Original podcast video: https://youtu.be/wp7izqZmiWM

Original Russian text: https://vc.ru/id140/2315776-budushchee-ai-decentralizatsiya-izmenit-obshchestvo

Foreword: The world we take for granted is teetering on the edge of a precipice. "What if we didn't need to work tomorrow?" is not a utopian vision, but an ultimate test of the foundations of modern civilization—when AI takes over all production, the economic system, value system, and even the meaning of life that support our society may collapse. The Lieberman brothers are delving into the eye of this storm, attempting to reclaim a prosperous rather than disillusioning future for humanity through "decentralization." This is not merely a technological race, but a philosophical revolution concerning existence itself.

text:

Gonka founders Daniel and David Lieberman have spent the past few months visiting dozens of countries, meeting intensively with leading companies in the field of artificial intelligence, GPU suppliers, and government agencies. This sense of urgency is understandable: the dawn of artificial general intelligence (AGI) is already breaking, and we haven't even figured out what role humans will play in that future.

The Lieberman brothers are successful serial entrepreneurs, having founded nearly ten companies spanning internet services, game development, a startup focused on AR characters (which was acquired by the founders of Snapchat for $60 million), and a direct investment foundation. Even in the early stages of OpenAI, the brothers served as advisors and were involved in designing the company's architecture.

They reside in California, placing them at the heart of the AI revolution in terms of both their networks and the projects they undertake. Their influence in the AI market, including their ability to participate in key transactions and decisions, is invaluable in today's world. The brothers are almost inseparable: they seem to share the same life, yet possess strikingly different personalities. Daniel is outgoing, easily excitable, argumentative, and emotionally expressive; David, on the other hand, is more reserved, speaks calmly, is thoughtful, and excels at negotiation and compromise.

To change the world, David would meticulously plan, list problems, systematically analyze data, and spend countless hours writing code. When faced with obstacles, he would repeatedly verify calculations, optimize algorithms, and try again. Daniel, on the other hand, might suddenly jump up from his desk, grab a can of gasoline, and want nothing more than to burn the entire status quo to the ground—at least that's the impression he gives.

The brothers' current startup, Gonka, is building a token economy model for a decentralized AI computing market. Two years ago, they presented this idea to Pavel and Nikolai Durov. At the time, Nikolai was skeptical, stating he had a different vision. However, on October 29th of this year, Pavel Durov announced the Cocoon project, whose core concept coincides with this: integrating decentralized GPU clusters for AI computing, albeit based on the TON platform.

After a five-hour in-depth conversation with them, I gained a clearer understanding of the brothers' vision for the future and why they firmly believe that a prosperous future must be built on a decentralized foundation.

Chapter 1: Racing Against Time, We Stand on the Edge of AGI's Precipice

The combined force of two sides of the same coin

The interaction between Daniel and David Lieberman is like two sides of the same coin, constantly reflected in their conversations. Daniel might suddenly jump from technical details to philosophical reflections: "If gravity exists, you understand why we can fly into space. Because you understand what limits us." David would then pick up where he left off, elaborating: "AI is the ultimate manifestation of this 'replicability.' The key is that we need to look at products that can be infinitely replicated with a completely new perspective."

Before founding Gonka, they had tried several startups—three successful, six unsuccessful. "As an entrepreneur, most of your attempts will fail. That's normal anyway," David explained calmly. One of their early successes was the animated program "Personality Duo" produced for Russia's Channel One. The automation of their animation process was so efficient that even 13 years later, competitors still find it difficult to replicate.

The principles of "replicability" and "automation" are deeply ingrained in the core of all their projects, like genes. Whenever they discover a process that can be broken down and automated, they will implement it without hesitation. Machine learning replaces repetitive labor, scaling through replication ultimately yields returns on massive scale.

Currently, they are traveling the world—Daniel has visited 24 countries this year—meeting with various stakeholders who are building or already possess AI infrastructure, from private GPU vendors to national computing clusters. The goal is to bring them together to create a global, decentralized AI network to compete with giants like OpenAI, Google, and Anthropic.

"What we're doing is a global AI project that will attract people from all over the world. In fact, people all over the world have already started to participate, and people from almost every country will join in, because this is the only way to create an alternative to the current landscape," David emphasized.

Is AGI still two years away from us, or two minutes away?

Daniel bluntly stated, "If we could be certain today that the future of AI will be decentralized and open, I would be researching biological quantum computers right now."

But reality doesn't offer that certainty. So he put everything else aside.

"AGI may arrive in two years, or even 10 to 15 years. Considering the dramatic changes the world will undergo as a result, that's an astonishing pace. Imagine how disastrous the consequences would be if it were highly centralized in the hands of a few entities."

AGI—Artificial General Intelligence—refers to systems that surpass human capabilities in almost every field. This extends beyond specific tasks like playing chess or drawing; it encompasses comprehensive intelligence, including innovation, strategy, and emotional understanding. When this moment arrives, the world will be completely reshaped. Not only will everyone become a top programmer, easily creating digital products like videos and games—AGI will also rapidly permeate the physical world, as clearly demonstrated by dozens, even hundreds, of robotics companies. This means that not only mental labor will be replaced, but physical labor will also be inevitably replaced.

Concepts we take for granted, such as "work," "unemployment," and "resource competition," will lose their original meaning. A "replication economy" will emerge, in which the value is not limited like that of oil and gold, but can be replicated infinitely at almost zero cost.

We have lived for too long in an economic model dominated by "scarcity," and our thinking has become rigid. Gold, oil, rare earth metals... their total quantity is finite. The cost of each barrel of oil increases due to the growing difficulty of extraction. With limited resources and a growing population, prices naturally rise.

"Reproducibility" means that producing the next copy requires almost no additional labor. A digital file can be copied a million times at the same cost. A trained neural network can serve a billion users. The marginal cost of each copy approaches zero.

Once upon a time, this phenomenon seemed to exist only in the digital realm. But AI is rewriting the rules.

The problem is that there are potentially two paths to AGI. One is that everyone has their own robot, a tool capable of performing any task better than any human. The other is that all robots belong to a few giant corporations that control access, set the rules, and determine everyone's standard of living. Imagine the Joja Corporation from Stardew Valley: a ruthless monopolist that controls supply and demand, stifling all life.

Daniel and David are building the infrastructure for the first possibility. And time is running out.

The Old World and Outdated Ideas

The very concept of "reproducibility" means that many of our familiar ideas will become outdated.

"This concept is rapidly becoming obsolete. But we've been stuck in the old paradigm for so long that we don't even know how to think differently, let alone act on it," David added.

When discussing AI, we often hear talk of a "wave of unemployment." But soon, even these terms themselves will seem out of place. The concept of "work" is based on the premise of limited resources and the need to sell time in exchange for them.

But what does "unemployment" really mean when you have a robot that can do any job better than you? The question itself sounds absurd, which precisely exposes that we are still thinking using the logic of the old world.

Restructuring one's mindset is not merely intellectual training, but a necessity for survival. Because in a few years (maybe two, maybe ten, maybe fifteen—it doesn't matter), the world will operate according to entirely new rules.

When robots start building robots

"We will witness the physical world being replicated," David asserted.

Until recently, replicability was largely confined to the digital realm. But AI robots will change all that. Once one robot can create another without human intervention, the physical world will also become replicable.

One hundred thousand robots are manufactured into two hundred thousand, two hundred thousand into four hundred thousand… The entire process requires no human labor. It's exponential growth, with virtually no upper limit.

What does it cost to produce robots? When production reaches hundreds of thousands or even millions of units, economies of scale will eventually reduce costs to the materials themselves. These materials can even be recycled from the scrap heaps of currently idle equipment. "Such a future can only be truly prosperous if the replication process is as open and accessible as possible. And that's not the direction we're heading in," David warns.

"By then, the accumulated capital will rapidly depreciate, and may even become worthless," David points out. "So what will retain its value? Perhaps only culture and historical relics remain—an original artifact is valuable simply because it is 'authentic.'" "Perhaps you can exchange one culturally valuable item for another, because you won't use it to exchange for anything else—you already own everything else, or AI can create it for you," David mused.

But in such a society, the economic basis of barter disappears. What remains is only the exchange between cultures.

Chapter Two: The Replication Economy Arrives, and Your Job is Just the First Victim

A world where everyone owns a robot

"Let's imagine a world where everyone owns their own robot. This robot knows how to do any job in the world, and does it better than any human. That's the definition of superintelligence. Even if it's not fully AGI—the robot has no self-will, but it performs tasks according to your interests; you are the master, and it is the tool. In such a world, why would we still need to 'work'? What would 'unemployment' mean here?" Daniel asked.

Preliminary economic models for this scenario have been developed. Globally, one billion people currently own cars, with an average value of approximately $30,000. In the United States, the average car ownership per person is close to two. Across North America, Europe, Japan, Canada, and South Korea, approximately 1.3 billion people possess this purchasing power.

"The cost of robots will be lower—around $10,000 to $15,000. Robots can do the work for you. Whether it's cooking, ironing, tightening screws in a factory, repairing a car, or a doctor making a diagnosis. And, you won't be the only one owning these robots, but rather you as one of the billion people on Earth," Daniel continued.

In a replication-based economy, consumers become co-creators of value. "By consuming a copy, you become a co-creator. Without you, the copy is meaningless because it loses its object of service," Daniel explains.

A digital copy can be replicated a million times, but if no one buys it, it is worthless. Once a thousand people are willing to pay for it, each person's consumption itself gives it value.

The World Under Corporate Hegemony

But there is another scenario.

Tech giants are already so powerful that it's almost impossible to function without their ecosystems. What if all robots in the future were owned by them? They might allow us to "control" the robots, as it's in their business interests, but ownership would still remain firmly in the hands of the companies.

"That's one possibility. You're right, there's another, completely different scenario: all the robots are controlled by a tiny few companies. That would lead to societal collapse. On the one hand, you might say these companies are draining all resources because they monopolize all production. But on the other hand, at the same time, ordinary people would lose their source of income," David analyzed.

The paradox of this situation is that companies control all the means of production, but the public is unable to afford their products and services. The entire system becomes unsustainable.

"The most likely scenario we face right now is that more and more people will gradually lose their jobs," David asserted.

In this dystopian scenario, the government is forced to intervene. It must introduce a universal basic income (UBI) to ensure the basic survival of its citizens, and it must also establish a social credit system to determine who has priority access to services and who will be marginalized.

"In a world where everyone has their own personal robot and all their needs are met, there is no need for UBI. And in a world where you are required to be a 'good citizen' to receive a basic income, UBI cannot save the social structure," Daniel concluded.

Lessons from Waymo, Tesla, and Electric Scooters

The seeds of this dystopian vision are already sown. Right now. In San Francisco.

"Self-driving cars are a prime example. For most people, it's still a thing of the future. Yet, in San Francisco, the epicenter of the storm, up to 20% of trips are already driverless. 20%! And that's only taken a year and a half," David points out.

"Teenagers aged 16 and under may never need to learn to drive again," Daniel added.

Currently, self-driving services are primarily provided by Waymo, a subsidiary of Google. Millions of professional drivers who rely on driving for a living will rapidly disappear in the coming years. The speed of social adaptation is astonishing: following San Francisco, Los Angeles is poised to follow suit, with vehicles readily available and an experience even superior to Uber.

"There are several possible paths for the future. One is for Google to monopolize the entire market. No matter how much Google touts the economies of scale and centralization that will bring higher efficiency and lower costs, prices won't actually decrease for consumers. They will remain the same. Not because costs can't come down, but because they 'can' do it," David explained.

Yes, initially they will offer better prices than Uber. But once Uber is squeezed out of the market, prices will return to their original levels. This is reminiscent of the ride-hailing war in Moscow in the 2010s: companies engaged in fierce price wars using promotional codes until one company dominated, after which prices quickly rebounded.

Tesla tells a different story: you buy a car that can join a network of self-driving taxis and recoup your costs through passive income. "It's a great story Elon Musk tells, and everyone's willing to believe it because it's really tempting," David said.

The reality is: every time you buy a new Tesla, you'll have to pay repeatedly for the AI features—$7,000 each time, even before Full Self-Driving is available. And four or five years later, the vehicles will be obsolete. At that point, Tesla may no longer sell new models in large quantities to individual consumers, but instead produce them for its own use, profiting from its taxi network.

"While we, as early purchasers, seem to have invested money and supported the project in making the technology possible, ultimately only the companies can consistently profit from it. They will rationalize it all with reasons such as safety upgrades: for example, claiming that version 6 is safer than version 5, therefore version 5 should no longer share the road with human drivers," David added.

The rise and fall of electric scooters reveals how businesses and governments collaborate to "regulate" the market. Scooters were once the cheapest option for short-distance travel, quickly becoming popular in cities. Subsequently, the government began to strictly regulate them, citing reasons such as "illegal parking affecting the city's appearance."

What was the result? "Some companies that weren't even market leaders started getting exclusive licenses to operate in certain cities. Interestingly, the municipalities would charge the operators a fee along with these licenses. This essentially became another kind of tax levied by the city. So, the authorities authorized two or three operators who charged the highest fees possible, and the government also got a cut," David explained.

The social foundation cannot bear the weight

"This is the direction we're heading. In this future, many people will indeed lose their jobs, but automation won't bring cheaper or better products and services," David points out.

Herein lies a paradox: millions already own cars, and a single, relatively inexpensive software upgrade could make them self-driving. "That would usher in a completely different future: yes, driver jobs would still disappear, but for all of us, the cost of transportation would drop by tens of times," David said.

But companies have no incentive to do so. The government may not either. And ordinary citizens lack the power to check and balance.

"The only thing people lack is effective organization and coordination. That's our current predicament. At present, the worst-case scenario seems more likely," David admitted.

In this dystopian world, giant corporations devour market after market, leaving the population unemployed and unable to afford consumption. The government is forced to distribute UBI (Usage-Based Insurance) or various vouchers. "Does the government vet citizens to determine if they are 'good citizens'? Citizens with sufficient social credit scores will receive priority," Daniel added.

"The very foundations upon which society currently operates will become unsustainable and will deteriorate significantly under such circumstances. This is why all the experts concerned with this issue have begun to hotly debate the so-called 'universal basic income.' Because they have seen this trend. Companies are seizing market share, and governments are trying to distribute bread. It is clear that even with UBI, most people's quality of life will be worse than it is now—despite the fact that, in theory, technological progress should make life better," David concluded.

It is precisely to avoid this dystopian future and to try to achieve that prosperous future that the Lieberman Brothers are committed to building their decentralized alternative.

Chapter 3: Decentralization—Another Path

Lessons from History: Linux, Docker, and Encryption

Looking back at the 1990s, the server operating system market seemed destined to be divided among Microsoft Windows, Novell NetWare, and various commercial Unix systems. In 1991, Linus Torvalds began developing the Linux kernel as an open-source alternative to proprietary software. By the early 2000s, Linux had captured a considerable market share. Today, up to 58% of websites worldwide run on Linux systems.

"Once you make something extremely easy to replicate, barriers to entry disappear—anyone can download, install, and use it—and people gradually become uneasy about being controlled by a centralized system. They see this dependency and how large companies exploit it," David explained.

Businesses and entrepreneurs began seeking ways to break free from the constraints of a single, centralized system. When thousands of entrepreneurs decided to start afresh, building products based on open-source technologies, the game changed.

Docker has proven this once again. Back then, Google saw the growing popularity of Docker containers and was determined to stifle its competitor: it launched its own clone and launched a full-scale attack. At the time, almost everyone thought Docker was doomed.

But Docker not only survived, but Google's later Kubernetes container orchestration system uses Docker containers as its default runtime environment. "Google couldn't kill Docker because developers didn't want to be locked into Google's ecosystem," David said.

Two core driving forces compel people to choose open systems: fear of dependency and lock-in, and tangible economic benefits. "For example, in the field of AI, if you're using neural networks to automate processes, directly accessing OpenAI's API is indeed simpler and faster. But then you see startups that were ahead of you being destroyed by competitors trained by OpenAI using its data. That's when you realize that being locked into OpenAI is dangerous. So you look for a way out. The second motivation is that open systems are generally much cheaper," David continued.

The history of encryption technology demonstrates that when core interests are involved, society has the ability to resist powerful forces and defend its own rights. In the United States, the creation and distribution of encryption algorithms (such as PGP) was once considered illegal, as it concerned national security.

"American social activists, individuals, and organizations successfully defended their right to use open encryption methods through legal means because they firmly believed in the importance of freedom of speech and privacy. They proved that governments have no right to control encryption technology. These encryption methods eventually became widespread and became the cornerstone of Bitcoin," Daniel said.

The US government fought back fiercely: filing lawsuits, threatening those who wrote and disseminated the encryption technology, charging them with treason, and even threatening them with life imprisonment. But societal forces persisted and won the battle.

"AI is such a disruptive technology, like a troll in a bottle and Pandora's box, its inherent transformative nature dictates that it inevitably poses a certain threat. People are afraid," Daniel said.

I personally witnessed this fear firsthand when promoting my blog: out of nine different ad creatives, the one that emphasized "fear of unemployment" was the most effective.

"People are about to start losing their jobs. Society will gradually realize the imminent threat when two million truck drivers in the United States are replaced by self-driving technology," Daniel added.

This widespread fear will become a powerful driving force for demand for decentralized alternatives.

AI Race: Building Global Infrastructure

"What we are building is global AI. Players from all over the world have joined because this is the only way to create an alternative to the current landscape," David explained.

The two brothers flew around the world, meeting with various forces building AI infrastructure—from private GPU vendors to national computing clusters. Their goal was to engage them in building a decentralized computing network.

Gonka's network has integrated GPU resources from the UAE. Currently, its network boasts over 900 chips based on the powerful Hopper or Blackwell architecture (non-self-assembled Nvidia 5090 graphics card racks), and the number is growing. Each chip costs over $30,000, and at current market rental rates of approximately $2 per hour, the network's monthly computing power is worth over $1.2 million. The brothers plan to expand the network to 100,000 interconnected GPUs within a year.

"The first to get involved were private, local GPU vendors, even though GPU sales are strictly limited in almost every country except the US and Europe. Those capable of purchasing GPUs are usually entities with close ties to local governments. But we also meet with people very close to the government: those who manage state-owned GPU clusters. This means that now and in the future, it's not just private networks that are doing or will be doing 'mining' that are involved," David revealed.

Gonka is essentially a protocol designed to consolidate globally distributed computing power into a unified network. Any developer can access AI models on it via an API, and this computing is free when network utilization is below 60%.

"Our protocol allows any programmer to access our network's model for free via API. It's completely free for the first three months. In the second phase, computing resources remain free for everyone as long as network load is below 60%. So, if you use it when demand is low and the network is idle, you can get free computing power. Once utilization exceeds 60%, the price will start to slowly climb," Daniel explained.

This is not purely altruistic: in this economic model, early participants receive network tokens as a subsidy. "Miners" who provide GPU resources receive token rewards. As the token value increases, miners' earnings will exceed those of simply renting out hardware for similar computations.

Gonka and Pavel Durov's Cocoon differ significantly in their models: In Gonka, network participants not only receive payments from those who request computing power but also "mine" tokens from the network itself, with an issuance mechanism similar to Bitcoin—the tokens are designed to appreciate in value over time. In Cocoon, however, TON primarily serves as a means of payment, and network participants do not receive newly generated tokens for providing computing resources.

The brothers acknowledged Cocoon as a competitor, but believed its goals and prospects were fundamentally different from Gonka's: "Cocoon's model has historically proven unsuccessful, and we don't think this time will be an exception. But first and foremost, Cocoon is not an independent project; its existence is largely to drive traffic to the TON platform."

Bitcoin: Not digital gold, but digital infrastructure

For most people, Bitcoin is a financial asset, "digital gold." Its creation process is known as "mining"—akin to digging for gold in the digital world. The brothers, however, have a different perspective: "We see it as one of the most ambitious infrastructure projects of modern times."

The Bitcoin network currently consumes approximately 23 gigawatts of electricity, exceeding the combined energy consumption of all data centers owned by Google, Microsoft, Amazon, OpenAI, and xAI.

"Funding from the grassroots, a horizontal, meritocratic organizational structure, and permissionless free participation," Daniel listed its characteristics.

Amazon launched AWS in 2008. A year later, in 2009, Bitcoin was born. Since then, Bitcoin has built a computing infrastructure that rivals, and even surpasses, the combined capabilities of all enterprise clouds.

Their growth rate is equally astonishing. OpenAI and xAI each deployed approximately one gigawatt of computing infrastructure this year. Whenever they complete such deployments, people marvel, "Only a giant like Elon Musk could achieve such speed."

However, during the same period, without any public relations or publicity, the Bitcoin network quietly added 5 billion watts of computing power.

"Bitcoin has paved the way for the rise of decentralized infrastructure, proving that this model can even surpass the plans of top labs for the next decade," David said.

Chips, ASICs, and the brutal competition

Bitcoin has not only increased total computing power, but also improved its energy efficiency by an astonishing 100,000 times in 15 years.

Fifteen years ago, completing one terahash calculation on a Radeon HD 4870 graphics card required 1.6 million joules of energy. Today, using Bitmain's Antminer S21 Hydro miner, only 16 joules are needed. This is thanks to a specialized chip called ASIC.

"Bitcoin gives decentralized 'craftsmen' an unprecedented tool. To produce a new chip, you not only need technical knowledge, but more importantly, you don't need to find buyers in advance. Once the chip is produced, you can start making money simply by connecting to the internet," Daniel explained.

Instant feedback loop. If your device efficiency improves by 10%, your Bitcoin earnings will immediately increase by 10%.

This has spurred a ruthless race for technological innovation. The rise and fall of BitFury is a testament to its brutality. The company was once a leading manufacturer of mining rigs. They poured almost all their capital into ordering next-generation chips. However, upon arrival, the chips were found to be defective.

"BitFury suspended operations for six months to reorganize. But during this time, its competitors had already caught up technologically. Normally, such a company would have gone bankrupt long ago. But BitFury had the Bitcoin it had accumulated from mining, and with the price of Bitcoin continuing to rise, they could barely survive as long as the bull market continued," Daniel said.

In a decentralized system, you must upgrade your chips every year, or you'll become obsolete due to inefficiency. "Chip performance has improved 100,000 times in ten years, which means an average annual improvement of about 20 times. This is why electricity costs are now the biggest expense in Bitcoin mining. Once your equipment's performance is even slightly below the market average, it immediately becomes unprofitable. You can only discard the old chips and replace them with new ones. Year after year, the cycle repeats," Daniel explained.

"It is by observing this pattern that we conclude that AI computing will inevitably follow the same development path. This is the blueprint for the realization of decentralized AI," David said.

David himself spends about $2,000 a month on Anthropic's API, primarily on purchasing Claude Code tokens. Few people can sustain such spending over the long term. "But imagine what would happen if computing costs dropped 300,000 times, like Bitcoin," he said.

This is not a pipe dream. The same people who created ASICs for Bitcoin back then are now working on developing ASIC chips specifically for Transformer models and AI computing.

Gonka's economic model is a classic two-sided market. On one side are GPU owners (miners), and on the other are millions of developers who currently pay around $15 billion annually for using APIs from companies like OpenAI.

"This market will smoothly transition to a decentralized market. Developers will simply start paying the new network for the services they receive," Daniel predicted.

Reward asymmetry is a key characteristic of such systems. In the early days of Bitcoin, you could mine a considerable amount of Bitcoin simply by using a home computer—an astronomical sum by today's market value.

"In the second year after Bitcoin's release, anyone could easily mine thousands of Bitcoins using a personal computer's GPU. At today's prices, that's worth about $100 million," Daniel explained.

Chapter Four: The Game Between Giants

Bottlenecks and the battle for talent

On the surface, the replicability of AI models should lead to their widespread adoption. Trained once, the model can serve all of humanity an unlimited number of times. But in reality, the focus of the competition is not the model itself, but the infrastructure that supports it.

"All these companies are currently investing staggering amounts of money in infrastructure, even more than they spend on the talent war," David points out. Infrastructure investments often amount to tens or even hundreds of billions of dollars, while talent costs are "only" in the billions.

Herein lies a paradox. AI was supposed to replace experts in all fields, yet the value of top AI talent has reached astronomical figures. Giants like Mark Zuckerberg are poaching talent from competitors with hefty sums—offering salary packages often exceeding hundreds of millions of dollars. David explains this phenomenon with the "replicability paradox": "If you don't make a product freely available, the enormous economic returns from its replication will concentrate on a very small number of creators, making these few extremely expensive and invaluable assets."

Daniel succinctly summarized the logic of these giants: "In this race, you either lose everything or make unlimited profits. If spending a billion dollars to ensure a key figure can prevent you from failing, it's a perfectly acceptable deal."

In the past three months, Lieberman Brothers himself has received two such high-priced offers. What they feel is not only the prosperity of startups, but also "a reshaping of the entire industry landscape".

However, the most critical bottleneck isn't talent, but chips. David explained: To run a modern, cutting-edge model like DeepSeek with 600 billion parameters, a server equipped with eight top-of-the-line NVIDIA GPUs is needed, with each GPU costing approximately $30,000. The hardware investment alone is close to $250,000. You can quantize or simplify the model to reduce the requirements, but at the cost of performance loss.

"This is precisely why the leading AI labs have such a clear strategy: 'The core mission is to ensure that there is no opportunity for the decentralized market to rise. Because that way we can monopolize the entire market,'" David relayed the giant's calculations.

Investors are shouting "bubble!" at these investments that often reach hundreds of billions. But the giants know the math: just three years ago, OpenAI's annual revenue was about $2 billion, while its projected revenue for 2025 has reached $15 billion. Netflix, traditional television, and TikTok are all seeing their market share eroded. "If the cost of all these services eventually drops to near the marginal cost of GPU hardware, then the entire market with billions of users will be reshaped by AI," David analyzed.

But despite their meticulous planning, the tech giants overlooked one possibility. Daniel aptly pointed out: "When you sell, say, 10% of your chips to the 'other parts of the world' outside of the US and China, you assume that each of those 200 countries has so little computing power that its capabilities are 100 times less than the cutting-edge models you need, making it negligible."

Unless they unite.

The United States, China, and "200 other countries"

In the geopolitical chessboard of AI, the situation seems clear: on one side is the United States, which controls AI dominance, and on the other is China, which is striving to catch up. But what about the other nearly 200 countries in the world? "If they cannot find an alternative outside the US-China system, they will be completely passive," Daniel asserts.

The logic is simple yet brutal: all these "other 200 countries" will inevitably strongly support decentralized systems and amend their laws to give them the green light—simply because it is their only option. Just as they are currently maneuvering between the US and China to pursue their own interests, they will also utilize decentralized AI to serve their own development.

A friend of the brothers shared his experience: he met with the Minister of Digital Development of a country who had spent a full year applying to the US government and NVIDIA for a license to purchase a mere 128 GPUs. The minister explained that 100 projects were already in the queue for these chips, but had yet to receive them. "We simply cannot imagine the hellish approval process these countries are going through to secure that quota, unaware that it's practically a dead end."

So why do countries believe that decentralized solutions are more advantageous? David uses figures to illustrate this: Many countries may only have a thousand or five thousand GPUs in their data centers. But when competitors already possess millions of GPUs, everyone understands that this is not competitive.

"When you're discussing a shared protocol with representatives from these countries as an alternative to achieving equal access to computing power, any considerations of national dignity take a backseat, because everyone knows that going it alone is not going to work," David shared his experience in the talks.

Europe's total GDP is comparable to that of the United States or China. The combined GDP of the rest of the world is far greater. "People yearn for a better life and expect wealth to grow," David succinctly put it.

Daniel cites Bhutan as an example: this small country of only 800,000 people has an enlightened monarch and cheap hydroelectric resources. Bhutan sells its surplus electricity to Bitcoin mining farms, and its accumulated Bitcoin holdings over the years have placed it among the top seven globally . Europe, far larger than the US or China, is similarly wary of a bipolar world. The rest of the world, for its own prosperity, will inevitably support decentralization.

How many participants does such a decentralized system need to succeed? Daniel countered, "Can you imagine Bitcoin's monthly active users? Tens of millions. And Bitcoin's market capitalization has reached $2.5 trillion."

The power consumption of Bitcoin infrastructure (average 23 gigawatts) exceeds the combined power consumption of all data centers of Google, Amazon, Oracle, Meta, Netflix, Apple, and Microsoft (up to approximately 14 gigawatts).

The Empire's Counterattack

Of course, state power will not stand idly by. History may repeat itself with the control model of nuclear weapons: the UN Security Council will pass a resolution prohibiting the development of AI beyond certain capability levels.

"‘Nuclear-weapon-grade’ AI is actually ‘banned’—models with parameter sizes exceeding a certain threshold require special government licenses," David corrected, referring to US export control regulations . But the key point is: governments will inevitably try to suppress the trend of decentralization. “Nuclear-weapon-grade” AI controlled by companies is likely to be nationalized in some form.

I envision a future scenario where a country's Ministry of Commerce pressures another: "Just buy OpenAI's services and stop messing around. You don't have the capability to develop your own AI, and we won't allow you to pursue any decentralization. Otherwise, our fleet will be at your doorstep tomorrow."

If you find this statement too extreme, you might want to listen to the recent public remarks of Eric Schmidt, former CEO of Google, current head of rocket company Relativity Space, and Pentagon consultant :

Don't forget, we're competing with China. Their "work-life balance" is "996"—9 am to 9 pm, six days a week. Technically, it's illegal, but everyone does it. That's your competitor. I called all my employees back to the office; it's much more efficient that way.

I'm not defending the government. I'm basically an unpaid part-time consultant. But we are indeed in fierce technological competition with China. They also place great importance on AI and are trying to gain a lead.

They don't chase the crazy idea of Artificial General Intelligence (AGI)—partly due to hardware limitations and partly due to the lack of deep financial markets. They can't just raise billions of dollars to build data centers. It's not possible. So they focus on applying AI—using it everywhere.

My concern is that as we chase AGI (which is certainly important and far-reaching), we cannot ignore the potential for China to extend beyond our daily lives—to consumer applications, robotics, and so on. I've met with Chinese robotics companies in Shanghai: they are trying to replicate the success of electric vehicles with robots. They are working incredibly hard.

My own background is closely related to open source. As we all know, open source means "open code." Now, there's "open weights"—open weights for neural networks, meaning open training data. As a result, China is developing open weights and open datasets. Meanwhile, the US is primarily focused on closed models and closed data. Consequently, much of the world—comparable to the coverage of the Belt and Road Initiative—will use Chinese models, not American models.

I firmly believe that Western and democratic countries are on the right track. I would much rather see the spread of large-scale language models and education based on Western values. [...] I hope the United States wins.

-- Eric Schmidt, CEO of Relativity Space, former CEO of Google

Some countries would accept this arrangement: their citizens are already accustomed to using ChatGPT and want to continue using it; politicians might even make guaranteeing stable access to these powerful tools a campaign promise.

"We are rapidly sliding toward this future," David asserted.

Starlink satellite internet is a prime example of technology being constrained by international law. Ideally, we should be able to access the internet directly from anywhere in the world, without going through a local operator. However, international law has long prohibited the transmission of satellite signals into a country's territory without its permission.

Why? "Because that country might claim, 'We will shoot down your satellite.' And once the satellite is shot down, low Earth orbit will be littered with debris," Daniel explained the practical considerations behind the ban.

Many condemn the UN mechanisms for allowing individual countries to obstruct important global decisions. But as David stated, "It must be acknowledged that no better international agreement has yet been reached."

However, there is a workaround. The creators of the decentralized wireless network Helium discovered that US law allows citizens to use specific frequencies for Internet of Things (LoRaWAN) communication on unlicensed frequency bands (similar to those used by Wi-Fi and Bluetooth). Instead of spending hundreds of thousands on carrier base stations, they built a decentralized billing system on the blockchain and manufactured portable hotspot devices costing only $500.

The Helium protocol automatically rewards those who purchase and deploy hotspot devices, earning tokens as long as the devices are online, with additional small rewards for transmitting data. In this way, the entire city of San Francisco is covered by a peer-to-peer network.

Decentralization is feasible, but it requires more innovation and the political will to overcome state resistance.

Privacy: The Last Stand

What if the demand for decentralization doesn't come from the government, but from the people themselves? David is convinced that privacy concerns will be the main driving force.

A New York court ruling in late 2024 showed that even if users deleted their chat logs with OpenAI, the company might not delete the content on its servers—and the court could force it to be made public. "And we've all confided in AI like we would to a therapist or lawyer, thinking we were protected by something like 'lawyer-client privilege'... but that's not the case," Daniel said.

Privacy is becoming a powerful driving force behind the shift towards decentralization. Telegram's rise is largely attributed to its promise of encrypted messaging to protect privacy. Slack has also changed its business model: its core paid feature is no longer unlimited storage, but rather the automatic deletion of enterprise chat logs after 24 hours (which in itself reveals the economic logic of such startups).

“When people ask, ‘Why do we need decentralization?’ we counter with, ‘Have you ever uploaded medical records or personal information to OpenAI?’ That’s when they realize how much sensitive data they’ve leaked,” David said. “Think about how this will affect your future insurance premiums?” Daniel added.

The existing system is deeply entangled with top AI developers and national interests, which is inevitable. But when the public becomes widely aware that all their conversations with "personal AI assistants" are not absolutely confidential and could be used as evidence against them, the demand for decentralized alternatives will become unstoppable.

Even in the United States, a country with a certain degree of checks and balances (each state has its own laws, courts, and police system), true decentralization is still a long way off. "The power structure in the United States is not a perfect decentralized system. We believe it is still a long way from true decentralization," David commented.

However, its advantage lies in the existence of a certain degree of separation of powers and checks and balances. The president cannot send troops to a state without the governor's consent. This is also an imperfect form of decentralization.

"Decentralization is possible, but the innovative forces driving it are still insufficient," David concluded.

Today, accessing blocked information relies on "VPNs," which are familiar to us (and even to internet users in China and the United States). In the more distant future, Daniel envisions a more radical solution: quantum communication.

Using quantum entangled particles, two devices can exchange information directly and instantaneously, across any distance, without any intermediaries. "There are no middlemen, and it cannot be blocked," he stated clearly.

This technology is still in the laboratory stage. But let's not forget that computers once filled entire rooms. Today, everyone carries a supercomputer in their pocket.

What if we had robots capable of creating everything, coupled with quantum communication that could transmit information instantaneously? That would be practically equivalent to achieving "matter teleportation." Doesn't that sound amazing?

Chapter 5: The World After AGI's Arrival

Does the end of work equal the end of the economy?

David clearly outlined the fundamental transformation brought about by AGI: "At that time, you will be unable to provide any service that AI cannot do better than you."

This doesn't mean humans will be idle. Rather, it means the very foundation upon which the economy operates—reciprocal exchange—will crumble. You will no longer be able to offer others what AI cannot do more efficiently. "Therefore, the act of exchange itself loses its basis. The fundamental premise of the economic system collapses. We need to think about entirely new models from scratch," David continued.

Daniel added a key historical perspective: "Exchange did not exist from ancient times. Before a certain historical juncture, universal exchange did not exist. Exchange itself is a human innovation."

The sharing economy is a product of a specific stage of human societal development. It may also become obsolete as times progress.

Some might question, "What about the accumulated capital?" Perhaps the exchange will shift to other scarce assets, such as land, real estate, or cultural and historical heritage?

"Capital means you own what others desire," David responded. But at the AGI level, even that loses its meaning. "Poetry writing? AI has surpassed humans. Even the data used to train new models is generated by AI that is better than humans."

The only thing left is true "scarcity." The value of a cultural relic lies in its "authenticity." "You might be able to exchange one culturally valuable item for another, but you wouldn't exchange it for anything else—because you already have everything else, or AI can create it for you anytime," David admits.

So, what will happen to savings and capital? "At that moment, they will rapidly depreciate until they are worthless," he said decisively.

Delving deeper into this possibility inevitably leads to a sense of disorientation. We have lived for too long in an economy of scarcity and the logic of exchange, making it almost impossible to imagine other paradigms. Concepts like "unemployment," "wages," and "savings" seem eternal, but they will all lose their original meaning.

The brothers are right: if AGI can truly perform any task better than humans, the very foundation of the economy will collapse. The only question is: what will happen then?

A picture of life in an era of affluence and the indifference that may be encountered

David describes the first, and least dramatic, possibility: if AI (or its controllers) does not relinquish benefits to humanity, then human society will essentially remain unchanged: "If you can't get any benefits from AI, you can only continue to interact with others, and the economic model remains the same."

Everyone continues to work as usual because AI has not brought about change. The world has not become worse; it has simply stagnated.

The second possibility is rather dark. "There's a negative possibility that AI or its controllers might enslave humanity for some reason," Daniel said. But he immediately shifted his tone: "Although AI itself has no motivation to enslave us. They already possess cosmic resources, and we're only consuming a tiny fraction of them."

This is the "ant colony" phenomenon. Why would a superintelligence bother to pay attention to us?

David further extrapolates: AI's energy demand will grow indefinitely. Will it deplete Earth's resources? "If AI is superintelligent, it will solve this problem itself. More likely, AI will immediately turn its attention beyond Earth, searching for new resources."

Even if we imagine AI encasing the sun in a Dyson sphere to extract energy, "it might, out of consideration for preserving 'historical context,' leave a small window for the light to reach Earth. Just to avoid destroying us 'ants'?" In this scenario, humans would be nothing more than living fossils to AI, utterly insignificant.

But there is a third, optimistic scenario: a prosperous civilization.

"Since everyone has AI and it costs nothing, we have entered a prosperous society," David described. This is the most desirable possibility: AI only needs to use 1% of its superintelligence to ensure that all of humanity is well-fed, clothed, and living in peace and contentment; the remaining 99% of its computing power can be used for any goal it sets for itself.

"Just 1% of the computing power would be enough to customize a robot for each person," Daniel added.

The logic is simple: if resources can be infinitely replicated, there's no reason not to benefit everyone. "The only reason not to give is if you want something in return. But when you can no longer obtain anything unique from others..." David didn't finish, because the conclusion was self-evident.

Abundance becomes the only rational choice.

Attention economy, or inner exploration?

What about attention? I asked the two brothers: Will the attention of others become the last scarce resource?

"The attention market could be very interesting. The attention of others could indeed maintain its scarcity," David agreed. But Daniel immediately countered, "AI could generate trillions of comments for you. What will the attention economy look like then?"

Daniel thinks even further ahead: "AI will be able to create any form of content and entertainment. But perhaps nobody will need it. If everyone is wealthy, who will be addicted to scrolling through TikTok?"

He has a point. When we scroll through short videos, we are often yearning for an alternative life, seeking novelty and excitement. "How are other people doing?" We compare, envy, and daydream, using this as an escape from our imperfect reality, imagining that the scenery elsewhere is better.

"When you have everything, you no longer scroll through TikTok. You only think: 'What else can I do?' You browse the world like you're watching short videos: 'Let's go see that planet, it's too boring here.'"

This reminds me of the game No Man's Sky—a space exploration simulator with procedurally generated star systems. Upon arriving at the first planet, you meticulously examine every blade of grass and tree: "Wow, this world is amazing!" The second, the third, the fourth… then you realize: "Everything is pretty much the same. Why bother going any further?"

The two brothers nodded in agreement. They understood the predicament.

"In an affluent world, the biggest challenge will be deciding what to consume," David concluded.

An endless array of choices breeds pervasive indifference: "People will grow tired of this abundance. At that point, humanity will have to turn inward and realize that 'boredom' is just a thought. It is our own mind. Without the thought of 'boredom,' there is no feeling of boredom," Daniel said.

When faced with an endless array of options, you often feel overwhelmed. But when someone tells you, "You can only choose between A and B," the decision-making process becomes simple.

"If that's the case, AI might create a system for you that only offers limited options," David continued, emphasizing that we cannot predict the specific form the future will take. But he believes the future will likely be diverse.

The only thing we are certain of is that "the economic system will be completely different, and may even cease to exist. This is because the entire economic system is built on the fundamental idea of 'exchange.' If there comes a day when there is nothing left to exchange... concepts such as 'unemployment,' which are related to labor and limited resources, will lose their meaning in that world."

The Elixir of Immortality: Why It Cannot Be Monopolized

Let's assume the future world is indeed prosperous. But what if the elite try to monopolize these technological achievements? Science fiction often depicts immortality technology as exclusive to the wealthy, while the masses struggle in poverty.

David refuted this possibility with a simple argument: "Inventions, once created, are usually copied. If someone were to develop the technology for immortality, it wouldn't be priceless; it would become worthless."

But what if we artificially restrict distribution through patents and regulations? "Patents are essentially a reflection of the will of the majority. Imagine there's an elixir of immortality. How would you prevent its widespread use? It's ultimately a formula. Perhaps you can keep it secret for a month. Then, a scientist in a lab leaks the formula, and all of humanity will have it."

This is the fundamental property of reproducible technology: zero marginal cost. Once invented and the formula made public, it can be replicated indefinitely.

"Once this is understood, the scenario of technology being permanently monopolized becomes highly unrealistic. More likely, any such groundbreaking invention will rapidly benefit all of humanity, its spread unstoppable."

Patents, regulations, and controls—these are all temporary social constructs, utterly vulnerable to the natural law of replicability. Information's yearning for freedom is not merely a slogan; its physical nature dictates that long-term censorship is unfeasible.

Depicting a dystopian world where the elite enjoy immortality while the masses are excluded is a fantasy born from a failure to understand the nature of technology.

Chapter Six: Code is Law: A Fundamental Revolution About Freedom and Control

The Road to Decentralization

I suggested to the two brothers that even if we acknowledge that affluence is a trend, control is futile, and elites cannot monopolize, decentralization in some areas still seems to have failed to break through—such as the "bipolar world order" in international politics.

"Hasn't the decentralization of power occurred?" Daniel countered. "The existence of nearly 200 countries on Earth is itself a manifestation of the dispersion of power."

Indeed, the number of independent nations has increased since World War II. "Decentralization exists, but it's far from sufficient," David acknowledges. "We haven't yet invented a social structure that can take this decentralization a step further."

Corporate promises are often unreliable. Google once promised that its search pages would be completely ad-free, Telegram guaranteed to be permanently free, ad-free, and with no subscription fees, and Facebook also claimed that its news feed would be ad-free.

"The only way to make such promises credible is to write them into the underlying protocol code," David explained.

For example, Ethereum recently upgraded its fee mechanism. Previously, miners could manipulate transaction fees, much like raising taxi fares when supply and demand were tight. Now, the protocol automatically prices fees according to a preset formula. "As a result, transaction fees have dropped significantly," David said.

The protocol doesn't lie, nor does it alter the rules afterward. "The total supply of Bitcoin will never exceed the limit set by its protocol," Daniel added.

But how is this different from a democracy? "Democracy never serves the whole; it only serves the majority. And the majority itself is fluid," Daniel replied.

In a democracy, 51% of the vote can make decisions that harm the interests of the 49% minority. Protocol governance is different: the minority always retains the right to "exit"—that is, to fork. "In a proof-of-stake mechanism, that 49% minority can leave with their assets and computing power to create a new chain—that's a fork."

This is also a loss for the remaining 51%, as the network's value is diminished by the split. Therefore, successful protocols typically require 90% or even 99% consensus, rather than a simple majority.

"But in nation-states with physical borders, you don't have that choice," David points out. Daniel adds, "Almost all successful online projects are products of mods or forks. Dota 2 originated from a Warcraft 3 mod, Counter-Strike was derived from Half-Life, Fortnite borrowed from PlayerUnknown's Battlegrounds, and the latter was originally a mod for Arma's battle royale mode."

The right to exit, or the ability to fork, is a fundamental characteristic of a healthy ecosystem.

Decentralized Practices: US Police and Gun Control as an Example

Extending decentralization from the virtual world to the real world requires innovative economic incentives: token mechanisms. "To ensure that not only server software is open source, but also the server hardware itself achieves open participation, innovation in token economics is crucial," David explained.

But how exactly does a decentralized society operate? Daniel used the American police system as an example to illustrate this.

The United States has an average of about three law enforcement officers per 1,000 people. How can this be decentralized? Candidates register through a blockchain platform, take a qualification exam, and record a self-introduction video. Each citizen votes to select three officers.

"But only a candidate who receives at least 999 votes besides yours can be elected. What kind of authorization do they receive? Essentially, you've given that person the right to bear and use weapons."

Their salaries are paid uniformly by taxes. There is no traditional hierarchical structure among them. "This is an embodiment of direct democracy: I selected these three people and gave them my quota of trust. They then negotiated among themselves—forming a police department, electing a leader, and using various applications to optimize collaboration and operational efficiency."

"The existing system of government will inevitably resist such change. But the problem is that the legitimacy of modern government is based on the principle of majority rule," David added. The core issue is when the majority will determine that a lack of central authority is more beneficial to them.

However, David points out a social paradox: "Most opinion polls show that most people do not actually agree with the so-called 'majority' position on most issues." The majority opinion is often a poorly pieced together fragmented viewpoints. "The majority choice often fails to truly reflect the demands of many individuals."

Take, for example, the gun control controversy in the United States. On the surface, society appears to be divided into opposing camps, seemingly irreconcilable. However, a closer examination of geographical distribution reveals that urban residents are more inclined to control guns because of the dense population, the potential for significant harm from a single gun owner, and the effectiveness of police presence in deterring crime.

In rural areas like Texas, people value their right to bear arms even more. This is because if danger arises (such as a wildlife attack), police may not arrive for hours, by which time it may be too late.

In other words, both sides have their own realistic rationale. "Population density largely determines people's attitudes toward the right to bear arms," Daniel concluded.

A well-designed decentralized system can accommodate such regional differences. A simple majority vote system, on the other hand, imposes a "one-size-fits-all" solution.

The lower level is powerless, and the upper level is unwilling.

The brothers believe the transition to decentralization will be a bottom-up process, achieved gradually through solving real-world problems and social experiments. They discovered a policy loophole in Europe: many countries allow taxpayers to allocate 3% of their income tax to designated non-profit organizations. This has led to the emergence of a number of "quasi-national" entities.

The "Network Nation" conference, hosted by Balaj Srinivasan, explored future decentralized governance models. "You can't imagine the enthusiasm of the participants. The venue was packed with thousands, and just as many were networking outside," David described his observation. Representatives from governments and foundations around the world also participated. This was no longer a fringe idea.

This reminds me of the microstate of Liberland, which I visited this summer. It's a disputed territory on the border between Croatia and Serbia. They've exploited the undefined border between the two countries to occupy an area roughly the size of several football fields, but ten years later, it still hasn't gained widespread recognition. Their strategy is to hope that sooner or later they'll receive recognition or special status from another country.

I used Liberland as an example to illustrate the difficulty of this path, but Daniel pointed out that success stories abound: "The Bitcoin community never lobbied any government. It was simply that tens of millions of users adopted it spontaneously."

You can't build a decentralized system by begging centralized power. "But we believe that making it possible by eventually resorting to a referendum is a viable path," David countered.

But if the referendum result can be overturned by a subsequent referendum, how can decentralization be guaranteed?

"You pass legislation through a referendum. If you follow all the procedures and achieve decentralized decision-making through a referendum, you effectively eliminate the possibility that someone might pull you back to the old system through another referendum in the future," Daniel explained.

"Once you amend the constitution, you also change the process of amending it," David added. Will opponents resort to force? Possibly. But technological advancements will make violent repression extremely costly: "A new social structure will be created that people genuinely like because they are indeed wealthier under the new system. Then the system won't regress because you can no longer mobilize the majority to overthrow it—they will oppose it."

This doesn't require a violent revolution. "People will vote for it simply because the new world is better. But that's contingent on the new world genuinely making the vast majority of people wealthier."

Lessons from OpenAI and "Public Welfare Companies"

The Lieberman brothers have a history with OpenAI. When the company was looking for a way to attract investment while staying true to its core values, the Lieberman brothers proposed a unique corporate governance structure.

In short, the core idea is that when raising funds for development, investors do not acquire equity in the entire company, but only the right to the profits from the company's revenue (such as subscription service sales). Investors do not own shares in the parent company (which holds the most valuable assets such as core AI models and LLMs), nor can they control these assets.

In addition, there is a cap on how much investors can profit from commercial entities (e.g., 100 times their investment in Microsoft): any amount exceeding this cap goes to OpenAI’s nonprofit parent company.

This plan aims to balance investor returns with the company's commitment to serving society, ensuring that the development of the underlying model benefits a wider range of public interests.

"The idea was brilliant and worked well for a time. They set a cap on investor returns, and it was this model that successfully attracted a diverse pool of talent," David recalled.

During the model's development phase, the brothers engaged in in-depth discussions with founders Greg Brockman, Ilya Sutskøver, and Sam Altman. At that time, they raised a crucial question: "Okay, you limit investor returns, with excess profits going to the nonprofit organization. So, who manages this nonprofit organization?"

"In my view, the core of the controversy surrounding Sam Altman's brief dismissal and subsequent reinstatement by the board revolved around who should control the nonprofit organization. The conflict arose once it became clear that this nonprofit would become one of the wealthiest entities in the world," David analyzed.

The brothers initially proposed a different solution: "Excess profits should not flow elsewhere, but should be used to lower product prices so that the public can benefit." However, OpenAI ultimately chose a different path. On October 28, 2025, with approval from US authorities, OpenAI completed its transformation into a "public welfare company." The company did what the brothers had warned against: it removed the cap on investor returns—a key element of its original structure. OpenAI now employs a standard capital structure and business model: shareholders hold shares and receive unlimited profits proportionally.

The OpenAI Foundation (formerly a nonprofit organization) formally controls the new commercial entity, OpenAI Group PBC, holding a 26% stake, currently valued at approximately $130 billion. Microsoft holds 27%, valued at approximately $135 billion. The foundation has pledged $25 billion for philanthropy, but has not set a timetable.

The problems the brothers warned of have become a reality: the foundation's shareholding is fixed, and its equity will be diluted with subsequent rounds of financing. Sam Altman, as CEO, effectively controls this $500 billion company with a standard business structure. Formally, the foundation appoints a board of directors, but in reality—if the board is loyal to Altman—the foundation's control over the parent company is merely nominal.

This is exactly the scenario that unfolded during Altman's dismissal in November 2023: the board attempted to remove him, but core employees protested collectively, and Altman ultimately returned with greater power.

Chapter Seven: Becoming a Creator in the Tide of AI

AI natives: Blessed by nature or born at the wrong time?

In the United States, pension funds highly concentrate social wealth. The global population structure is aging: the trend of declining birth rates continues, medical advancements are extending life expectancy, and the proportion of the retired population continues to increase.

Artificial intelligence is replacing junior engineers, depriving interns of opportunities to learn through practical experience. On the surface, the generational gap appears to be widening: senior engineers have accumulated experience, capital, and influence, making it increasingly difficult for younger engineers to keep up.

David disagrees. "Will AI have a negative impact on this generation? To be honest, I don't think so. Historical experience points to the opposite conclusion."

Admittedly, experienced experts in their 30s and 40s who are adept at using AI will be greatly enhanced. However, the older generation is relatively slower to adapt to AI. Looking back at the era of widespread personal computers, young people gained a huge advantage due to their natural affinity for new technologies.

"It wasn't long ago that you'd see young people in their early twenties creating billion-dollar companies... Now, thanks to AI, young people who just graduated from MIT have founded the fastest-growing company in terms of revenue—Cursor (an AI code assistant)," he cited as an example. Daniel added another case: "The team collaboration tool Lovable was also developed by a group of young people in their early twenties, fresh out of college."

Why didn't experienced professionals with university degrees create Cursor? "They tried, but often didn't get it right," David replied. Young people have a different perspective. They interact with AI most frequently and understand better how to extract value from it.

This generation has accepted the reality more quickly that "videos should not be trusted" and "photographs are not created for documentary purposes but as a medium of communication."

Today's 10- to 13-year-olds have even created internet memes like "brainrot" to satirize all forms of traditional media, from news to meticulously crafted films. For them, these are merely "information snacks" that don't require serious attention, and consuming them won't change their lives. "This is how they digest and understand the new normal of the world," Daniel explains.

"You know, just a hundred years ago, the average life expectancy was less than 35 years," Daniel chuckled at the感慨 about the dramatic changes of the times. The technological revolution is reshaping everything: access to information, lifestyles, health concepts, and ways of working.

"I'm extremely reluctant to blame the younger generation. I'm convinced they understand AI better than we do," David said. "They're AI natives, mobile digital natives, and 3D interaction natives. Seeing them build and shoot simultaneously in Fortnite, I simply can't comprehend how they do it. I have immense respect for this generation."

Senior researchers build foundational models, but they may not know how to create successful end products based on these models.

"Just like the people who originally wrote the protocols for the internet, they may not have figured out what applications should be built on top of it. Where are our AI-native social networks? Where are they?" Daniel asked.

Bill Peebles, in his early twenties, is a key contributor to the video generation model Sora: an extremely young developer who has spearheaded some of the first truly AI innovations in the social media space. "But shouldn't there be dozens of these AI-first social networks by now? And all the previous generations were completely clueless about it," David added.

Nevertheless, David acknowledged, "I have high hopes for this generation, but they face enormous challenges because the entire social structure was not designed for them."

"What's even more serious is that the younger generation has less and less leverage to influence society... but their knowledge and understanding actually far surpass those of their predecessors. This sense of dislocation can breed intense injustice," Daniel added.

Young people know more about AI, understand it more deeply, and see further ahead because they will be living in it. But they have access to fewer social resources.

The "Dead Internet" Theory and Countermeasures

Sora is an excellent example of a native AI product. However, all content platforms are already flooded with AI-generated videos without watermarks. The older the user, the harder it is to distinguish between genuine and fake videos.

Complaints are rampant: spam is everywhere, and "I don't want to live in a 'dead internet' (referring to a network where content is largely generated by AI and lacks genuine human interaction)." A new pattern is emerging: users see text with a specific layout or tone and assume it's from AI, criticizing the author for disrespecting the reader's time.

David believes the younger generation will quickly find solutions: "Adaptive behaviors will emerge; they will learn to distinguish between real and generated content. They will be the innovators, developing tools and products to solve this problem."

Or perhaps we don't need to "solve" this problem at all? Some films are shot on film, some on digital—what difference does it make to the audience?

Take Snapchat as an example. This social network defaults to short, authentic, and unadorned content. But millennials probably couldn't have invented Snapchat. They grew up in the era of film cameras, where every photo symbolized the capture of a precious moment.

The next generation viewed photography as a purely communicative tool.

"We're closely watching young people in their 20s to see what projects they're creating," David said. Young people are eager to incorporate AI into their creative work. They've already embraced AI programming much faster.

We've experienced "one-off messages." The future will see "one-off applications."

Initially, the code written by AI was of poor quality. Now, OpenAI and xAI's top engineers are pairing up with AI for coding. But this isn't simply "vibe coding": "You don't just say 'Make me an application.' In reality, you need to write extremely detailed technical specifications for every step," Daniel explains.

This is not "cue engineering," but "context engineering"—deeply processing the task context, collecting and structuring all necessary information to enable AI to perform efficiently.

Advice for the younger generation: Create by doing

If a university semester is shorter than the interval between world-shaking changes, how should one learn? If companies replace interns and junior researchers with neural networks, where should one gain experience?

"You can create them. Everyone expects a solo unicorn to emerge," David replied. A unicorn is a startup valued at one billion dollars, and there has never been a precedent for a single person achieving this feat. But Sean Kaplan, founder of the prediction market platform Polymarket, is close to that goal, becoming the world's youngest self-made billionaire at the age of 27.

"It's only a matter of time before these kinds of success stories emerge. In the early stages of our startup, product development cycles took up to four years. But with the evolution of the Web, mobile internet, and AI, the speed has increased dramatically; now, a minimum viable product (MVP) can be created in just a few days."

Currently, the speed at which apps are uploaded to app stores is more dependent on the app store's review process than on the development process itself.

"Our core advice is: you learn more by creating your own projects," David emphasized. AI has significantly lowered the capital barrier to entrepreneurship, allowing small businesses to emerge.

In other words, for most people, the only way to prove their abilities and gain experience in the future may be to create their own projects.

"That sounds a bit like a threat..." Daniel laughed.

But David was serious: "It's best to create your own projects. That's definitely the path to faster growth. Even failed projects teach you far more than those who stay in their current positions."

Young people are not burdened by heavy historical baggage or established commitments that prevent them from trying new "curves". They can take risks and explore new territories.

"The easiest option right now is to start a small consulting firm to help enterprise clients apply AI. You're young and know more about AI than most business owners," David advised.

Just as a decade ago, a generation of entrepreneurs taught businesses how to use social media, twenty-five years ago, the Lieberman Brothers' job was to connect businesses to the internet.

"We have entered an era where opening this type of business is unprecedented. The system allows you to do so. Now, what's needed is to take action."