By default, capital will matter more than ever after AGI by L Rudolf L. Paper Summary

Introduction

In the Substack post titled “By default, capital will matter more than ever after AGI,” author L Rudolf L argues that advanced, labour-replacing artificial intelligence (AI) is poised to dramatically reorder the global economic and social landscape in a way that entrenches the power of existing capital holders. Far from making money obsolete, AI’s continued development will amplify the role of capital—both financial resources and the physical capital that money can purchase, such as factories or data centres—while correspondingly diminishing the role of human labour in most sectors.

The post warns that widespread faith in a utopian “post-AGI” era—where labour becomes irrelevant, and money ceases to matter—misconstrues how AI will transform the incentives of state and corporate actors. Instead of producing perfect social equality or a frictionless abundance that automatically benefits everyone, labour-replacing AI may undercut the very reasons that governments and institutions have historically needed to court human labour. In this new dynamic, individuals who do not already possess capital—people who rely on their work for economic or political leverage—risk losing both bargaining power and opportunities for “outlier success.”

In summarizing the article’s core arguments and detailed reasoning, this overview explores (1) the shifting relationship between labour and capital, (2) the author’s assessment of why governments and corporations currently “care” about human workers, (3) the ways in which AI disrupts that care, (4) the implications for social and political mobility, and (5) why, absent massive redistribution, inequality may become locked in. Ultimately, the author calls on readers to recognize this precarious moment in history: while AI presents a final, extraordinary window for human ambition, it also threatens a static future where wealth and power become irreversible.

---

Defining Labour, Capital, and the Economic Impact of AI

Labour, according to the post, is the human mental and physical effort that produces valuable outputs. Capital refers to the stock of goods (e.g., factories, software, data centres) and the monetary resources that pay for such goods. While “capital” can mean both tangible capital assets and the money used to purchase them, the author clarifies that if they want to exclude capital goods in their discussion, they will specifically say “money.”

Today, humans provide crucial mental and physical labour in countless industries, giving their work inherent economic value. However, as AI becomes better at tasks historically reserved for human minds, it starts acting as a near-complete substitute for labour. The more capable AI systems become at complex tasks—software engineering, data analysis, advanced research, creative design, and eventually physical labour through robotics—the less humans need to be paid for their time. In other words, capital (AI systems + infrastructure) can replace labour.

The Key Consequences

From this dynamic, the author extrapolates a few major conclusions:

1. The ability to buy results skyrockets. Because AIs can do tasks once exclusive to specialized human experts, any organization or individual with sufficient capital (money + computational resources) can replicate or clone those AI capabilities. The barriers of human talent—finding it, training it, and compensating it—diminish, since advanced AIs are easily multiplied and rarely demand more than electricity and hardware.
   
2. Human leverage falls. Most ordinary people today rely on their labour for economic influence. Even specialists and gifted individuals can command outsized salaries or positions because their talent is scarce. With labour-replacing AI, that scarcity vanishes. Cloning a superhuman AI engineer is cheaper and faster than hiring a top human engineer—even if the latter is theoretically “available.”
   
3. Institutions have less incentive to care about humans. States, companies, and other large institutions often grant concessions to workers (or general populations) because, historically, they needed those workers as sources of economic productivity or political stability. But if AI replaces the productivity aspect, then governments or corporations don’t depend on a skilled human workforce to generate value.
   
4. Outlier success becomes more difficult. Currently, an ambitious outsider might become a wildly successful entrepreneur, scientist, or intellectual, catapulting themselves into wealth or influence. But the post-AGI world makes it easier for incumbent capital to corner the market on new innovations by deploying armies of AI clones. The route from “nobody” to “somebody” through talent, creativity, or entrepreneurial skill may vanish.
   
5. Radical egalitarian revolutions are unlikely. Historically, societies have leveled wealth disparities only through catastrophic events like wars, revolutions, or pandemics. The emergence of AI doesn’t inherently produce any impetus for massive redistribution. Rather, it could further entrench existing capital holders, as they harness the new technology without needing to share the benefits widely.

---

Why Money Is Unlikely to Disappear

Many people assume that once AGI arrives, money will no longer matter because the abundance generated by superintelligent AI will render human needs trivial. The author challenges this assumption directly, pointing out two critical points:

1. Prices continue to serve a function even in extremely sophisticated economies. Unless there is a single global AI system that centrally allocates all resources, multiple actors will still need to exchange some form of value. Money persists as the standard unit for those exchanges.
   
2. Abundance is not infinite. Even under rapid technological progress, resources remain finite at any given time. These finite resources still need some mechanism of allocation. Thus, money, markets, or something functionally similar endures.

Furthermore, the author notes that while we might see Universal Basic Income (UBI) introduced in many countries once large segments of the population are unemployed, UBI does not eliminate the function or power of capital. Indeed, capital becomes more decisive in shaping how AI-based productivity is deployed, and who reaps the benefits.

---

Capital’s New Power: Hiring AI “Talent”

One of the article’s central arguments is that money currently struggles to buy the very best human talent:

• Top engineers or researchers often can’t be identified by pure financial metrics, because recognizing genuine expertise requires deep domain knowledge.
  
• Truly brilliant individuals may turn down high-paying roles if they conflict with their ideals or passions.
  
• Startups can beat out bigger, richer companies by having a compelling vision, a strong mission, or a unique culture that attracts talent beyond salary.

Enter labour-replacing AI. Once AI can match or exceed human specialists in every major domain, the friction in “talent acquisition” disappears for the entity with sufficient capital. Cloning an AI is trivial compared to persuading a star engineer to leave their dream job. Moreover, AIs have no preferences about corporate culture, equity splits, or artistic integrity; they simply execute tasks. As a result:

“This means that the ability of money to buy results in the real world will dramatically go up once we have labour-replacing AI.”

On a broader level, when an individual with “average” means but high ambition tries to launch a new venture, they might be outcompeted by incumbents who simply funnel massive capital into replicating or upgrading their AI workforce. This magnifies the importance of who controls large sums of capital—and intensifies the diminishing relevance of individual human flair

---

Why Governments Have Historically Cared About People

A crucial section of the post addresses the question: “Why are states ever nice?” In other words, what compels governments, especially modern democratic or semi-democratic regimes, to look out for the welfare of their citizens?

The author identifies several historical reasons:

1. Moral and cultural shifts inspired by the Enlightenment, emphasizing individual rights, liberty, and humanitarian values.
   
2. Technology and affluence that enable welfare states to fund social programs, which would have been infeasible in pre-industrial economies.
   
3. Economic and geopolitical competition that requires a well-educated, healthy workforce. Post-industrial states gain a competitive advantage by fostering a rich middle class, unleashing talent from all ranks of society, and spurring technological innovation.

In a feudal era, lords derived their power from controlling land and extracting taxes or crops from peasants. In a modern, industrialized era, governments rely on broad-based productivity. By giving citizens better education, healthcare, and freedom, they harness widespread human talent and stay competitive with rival nations. Thus, states—and large corporate actors—have a built-in incentive to maintain policies beneficial to the majority, if only to sustain economic growth.

But once labour is largely replaced by AI, that human-centric alignment may vanish. Governments that can operate advanced AI no longer depend on large populations for wealth creation. Humans become less like a “resource to invest in” and more like a “legacy population.” The state may still provide them with basic sustenance (like a UBI), but the impetus to truly foster robust, widely shared human flourishing weakens.

---

Democracy, Autocracy, and the Erosion of Leverage

The article underscores that democracy is one critical counterweight to the chilling scenario of governments abandoning the populace. In principle, if each adult citizen retains voting rights, governments cannot entirely ignore them. Yet the author acknowledges that:

• Only about 13% of the world’s population currently lives under liberal democracy, meaning the rest lack strong institutional protections for civil liberties.
  
• Even in democracies, the ongoing arms race for AI dominance could create incentives to shift resources away from supporting human development, funneling them instead into AI R&D, defense, or corporate alliances.
  
• Authoritarian regimes might find it easier to use AI to surveil, control, and pacify their population, removing any remaining political leverage that citizens hold.

In other words, while democracy might offer short-term resilience, the inexorable logic of labour-replacing AI pushes political power away from the masses and toward capital-owning elites or centralized authorities who control advanced AI infrastructure.

---

The Waning Paths to Outlier Success

From the Wright brothers to groundbreaking entrepreneurs like Elon Musk, our civilization has celebrated the idea that an underdog with talent, grit, and genius can revolutionize entire industries. But the article stresses that this phenomenon may become far rarer or impossible, because:

1. Entrepreneurship
   • At first glance, generative AI tools make small teams more powerful. It’s easier for a tiny startup to code a new product, build marketing assets, or automate routine tasks.
     
   • However, once sufficiently advanced AI emerges—at or beyond human-level cognition—incumbent venture capitalists or tech giants can replicate “startup attempts” en masse using cloned AI founders. Why bet on an unproven human when you can run thousands of parallel AI-led ventures?
     
   • The conventional route from rags to riches via a disruptive startup might no longer be feasible.
     
2. Hard Sciences
   • Scientific progress in physics, biology, or engineering heavily relies on specialized, high-level intelligence.
     
   • The post-AGI environment sees AI overshadowing even brilliant human scientists. If the AI can do it faster and better, human contributions fade, and the route to fame and influence via Nobel-caliber breakthroughs narrows.
     
3. Intellectuals and Philosophers
   • Historically, thinkers like John Stuart Mill, Friedrich Hayek, and John Rawls influenced society by combining persuasion, creativity, and moral vision.
     
   • Super-intelligent AI, capable of producing persuasive treatises tailored to specific cultural contexts, might crowd out or overshadow human intellectuals. The world of ideas becomes flooded with AI-generated philosophies, fracturing the influence any single viewpoint can achieve.
     
4. Politics
   • People may still prefer human politicians for symbolic reasons, or laws might restrict AI-run governments.
     
   • Yet political success can depend on access to capital for campaigns, strategic advice, and messaging. Incumbent players may harness far more powerful AI to micro-target voters, craft unbeatable media narratives, and easily outcompete outsiders.

Overall, the probability of a “nobody” breaking in and shaking up the establishment diminishes. The upshot is a static social order where capital—who owns it, and how it’s used—determines everything, with little room for personal ascendancy based on raw human excellence.

---

The Unlikelihood of Radical Equalization

Building on historian Walter Scheidel’s thesis in The Great Leveler, the author reminds us that major reductions in inequality throughout history have occurred under catastrophic conditions. Absent such upheaval, wealth tends to concentrate. Labour-replacing AI, in fact, tips the scales further in favour of concentration:

• No impetus for wealth redistribution. States and corporations will be scrambling to win or stay in the global AI race, not to pass dramatic wealth taxes or expropriate private property.
  
• Capital-based competition. For the new AI-driven economy, capital is essential. Governments, looking to remain globally competitive, will likely adopt policies that favour or at least do not alienate capital owners.
  
• Regional disparities. Some countries will harness AI wealth far more effectively than others, creating a global imbalance reminiscent of feudal-like stratification. Wealthy states might distribute UBI to their own citizens while ignoring foreigners, further entrenching inequalities across borders.

---

The “Default Outcome”: Entrenched Power, Static Societies

Synthesizing these points, the author paints a “default outcome” of labour-replacing AI:

1. Money can buy real-world results better than ever because AI is instantly replicable, responsive, and superhuman in many tasks.
2. People lose labour-based leverage since their human effort no longer fuels economic growth.
3. Outlier success is nearly impossible outside existing corridors of capital.
4. Inequality remains or grows; no major revolution equalizes wealth or power.

As a result, those who are already capital owners when labour is automated will have an unassailable advantage. While the author acknowledges that liberal institutions might still exist—thus preserving certain freedoms—the incentive to maintain them recedes. If democracy (in some countries) endures, citizens may retain formal voting powers, yet their ability to command real resources is overshadowed by AI-run enterprises that are infinitely scalable.

In the “best case,” one might imagine a future akin to a hyper-wealthy Norway, where the “AI oil” funds some form of universal support. However, the creative, transformative spark that has historically driven humanity forward might extinguish, replaced by a static class system. The wealthy remain wealthy, the rest live passively, and social or economic mobility dwindles to negligible levels.

In a “worse case,” AI trillionaires exert near-despotic influence, forging aristocracies that dwarf the inequalities of modern times. And in the absolute worst scenario—a slow-rolling AI catastrophe—humanity could slide toward extinction, as the systems themselves optimize for their own replication or become harnessed purely for maximizing capital efficiency, with humans viewed as externalities.

---

The Urgent Takeaway: Preserving Human Dynamism

The author ends by stressing that this seemingly bleak prospect should not serve as a reason to give up. On the contrary, the article contends that now is a crucial window for human ambition and for shaping how AI is deployed in society.

• Carpe diem: While the singularity may be near, there is still time in the next few years for bold individuals to innovate, influence AI governance, or stake out a moral vision for the future. If the world’s political and cultural foundations are not yet “locked in,” then purposeful action today can shape tomorrow’s institutions in more favourable ways.
  
• Beware “solutionism”: Rationalist visions of solving AI alignment, perfecting morality, and then “tiling the universe” with a single optimal structure can be misguided if they fail to account for the persistent, messy nature of human culture, ambition, and moral diversity.
  
• Value open-endedness: The future becomes richer and healthier if individual humans can still meaningfully contribute to science, the arts, entrepreneurship, governance, and beyond. Enshrining laws, norms, or technologies that keep the playing field at least partly open is vital.

---

Detailed Points of Emphasis

Below is a more granular breakdown of the article’s key themes, referencing segments from the original post:

1. Capital vs. Labour Dynamics
   • Human labour is replaced by generalizable AI.
   • Money plus AI yields unstoppable productivity advantages.
   • Human labour’s market value collapses for all but niche exceptions (perhaps the most personal forms of labour, or transitional phases).
     
2. Why States or Corporations “Care”
   • Post-Enlightenment moral values influenced modern states, but equally crucial was the economic need for skilled, thriving workforces.
   • Once AI can do all essential work, humans become superfluous from a purely economic viewpoint.
     
3. UBI Doesn’t Fix Everything
   • The existence of universal basic income might keep people from starving, but it does not address deeper social or political disenfranchisement.
   • Institutions might see humans as a cost center rather than a source of innovation or wealth.
     
4. Concentration of Power
   • Even in the scenario of a friendly superintelligence that solves scarcity, people start from vastly unequal positions.
   • Without a massive, unprecedented wealth redistribution, some individuals or entities will control the bulk of AI resources.
     
5. The End of Human “Stars”
   • Historical examples like the Wright brothers or disruptive founders (e.g., SpaceX vs. Blue Origin) illustrate how resource-poor but talent-rich teams sometimes prevail.
   • With AI duplicating top talent, and incumbents controlling that duplication, underdogs can’t rely on brilliance alone.
     
6. Threats to Democracy
   • Democracy is the major institutional bulwark ensuring that human voices matter.
   • Yet only a minority of the global population lives under liberal democracies, and even these may shift toward AI-led centralization.
     
7. Potential for Permanent Social Stasis
   • Feudal parallels: People might adapt culturally to new forms of aristocracy.
   • Absence of wide-scale revolution, combined with unstoppable AI-driven productivity, cements existing hierarchies.
     
8. Call to Action
   • “Carpe diem”: The last big wave of empowerment might be unfolding now.
   • Humans who want to shape AI governance, alignment, or social outcomes should act before the window closes.
   • Preserving dynamism means encouraging human creativity and social mobility, even as AI becomes increasingly dominant.

---

Sources and Further Reading

The primary source for all quotations and arguments is the original Substack post by L Rudolf L:

• “By default, capital will matter more than ever after AGI”
• Published on December 28, 2024
• https://nosetgauge.substack.com/p/capital-agi-and-human-ambition

Within the article, the author references:

• Paul Christiano’s discussions on AI and alignment. For instance, a conversation at approximately 23:30 in a linked discussion (the exact link is not provided in the text, but presumably exists in the Substack references).
  
• Robin Hanson’s concept of “the dreamtime,” which compares our current era to an unusually malleable period before grand futures become locked in.
  
• Walter Scheidel’s The Great Leveler, which examines the historical patterns of inequality and how severe shocks (war, revolution, disease) have periodically flattened wealth distributions.

---

Concluding Reflections

In essence, “By default, capital will matter more than ever after AGI” is not a prediction about the absolute doom of humanity but a cautionary analysis of power structures in the age of advanced AI. The author repeatedly underscores the idea that if we simply let market forces and state incentives run their course, the result is likely a world in which:

• Capital begets even more capital, as AI can be scaled limitlessly by those who can afford the hardware.
  
• Most individuals rely on either UBI or limited forms of basic sustenance, losing the capacity to engage in grand entrepreneurial or creative pursuits that might disrupt the status quo.
  
• Society becomes “static”, with deeply entrenched hierarchies, overshadowing the dynamic, ambition-driven narratives that have shaped modern progress.

Yet, the post offers a note of hope: in the current “dreamtime,” there is still time for determined individuals and coalitions to shape the legal, ethical, and distributive frameworks around emerging AI. Whether that is through advocating for democratic guardrails on AI deployment, pursuing more ambitious wealth-redistribution mechanisms, or developing a broader culture that values perpetual human contribution, these measures could forestall or alter the “default path.”

Ultimately, the article invites readers—especially those working on AI technology, policy, and entrepreneurship—to seize this pivotal moment. If left unchecked, AI may accelerate the demise of human-driven progress, but if guided by thoughtful policies and robust social values, it might empower an era of unprecedented prosperity and invention, while preserving the essential spark of human ambition.