The Barbell — Top 20%, Bottom 20%, and the Squeezed Middle

the Crushed Middle Picture a barbell at the gym. Two heavy weights on the ends, connected by a thin bar in the middle. That is the economic shape AI is creating - and almost nobody sees it coming. The top 20 percent will thrive. The bottom 20 percent will actually be lifted up. And the middle 60 percent - the vast majority of people in developed nations - will get crushed during the transition. Not metaphorically crushed. Economically, psychologically, and socially squeezed in ways we have not seen since the Great Depression. It will be one of the biggest and fastest disruptions in the history of civilization. Every conversation about AI and jobs misses this. Optimists say new jobs will emerge for everyone, just like after every technological revolution. Pessimists say AI is coming for everything and we are all doomed. Both sides are looking at the wrong variable. The question is not whether AI creates winners and losers. Every technology does. The question is who specifically wins, who specifically loses, why the middle will bear the burden, and what everyone can do to prevent themselves from ending up on the wrong side.

The reality is that there’s a winnable path for everyone, but depending on

where you are today, your path will vary from extremely difficult, to extremely easy. I will explain to you exactly why.

The Shape of What Is Coming

On one end of the barbell, you have roughly the top 20 percent of the

socioeconomic ladder. Capital owners. Builders. Risk-takers. The people who already have resources and the knowledge to deploy them. AI is going to be absolutely incredible for these people. Not just good - transformational in ways most of them do not even fully understand yet. Especially if these folks have a natural inclination to “just do things”. On the other end, you have the bottom 20 percent. Depending on where you look in the world, these vary from people living in regions without basic infrastructure, clean water, stable food supplies, adequate healthcare, to people who are dependent on the welfare system to survive. I know this sounds counterintuitive, but AI combined with robotics and cheap energy is going to be phenomenal for them too. We are talking about finally delivering abundance at scale to populations that have been waiting generations for it that haven’t had the ability to join the other 80% of society. And then there is the middle. The 60 percent in between. The people who make up the vast majority of developed nations' populations. The people who work jobs, pay mortgages, raise families, and assume that working hard for someone else will continue to provide a reasonable life. This group - the 60 percent in between - is where it’s going to get ugly, and fast, unless drastic action is taken.

Why the Top 20 Percent Thrives

If you already own capital - whether that is money, equity, real estate, or

intellectual property - you have a unique advantage in the AI era. You can deploy that capital to purchase, control, direct, and use AI systems to create

new businesses, deploy ideas at scale, or build things that previously required hundreds of employees. The more capital you have, the easier it is to employ AIs to do your bidding. Today if I want to start a company, I need to hire people. Lawyers, accountants, marketers, developers, customer service reps. Each person costs money, takes time to find, has varying skill levels, needs management, can quit, gets sick, and has limited working hours. Now imagine I can spin up AI agents that handle legal document review, accounting reconciliation, marketing campaign creation, code development, and customer interactions - all for a fraction of the cost, at any hour, with consistent quality, and infinite patience. As I’m doing my final proofread, we are seeing a massive drawdown in the stock market (February 5th 2026) because the white collar class on LinkedIN has woken up to Claude Cowork, which is an agentic AI tool that can do 90%+ of what your typical SaaS products can do, but at a fraction of a fraction of the cost. This is already happening. The companies deploying AI agents today are seeing productivity multipliers of 5x, 10x, sometimes more in specific domains. If you work with AI on the daily, you already know this. This isn’t news to you. It’s beyond transformational. So if you already have capital - even modest amounts - you can put it to work in ways that were impossible just two years ago. Your money becomes more productive. Your ideas become more executable. Your competitive advantage compounds. All you have to do is “hire” these AI agents. And if you are a builder - someone who creates things, takes risks, experiments - AI becomes the ultimate force multiplier. You can prototype in hours what used to take months. You can test ideas at scale without massive upfront investment. You can compete with much larger organizations because the playing field is flattening in certain dimensions.

The curious, the risk-tolerant, the capital-endowed - these people are going to

be incredibly well off.

The Self-Reinforcing Advantage

There is another dimension to this that makes the top 20 percent advantage

even more pronounced. When you deploy AI successfully, you learn how to deploy it better. Your understanding deepens. You discover new applications. You develop intuitions about what works and what does not. And this knowledge compounds. A founder who has been using AI agents for two years has a massive advantage over someone just starting. They know which tools work, which prompts get results, which workflows to automate and which to leave manual. They have already made the expensive mistakes and learned from them. This is precisely why I’ve been spending way too much time using OpenClaw as of the last few days, which is an open source AI agent that lives in a computer that can do literally anything on it. Think of it as a digital human. It’s not perfect, but it’s 80%+ of the way there. I have my entire business life plugged into it - emails, financials, chats, discord servers, YouTube backend, X backend. Now this agent can execute actions in seconds that would take me much longer. “Hey Claw (that’s the agent’s name), so can you check and see which videos are underperforming on the channel from the last couple weeks, and see what kind of titles we can switch them out to that align with the latest news cycle? Can you also look for ways to improve SEO on the description and tags to capitalize on this? Oh and can you make this into a repeatable job every couple days so we can track previous performance and find new opportunities?”

Done. That job alone will easily increase my views by 10% per year. For a

channel that has earned over a million dollars in the last 4 years, that’s $100,000 that took about 30 seconds to execute. And that’s one use case out of probably… thousands? Millions? And it will simply get better with better models. The entire history of our chats is saved. All the jobs. All the learnings. All the files. All the connections. Everything. All I have to do is upgrade its brain every time a new model comes out. So the advantage goes beyond capital. It includes accumulated knowledge about how to leverage these systems effectively. And that knowledge is not equally distributed. The people already at the top are experimenting more, learning faster, and pulling further ahead. Meanwhile, the people in the middle are often still debating whether AI is actually useful or just hype. This knowledge gap is widening every month. And unlike financial capital, you cannot simply redistribute it. You have to develop it through experience. Which takes time. Which the middle does not have much of, because most are living paycheck to paycheck trying to live decent lives, but don’t have time for much else. You’ll quickly see that the ones that will win most in the AI age, are those with the most amount of time to spend with AI.

Why the Bottom 20 Percent Also Wins

The bottom 20 percent of the global population lives in conditions that

people in developed nations cannot really fathom. No reliable clean water. Food insecurity. Minimal healthcare. Housing that would be condemned in any Western city. Energy poverty that limits everything from cooking to education to economic activity. These are problems of delivery cost, pure and simple.

We know how to desalinate water. The technology exists. We know how to

produce housing at scale. We know how to manufacture basic medicines. We know how to generate power. The reason billions of people still lack these basics is that the cost to deliver them in remote or impoverished regions has been prohibitively high. Human labor in these contexts is the bottleneck. You need doctors who are willing to work in rural Uganda. You need construction crews who can build in remote Indonesian islands. You need engineers who will maintain water systems in villages across Bangladesh. And the economics rarely work because human labor is expensive to deploy where humans do not want to live. Now think about what happens when robotic labor costs $3-5 per hour instead of human wages, or less. When robots can be deployed anywhere without needing housing, healthcare, schools for their children, or cultural amenities. When AI systems can diagnose diseases from symptoms described in any language and transmitted through a basic smartphone. Desalination at scale becomes economically viable for coastal communities everywhere. Automated construction can build adequate housing in regions that cannot attract skilled builders. AI-powered telemedicine can provide diagnostics that would cost hundreds of dollars in a Western clinic for essentially nothing. All core human needs - water, food, shelter, basic healthcare, energy - become addressable at costs that were previously unimaginable. And beyond necessities, entertainment and education that used to be privileges of wealthy nations become universally accessible. The bottom 20 percent is going to experience abundant, life-changing technology that actually digs them out of poverty. Technology that simply makes it cheap to deliver what they need - something charity and government programs have never accomplished at scale. This is genuinely transformational. This is what I mean when I talk about the age of abundance becoming real for people who have never experienced it.

The Leapfrogging Effect

Something interesting happens when technology bypasses infrastructure

problems entirely. Africa largely skipped landline telephone systems and went straight to mobile phones. Many developing regions are skipping traditional banking and adopting mobile payment systems instead, or cryptocurrency like Bitcoin. The infrastructure that developed nations spent decades building is being leapfrogged by newer, cheaper alternatives. AI and robotics enable the same pattern across many more domains. You do not need to build a network of hospitals staffed by doctors to provide basic healthcare. You need smartphones with AI diagnostic apps and drones that deliver medications. The expensive legacy infrastructure that wealthy nations built over centuries becomes unnecessary. The bottom 20 percent does not need to follow the same development path the West did. They can skip directly to solutions that were not possible even five years ago. Solar panels with battery storage provide electricity without building power plants and transmission lines. AI tutoring systems provide education without building schools and training teachers. Automated farming systems produce food without the agricultural infrastructure wealthy nations developed over generations. The cost of meeting basic human needs is collapsing. And the people who benefit most from that collapse are the ones who currently lack those basics. This is why I remain optimistic about the bottom of the barbell even as I worry about the middle. The technology genuinely solves their problems in ways that previous development approaches could not.

Why the Middle 60 Percent Gets Crushed

So we have the top thriving and the bottom being lifted up. What happens in

the middle? The middle is where most people reading this book probably sit. You have a job. Maybe you have a college degree or specialized training. You work hard. You trade your time and skills for money, then use that money to buy the things you need. You are not wealthy, but you are not in poverty either. You are comfortable enough, or at least getting by. And AI is coming for you specifically. I am not trying to be alarmist. I am trying to be honest. The middle 60 percent represents people who spend most of their time building on behalf of others to earn a living wage. You write reports, analyze data, process transactions, design presentations, manage projects, answer customer questions, review documents, create content, handle logistics, work at a factory, warehouse, do construction - the list goes on. Every single one of those activities is being automated right now. Not in ten years. Not after some mythical AGI moment. Right now. The companies deploying AI are discovering that huge chunks of what they pay middle-management and professional salaries for can be done faster, cheaper, and often better by AI systems. And unlike previous technological disruptions, this one does not primarily affect blue-collar workers doing physical labor - at least at first. It hits knowledge workers. White-collar professionals. The college-educated class that thought their degrees and expertise would protect them. The accountant reviewing financial statements. The lawyer doing document review. The marketer writing copy. The analyst building spreadsheets. The project manager coordinating tasks. The recruiter screening resumes. The customer service rep answering questions. These are not unskilled jobs. These are jobs that require education, experience, and expertise. And AI systems are becoming capable of

performing them at 10x speed for a fraction of the cost. This is because AI is extremely good at one thing - Intelligence. And intelligence is at the root of every action a human makes, especially those that are high quality, repetitive, and valuable.

The Identity Crisis

For the middle class in developed nations like the USA, work is a source of

identity just as much as income. When someone asks who you are at a dinner party, you tell them what you do for a living. Your profession defines your social status, your sense of contribution, your understanding of your own value.

What happens when that work is suddenly unnecessary?

A lawyer who spent seven years in school and another ten building a career

does not just lose a job when AI handles legal document review. They lose an identity. A marketer who takes pride in their creative campaigns faces an existential question when AI generates better copy faster. An analyst who believed their value came from deep expertise confronts the reality that expertise is now a commodity. I'm aware that telling a 45-year-old lawyer to reinvent themselves is easy for me to say. I'm sitting here writing about a transition that will likely devastate many people who did nothing wrong - people who followed every rule, made every responsible choice, built careers exactly the way society told them to. The psychological weight of this transition is real, and I don't want to minimize it. Some of this will be brutally unfair to people who did everything right by the old rules. A generation was promised that education plus hard work plus showing up every day equaled security. That promise is breaking, and the people who kept their end of the bargain are the ones paying the price.

The cruelty is that preparation doesn't guarantee survival. You can see this

coming, understand it completely, take every reasonable step - and still get crushed by forces outside your control. I'm not writing this chapter to pretend otherwise. Some of what's coming cannot be fully prepared for. Some of the disruption will simply have to be endured. That said, I refuse to stop at acknowledgment. Compassion without action is just spectatorship. Understanding the pain doesn't help if we don't also talk about what can actually be done. This is not something you can solve with retraining programs. The psychological disruption in the middle class will be severe. Depression. Anxiety. Loss of meaning. We are already seeing this among knowledge workers who have been displaced, and it is only the beginning. Financial stress is one thing. But losing your sense of who you are and what you contribute to the world is something else entirely. And the middle 60 percent is about to face both simultaneously.

The Transition Problem

Now, some people will say what they always say: new jobs will emerge. Technology creates more jobs than it destroys. We have heard this argument after every major technological shift. And they are not wrong. Eventually. In the long run. New types of jobs will exist that we cannot even imagine today. But the transition period is going to be brutal. When manufacturing automation hit the Midwest in the 1980s and 1990s, entire communities were devastated. It took a generation for those regions to partially recover - and many still have not. And that disruption affected a relatively small segment of the population: factory workers with specific skills in specific locations.

What happens when the disruption affects a significant chunk of 60 percent

of the workforce almost simultaneously? We are talking about accountants in every city. Lawyers in every state. Marketers in every industry. Analysts in every company. Customer service reps in every sector. All facing the same pressure at the same time. The problem is not that there will not eventually be new jobs. The problem is the timeline mismatch. It takes years to retrain a workforce. It takes years to reform an education system. AI capability is improving on a curve measured in months. A 45-year-old middle manager with a mortgage, two kids in school, and parents approaching retirement age cannot simply pivot to a new career in AI-adjacent fields. They do not have time to go back to school. They cannot accept an entry-level salary while learning new skills. Their financial obligations are fixed even as their income becomes uncertain. This is what I mean by crushed. Not that these people will starve - at least not initially. But that they will face economic pressure, status loss, identity disruption, and financial stress at a scale we have not seen since the Great Depression. And without massive, competent government intervention, this leads to collapse. Not metaphorical collapse. Actual societal collapse - institutions failing, social contracts breaking, political extremism rising to fill the vacuum. History shows what happens when large populations face rapid economic displacement without adequate support systems. It's not pretty. The barbell can deliver abundance to both ends while the middle falls apart, and if that happens, the whole thing tears apart.

The Government Question

This is where my position diverges from both the tech optimists and the

traditional pessimists.

I am not against technology. I am extraordinarily bullish on what AI, robotics, and abundant energy will accomplish. The age of abundance is real and achievable. But getting there requires herculean government execution. And my confidence in government's ability to execute is very low. I will go deep on what government should do - and why I doubt they will deliver - in Chapter 8. For now, just understand the stakes: the middle 60 percent depends on institutional responses arriving in time. And institutions move slowly. What I see is a scenario where the technology advances faster than governance adapts. Where the top 20 percent accumulate more because they can deploy AI immediately while policy debates drag on for years. Where the bottom 20 percent benefits because their needs are simple enough that technology solves them directly. And where the middle 60 percent gets squeezed because they depend on institutional responses that do not arrive in time to help them transition.

What Makes This Different

People keep comparing AI to previous technologies. The printing press. Electricity. The internet. And in some ways, these comparisons are valid. Every major technology created winners and losers. But there is something different this time, and it is important to understand why the middle specifically gets hit. Previous technologies primarily replaced physical labor or augmented human capability in specific domains. The printing press did not replace storytellers - it amplified them. Electricity did not replace human decision-making - it powered tools that humans operated. Even early computers required human programmers, operators, and analysts to be useful. AI is different because it replaces judgment. It replaces analysis. It replaces the cognitive work that educated professionals have been trained to perform.

A factory worker displaced by automation could theoretically move into an

office job. The work was different enough that human comparative advantage remained clear. A knowledge worker displaced by AI faces a harder question: What can I do that the AI cannot? And for many tasks, the honest answer is becoming nothing - or at least nothing at a cost the market will pay. This is why the barbell forms. At the top, you have people who can own and direct AI systems - they thrive because they are not competing with AI, they are leveraging it. At the bottom, you have people whose needs can be met through automated delivery of basic goods and services - they benefit because their problems become solvable. And they never experience a quality of life that the 80% did, so any improvement will feel like a blessing. In the middle, you have people whose primary economic value was performing cognitive & physical tasks for wages that would often times come from the top 20%. And that value proposition is collapsing.

The Aggregation Problem

There is another dynamic that accelerates the barbell effect. In a world where AI amplifies capability, success aggregates even more than it already does. If one company can use AI to be 10x more productive than competitors, that company does not just win a little more market share. It potentially wins everything. Look at how technology markets have already played out. Google did not just win search - it dominates search. Meta did not just win social - it controls most social media. Amazon did not just win e-commerce - it is the default destination for online shopping. AI intensifies this winner-take-all dynamic across every industry. If an AI-augmented law firm can handle cases at one-tenth the cost of traditional competitors, why would clients go anywhere else? If an AI- augmented accounting firm can process returns faster and more accurately at

lower prices, why would anyone choose the expensive human-heavy alternative? The companies that deploy AI effectively will capture entire markets. The companies that do not will disappear. And the employees of those disappearing companies will be looking for jobs in markets with fewer and fewer employers, unless these new AI companies require more humans to help them operate. But what’s the likelihood of that happening when AI can (and will) do everything a human can do better, smarter, faster, and cheaper? Individual job displacement is only part of the story. The bigger issue is market structure. The number of employers in many industries will shrink dramatically because AI rewards scale and punishes companies that cannot leverage it. Fewer employers means less competition for labor means lower wages for whatever jobs remain. Another mechanism that squeezes the middle.

China Will Have an Easier Time

One prediction I will make that people find uncomfortable: China will

navigate this transition more smoothly than the West. Not because their technology is better. Not because their government is more competent in some absolute sense. But because Chinese society is already structured for collective adaptation in ways that Western societies are not. In China, the government can mandate retraining programs and people participate. They can redirect labor to new sectors and workers comply. The social expectation is collective sacrifice for collective advancement. Individual disruption is expected to be absorbed by extended family networks and community structures. In the West, we have optimized for individual autonomy. Which is wonderful for innovation and personal freedom. But it also means that when economic

disruption hits, the expectation is that individuals figure it out themselves. We do not have the social infrastructure for collective adaptation. A displaced knowledge worker in Shanghai has a multigenerational household to absorb the shock, a government that will aggressively manage the transition, and a culture that accepts temporary personal sacrifice for national advancement. A displaced knowledge worker in San Francisco has a nuclear family they are solely responsible for, a government deadlocked in partisan warfare, and a culture that blames them for not adapting faster. Of course I’m generalizing here, but you get the point. I am not saying China's system is better. I am saying it seems better suited to manage this specific type of transition. And that should concern Americans who assume our system's superiority will protect us.

The Timeline Nobody Wants to Hear

So what is the timeline we are talking about? When does this barbell fully

form? It is already forming. The pressure on knowledge workers is visible today in layoffs across tech, media, finance, and professional services. Companies are discovering they can operate with fewer employees by deploying AI tools. Within the next 2-5 years, I expect the pressure to become acute. Not universal - some roles will remain protected longer than others. But enough displacement that it becomes a major political and social issue. By 2030, I think we will have clarity on whether governments managed the transition well or failed. Either new support structures will be in place and functioning, or we will be in the middle of social unrest that makes recent political polarization look mild. The barbell is forming - that part is already happening. What remains uncertain is whether we reach a new equilibrium where all three groups find sustainable positions, or whether the middle collapses and takes civilization

with it. This is the mechanism that determines abundance or collapse. The technology is neutral - it can deliver either outcome. The barbell is the lens through which to see which way we're heading.

What Should You Do With This Information

I do not write this to be pessimistic. I write it because understanding the

shape of what is coming is the first step toward navigating it. If you are in the top 20 percent - or have a path to get there - you need to be deploying capital and AI systems aggressively. The compound advantage is real and growing. Do not wait. If you are in the middle 60 percent, you need to be honest with yourself about your vulnerability. Is your job fundamentally about cognitive tasks that AI will soon perform cheaper and faster? If yes, you have 2-5 years to reposition, fully embrace AI, or both. Not 10. Not 20. Two to five. What does repositioning mean? It means moving toward capital ownership, even in small amounts. It means developing skills in areas where human judgment and physical presence remain essential. It means thinking like an entrepreneur even if you stay employed. It means understanding that the social contract that promised stability in exchange for showing up and working hard is breaking. And it means pressuring your government to actually prepare for this. Not partisan culture war nonsense. Real policy about retraining, safety nets, education reform, and transition management. Vote for competence over ideology. Demand actual plans, not slogans. The age of abundance is coming. I genuinely believe that. On the other side of this disruption is a world where most human needs are met cheaply and easily. Where people have more freedom to pursue what they find meaningful. Where the global poor finally share in prosperity. But the barbell means we do not all get there at the same time, through the same path, with the same experience.

Some will ride it. Some will be lifted by it. And some will be crushed under its weight - at least during the transition. Which group you end up in depends largely on decisions you make now, with the information you have now, while there is still time to position yourself. That is the uncomfortable truth about what's coming. And I would rather tell you that truth than pretend we are all going to be fine.