible to implement, but big data and algorithms make it increasingly feasible.
Data-driven profiling (tech): Data-driven profiling uses collected information about individuals—browsing history, location, purchase patterns, demographics—to predict their preferences, financial status, and price sensitivity. This enables personalized pricing, targeted advertising, and discrimination. The more data collected, the more accurate and exploitative the profiling becomes.
Revealed preference theory (econ): Revealed preference theory claims that people's true preferences are shown by their choices, not their stated opinions. Economists use this to argue that if someone pays a high price, they must value it that much. Critics note this ignores power imbalances, limited options, and the difference between accepting a bad deal and preferring it.
Data broker industry (tech): Data brokers collect, aggregate, and sell personal information about individuals without direct relationships with them. They compile data from public records, online tracking, purchases, and other sources to create detailed profiles. This information is sold to advertisers, employers, insurers, and others for targeting and discrimination.
Privacy as infrastructure (philosophy): Privacy isn't just about individual preferences but serves as social infrastructure that prevents exploitation, discrimination, and manipulation. Without privacy, every interaction becomes an opportunity for extraction. Privacy laws create boundaries that enable trust, autonomy, and dignity in economic and social life.
GDPR implementation (law): The General Data Protection Regulation is the EU's comprehensive privacy law requiring consent for data collection, rights to access and deletion, and restrictions on data use. However, enforcement depends on national regulators, and Ireland (where US tech companies are headquartered for tax reasons) has done minimal enforcement, leading to cookie popups instead of real protection.
Consent theater (tech): Consent theater refers to privacy mechanisms that appear to give users control but are designed to be confusing, burdensome, or manipulative, ensuring most people click through without reading. Cookie consent popups are the prime example—legally required but practically useless, creating compliance without protection.
Reverse engineering (tech): Reverse engineering means analyzing a product to understand how it works, often to create compatible products, modifications, or improvements. Historically legal and essential for competition and innovation, it allows new entrants to build on existing technology. Software reverse engineering is now largely illegal under anti-circumvention laws.
DMCA Section 1201 (law): Section 1201 of the Digital Millennium Copyright Act makes it illegal to circumvent technological protection measures, even for legal purposes. This means you can't modify software or devices you own if doing so bypasses any access control. It's been used to prevent repair, modification, and competitive products.
API and protocol compatibility (tech): APIs (Application Programming Interfaces) and protocols are the technical standards that let different software systems communicate. Interoperability requires either open standards or the ability to reverse-engineer proprietary ones. When companies can legally block compatible products, they can prevent users from leaving while maintaining connections.
Right to repair movement (law): The right to repair movement advocates for laws requiring manufacturers to provide parts, tools, and information needed to repair products. Manufacturers use software locks, proprietary parts, and legal threats to prevent independent repair, forcing consumers to use expensive authorized service or buy new products. This extends to software modification.
Common carrier obligations (law): Common carriers are businesses that offer services to the general public and must serve all customers equally without discrimination. Historically applied to railroads, phone companies, and utilities. Common carrier status prevents companies from favoring their own services or extracting rents by discriminating among customers.
Structural separation (law): Structural separation means prohibiting companies from operating in multiple layers of a market where they could favor themselves. For example, preventing Amazon from both running a marketplace and selling products on it, or requiring Google to separate search from content. This prevents self-preferencing and conflicts of interest.
Continuous antitrust enforcement (law): Rather than one-time breakups, continuous enforcement means ongoing scrutiny and intervention to prevent anti-competitive behavior. This includes blocking acquisitions, preventing exclusionary contracts, and maintaining competitive pressure. The goal is keeping dominant firms insecure in their position so they must keep serving customers well.
Precautionary principle (philosophy): The precautionary principle says when an activity raises threats of harm, precautionary measures should be taken even if cause-and-effect isn't fully established. Applied to tech, this means not waiting for definitive proof of harm before regulating potentially dangerous practices like exploiting children or extreme surveillance.
Constrained choice (philosophy): Constrained choice recognizes that decisions made under limited options, coercion, or desperation don't reflect genuine preferences. Selling a kidney to make rent reveals desperation, not a preference for having one kidney. Similarly, accepting surveillance to use social media reveals lack of alternatives, not approval of surveillance.
Opt-in vs opt-out defaults (psych): Default settings dramatically affect behavior—most people stick with whatever comes pre-selected. Opt-out systems (where you must act to refuse) get much higher participation than opt-in (where you must act to accept). This shows that behavior often reflects friction and defaults rather than true preferences.
Preference falsification (sociology): Preference falsification occurs when people's public choices differ from their private preferences due to social pressure, power dynamics, or lack of safe alternatives. What looks like revealed preference may actually be adaptive behavior under constraint. True preferences only emerge when people have genuine freedom to choose.
Economic efficiency definition (econ): Economic efficiency means maximizing output from given inputs, or achieving outcomes at lowest cost. Pareto efficiency means no one can be made better off without making someone worse off. However, efficiency is value-neutral about distribution and doesn't account for dignity, autonomy, or quality of life.
Optimization for whom (philosophy): Optimization always serves particular goals and interests. What's optimal for a firm (maximum extraction) differs from what's optimal for workers (good wages and conditions) or society (flourishing and dignity). Treating firm-level optimization as universal optimization obscures whose interests are being served.
Instrumental vs intrinsic value (philosophy): Instrumental value means something is valuable as a means to an end, like money being useful for buying things. Intrinsic value means something is valuable in itself, like human dignity or meaningful relationships. Efficiency thinking treats everything as instrumental, but some things have intrinsic value that shouldn't be optimized away.
Gardener metaphor for regulation (philosophy): The gardener metaphor frames economic regulation as cultivation rather than hands-off neutrality. Gardeners make value judgments about which plants to encourage, which to prune, and what kind of garden they want. Similarly, regulators should actively shape markets toward desired social outcomes rather than pretending neutrality.
Attention as scarce resource (econ): Human attention is fundamentally limited—we have only so many waking hours and can focus on only one thing at a time. The attention economy treats this scarcity as an opportunity for extraction, with companies competing to capture and monetize attention. Unlike material resources, attention can't be expanded or stored.
Persuasive technology design (psych): Persuasive technology uses psychological principles to change user behavior, often to increase engagement or purchases. Techniques include variable rewards (like slot machines), social proof, scarcity cues, and commitment devices. Originally studied for positive applications like health behavior, it's now primarily used for commercial manipulation.
Commodification of experience (sociology): Commodification means turning something that wasn't previously bought and sold into a market commodity. The attention economy commodifies human experience itself—every moment becomes an opportunity for commercial extraction. This transforms life from something lived into something mined for value.
Positive vs negative competition (econ): Positive competition improves products or creates value—competing to make better movies or more useful tools. Negative competition exploits vulnerabilities or harms users—competing to be more addictive or manipulative. Markets don't automatically favor positive competition; without regulation, negative competition often wins because it's more profitable.
Blogging era culture (history): The blogging era (roughly 2000-2010) featured decentralized publishing where individuals controlled their own websites and audiences. Blogrolls and RSS feeds let readers follow writers directly without algorithmic intermediaries. This created diverse voices and communities without platform control, though with limited reach compared to social media.
Chronological social media (tech): Early social media platforms like Twitter and Facebook showed posts in reverse chronological order from people you chose to follow. This gave users control over their information diet and made the platform a neutral tool rather than an active curator. The shift to algorithmic feeds transferred control from users to platforms.
Technological determinism (philosophy): Technological determinism is the belief that technology develops according to its own logic and determines social outcomes. The counter-view is that technology is shaped by human choices, power structures, and policy. Disputing nostalgia means arguing that current outcomes weren't inevitable but resulted from specific decisions that could have gone differently.
Decentralized web architecture (tech): The early web was architecturally decentralized—anyone could create a website, and open protocols like HTTP and email connected them without central control. This contrasts with today's platform-centric internet where a few companies control most activity. Decentralization enabled innovation and user control but lacked the convenience of platforms.
Infrastructure as public good (econ): Infrastructure like roads, utilities, and communications networks are foundational systems that enable other economic and social activity. They're often natural monopolies and have public good characteristics—non-rivalrous and hard to exclude people from. This justifies public provision or heavy regulation to ensure broad access and fair terms.
Essential facilities doctrine (law): Essential facilities doctrine says that if a facility is essential for competition and can't be reasonably duplicated, its owner must provide access to competitors on reasonable terms. Applied to platforms, this would mean companies that control essential infrastructure for commerce or communication must allow fair access rather than extracting monopoly rents.
Civic infrastructure (sociology): Civic infrastructure includes the physical and institutional spaces where public life happens—forums, town squares, libraries, parks. These shape what kind of community is possible. Digital platforms now serve similar functions, and their design and governance determine what kind of public sphere and democracy we have.
Network effects and natural monopoly (econ): Natural monopolies arise when one provider can serve a market more efficiently than multiple competitors, often due to high fixed costs and low marginal costs. Network effects create natural monopoly tendencies in platforms—one large network is more valuable than several small ones. This justifies treating them as utilities rather than competitive markets.
Profit margin economics (econ): Profit margin is the percentage of revenue remaining after costs. Small businesses typically operate on thin margins—often 5-10% net profit. When platform fees rise from 20% to 50% of revenue, they can eliminate profitability entirely. This forces businesses to raise prices, cut quality, or exit the market.
Entrepreneurship barriers (econ): Entrepreneurship barriers are obstacles that prevent people from starting businesses, including capital requirements, regulatory complexity, and access to customers. While platforms initially lowered some barriers by providing infrastructure, rising fees and platform power now create new barriers that favor large sellers over independent entrepreneurs.
Economic concentration effects (econ): Economic concentration means wealth and market power accumulating in fewer hands. When platforms extract most of the value from transactions, wealth flows from many small businesses to a few platform owners. This increases inequality, reduces economic dynamism, and concentrates political power.
Middleman extraction (econ): Middlemen facilitate transactions between producers and consumers, taking a cut for their services. When middlemen have market power, they can extract disproportionate value relative to the service provided. Platform fees represent middleman extraction—Amazon takes 50% while sellers do the work of sourcing, storing, and shipping products.
Organizational bureaucracy (sociology): As organizations grow, they develop bureaucratic structures with multiple layers of management, formal procedures, and risk aversion. This makes them slower to innovate and less able to pursue radical ideas. Large companies often can't compete with nimble startups on innovation, so they buy them instead.
Innovation through acquisition (econ): Rather than developing new products internally, large companies increasingly innovate by acquiring startups that have already proven a concept. This can be efficient but becomes problematic when it's the primary innovation strategy, as it concentrates power and eliminates potential competitors before they can challenge the incumbent.
Internal product development failure (tech): Large tech companies often struggle to launch successful new products despite enormous resources. This stems from bureaucracy, internal politics, lack of urgency, and misaligned incentives. Google famously kills most internal projects while its successful products (YouTube, Android, Maps) were acquisitions.
Brandeisian antitrust (law): Louis Brandeis argued that bigness itself is a problem beyond just economic efficiency—large corporations gain political power, stifle innovation, and threaten democracy. The 'curse of bigness' refers to these broader harms. Modern Brandeisian antitrust revives this view, arguing size limits are necessary even when prices stay low.
Talent acquisition strategies (econ): Companies compete for skilled workers through recruiting, compensation, and workplace culture. Acquihiring is buying an entire startup primarily to hire its team, often paying millions per employee. This is cheaper than competing for talent individually and signals that the company can't attract top talent through normal hiring.
Venture capital incentives (econ): Venture capital invests in startups expecting most to fail but a few to return enormous profits. When acquihires become common exits, VC incentives shift from building sustainable businesses to creating attractive acquisition targets. This distorts entrepreneurship toward impressing potential acquirers rather than serving customers.
Product shutdown patterns (tech): When acquirers shut down acquired products, it signals the acquisition wasn't about the product's value to users but about eliminating competition or acquiring talent. This destroys value for users and the broader ecosystem. Frequent product shutdowns indicate a dysfunctional innovation ecosystem.
Startup ecosystem health (econ): A healthy startup ecosystem produces companies that grow into independent businesses, creating jobs, innovation, and competition. When most startups aim for acquisition rather than independence, the ecosystem becomes extractive—feeding talent and ideas to incumbents rather than creating new centers of power and innovation.
Regulatory capture mechanisms (econ): Regulatory capture occurs through revolving doors (regulators moving to industry jobs), lobbying, campaign contributions, and information asymmetry (industry knows more than regulators). Captured agencies make rules that benefit industry over the public. Capture is easier when industries are concentrated because fewer players need to coordinate.
Lobbying and political influence (law): Lobbying means attempting to influence government decisions, typically through direct advocacy, campaign contributions, and providing information to lawmakers. Concentrated industries can spend heavily on lobbying because they have monopoly profits to deploy. This translates economic power into political power, blocking regulation.
Congressional dysfunction (law): Congress increasingly fails to pass legislation even on popular issues due to polarization, filibuster rules, committee gatekeeping, and dependence on campaign contributions. Tech companies exploit this by preventing votes on privacy and child protection laws despite overwhelming public support, using lobbying to maintain the status quo.
Cartel coordination (econ): Cartels are groups of companies that coordinate to reduce competition, typically on prices or output. Consolidated industries function like cartels—few players can easily agree on lobbying priorities and regulatory preferences. This makes regulatory capture easier because the industry speaks with one voice and has resources to deploy.
European federalism structure (law): The EU is a federation where member states retain sovereignty but delegate certain powers to EU institutions. EU laws must be enforced by national regulators, creating variation in implementation. This structure allows individual countries to undermine EU-wide policies, as Ireland does with privacy enforcement to protect its tax haven status.
Tax haven dynamics (econ): Tax havens are jurisdictions with low tax rates that attract companies to establish nominal headquarters there to avoid taxes elsewhere. Ireland offers low corporate taxes, attracting US tech companies. This creates incentives for Ireland to avoid strict regulation that might drive companies away, undermining enforcement of EU laws.
Lead regulator principle (law): Under GDPR, the regulator in the country where a company has its EU headquarters takes the lead on enforcement. Since US tech companies are headquartered in Ireland for tax reasons, the Irish Data Protection Commissioner leads enforcement. Ireland's weak enforcement effectively nullifies GDPR for major platforms.
Compliance vs protection gap (law): Compliance means following the letter of regulations, while protection means achieving the law's intended purpose. Cookie consent popups provide technical compliance with GDPR but don't protect privacy—they're designed to be annoying and confusing so users click through. This gap between compliance and protection indicates regulatory failure.