Your Personal Roadmap

I've spent the last eleven chapters explaining what I think is coming. The

convergence of AI, robotics, and energy. The death of legacy industries. The barbell that will reward the top and bottom while crushing the middle. The data moats that will determine AI winners. The investment thesis that can protect those with capital. The transition that most people haven't yet recognized as urgent. Now comes the hard part: What do you actually do about it? I'm not here to give you generic advice about "learning to code" or "staying curious" or whatever platitudes usually fill this kind of chapter. That stuff is largely useless. What I'm going to lay out is specific, practical, and uncomfortable. Some of it will apply directly to your situation. Some of it won't. But all of it comes from the same framework I use for my own life and investments. This transformation is happening whether you're prepared or not. You cannot avoid it. The only question is whether you position yourself for abundance or get swept away by collapse.

Starting With Brutal Honesty

Before we get into specifics, you need to answer one question honestly: Where do you actually sit on the socioeconomic ladder? Not where you think you should be. Not where you were five years ago. Not where you hope to be in five years. Where are you right now, in terms of capital ownership, skill set, and positioning relative to AI disruption? I've talked a lot about the top 20%, middle 60%, and bottom 20%. Most people reading this book probably fall into the middle 60%. I mean that descriptively, not as an insult - just statistics. If that's you, then some of what I'm about to say will be uncomfortable. But discomfort now beats devastation later. If you're in the top 20% - meaning you have meaningful capital to deploy, ownership stakes in businesses, or the ability to build at scale - then your roadmap looks different. You have more options, but also more responsibility to use those options wisely. And if you’re in the bottom 20% - get ready, because your life is about to transform in ways you could never have imagined, for the better.

What It Means to Be in the Top 20%

When I say top 20%, I'm not just talking about income. A doctor making

$400K per year who spends it all is not in the top 20% for purposes of this

discussion. Income is not the same as capital.

The top 20% means: - You have deployable capital that can work for you while you sleep - You have ownership stakes in businesses or investments - You have the ability to hire talent or deploy AI systems to multiply your output

- Your income is not purely dependent on trading hours for dollars

If you're not sure whether you qualify, you probably don't. And that's fine -

the roadmap for the middle 60% is where most people need to focus anyway. The Top 20% Roadmap: Deploy, Don't Just Consume If you do have capital and positioning, your biggest risk is complacency. You might think you're set. You might think your current advantages will persist. They won't - not unless you actively deploy them. The AI revolution is going to be incredible for capital owners. But only for those who actually use their capital to participate in the transformation, not just preserve what they already have.

Deploy Capital Into AI-First Companies

I've laid out my framework in previous chapters. Companies that meet the

four criteria: misunderstood by the market, disrupting large legacy industries, rockstar leadership, and 10x potential over five to ten years. The specifics will change over time. Tesla was the obvious answer in 2012. It might still be the answer now given Robotaxi and Optimus. Lemonade looks compelling to me for insurance disruption. You might find other opportunities that fit the framework. But the key is deployment. Capital sitting in index funds is not going to capture the AI opportunity. Index funds are designed to capture average returns across the entire market - including all the legacy companies about to get disrupted. When Blockbuster got destroyed, index fund holders owned Blockbuster. When Kodak collapsed, index fund holders owned Kodak. The same will happen to legacy auto, legacy finance, legacy healthcare, legacy legal, legacy everything. Your diversified portfolio owns all of that. I'm not saying put everything into one stock. That's my approach, but I've been clear it's high-risk and requires constant thesis monitoring. What I am saying is that passive diversification is a strategy for capturing average returns

while owning a lot of future losers. You will not capture transformational returns that way.

Concentrate When Conviction Is High

I've made this case already, but it bears repeating for the top 20%. Diversification is a hedge against ignorance. If you've done the work, if you actually understand a business and its trajectory better than the market does, then diversification destroys returns. This is uncomfortable for most people. We're trained to spread risk. But spreading risk also spreads return. If you want outsized outcomes, you need outsized positions. The caveat - and I can't stress this enough - is that concentration only makes sense when you've genuinely done the work. If you're concentrating based on FOMO or hot tips or social media hype, you're gambling, not investing. Concentration without conviction is just recklessness.

Build, Don't Just Consume

The final piece for the top 20% is the hardest. Capital and investments are

important, but they're ultimately passive. The real opportunity is in building. AI is a 10x multiplier for builders. Maybe 100x. Someone who would have built a $10M company can now build a $100M company with the same effort. Someone who would have needed a team of twenty can now execute with a team of five plus AI agents. If you have the capital to fund your own projects, this is the time to do it. If you have the positioning to take entrepreneurial risk, this is the moment. The window won't last forever - eventually AI capabilities will commoditize and the arbitrage opportunity closes.

Right now, knowing how to deploy AI effectively is rare. In five years, it will

be table stakes. The builders who move now capture the value. The builders who wait compete in a crowded field. The world needs your creativity. Your innovative spirit. Your desire to make the world a better place. To help others. No more waiting for others to do it on your behalf. It’s now your turn. You have the ultimate side kick in AI. Don’t waste the opportunity.

The Middle 60% Roadmap: Move Urgently

Now for the harder conversation. If you're in the middle 60% - trading

cognitive labor for salary, not much capital accumulation, skills primarily valuable to an employer - your situation is more urgent. I don't say that to scare you. I say it because pretending otherwise would be doing you a disservice. The middle 60% faces the most disruption over the next five years. The question is whether you're disrupted or whether you disrupt yourself first.

Urgently Develop AI-Adjacent Skills

The most valuable skill right now is knowing how to deploy AI in specific

domains. Building AI is secondary. The world is full of AI researchers and ML engineers. What's scarce is people who understand a specific industry deeply AND know how to apply AI to transform it. Are you a lawyer? The lawyers who thrive will be the ones who figure out how to use AI to do 10x the legal work, not the ones who pretend AI doesn't threaten them. Are you a marketer? The marketers who thrive will be the ones who use AI to create campaigns at unprecedented scale and personalization. Are you a software developer? The developers who thrive

will be the ones who use AI to multiply their output by an order of magnitude. Forget "learning prompt engineering" or whatever the current buzzword is. You want to become the person in your field who actually figures out how AI changes everything. That person is incredibly valuable. The person who just does the same job the same way they did five years ago? They're the person who gets replaced.

Move Toward Capital Ownership

This is the structural challenge for the middle 60%. You're trading labor for

income, but AI is about to devalue a lot of cognitive labor - and soon after, physical labor. The solution is to start accumulating capital, even in small amounts. I know this sounds like "just stop being poor" advice. It's not. The point is directional movement, not immediate transformation. If you're currently saving nothing, save something. If you're saving into index funds, consider whether concentrated positions in AI-first companies might make more sense for the portion you can afford to put at risk. If you're paying down low-interest debt when you could be investing, reconsider. Every dollar of capital you own is a dollar that can compound without your labor. Every dollar you own in AI-first companies is participating in the rebalancing on the right side. The gap between capital owners and labor sellers is about to widen dramatically. Even small movement toward capital ownership changes your trajectory. At the very least, you’ll gain momentum.

Understand the Timeline

Five years. That's the window. Maximum. I'm not saying you'll be unemployed in two years. But I am saying that the structural shifts happening now will accelerate dramatically. The skills that are

valuable today may not be valuable in 2030. The companies that are stable today may not exist in 2030. This is an urgent repositioning - you have maybe five years, not ten. The people who spend the next five years gradually adapting will find themselves adapting too slowly. The people who treat this as an emergency will be positioned when the acute phase hits. I wish I could tell you there's plenty of time. There isn't. Not for the middle 60%.

Don't Wait for Government to Save You

As I argued in Chapter 8, governments will likely fail to execute well on the

AI transition. Not because the people in government are malicious - because

the structural incentives don't support it. The disruption is happening at AI speed while policy moves at bureaucratic speed. So yes, advocate for good policy. Vote for leaders who understand the stakes. But don't make your personal plan dependent on government executing well. Plan as if you're on your own, because you probably are.

If You're Starting From Zero

Everything I've written so far assumes you have at least some capital to

deploy, some financial margin to work with, some runway to make moves. But what if you don't? What if you're reading this while worrying about making rent next month? What if "invest in AI-first companies" sounds as realistic as "buy a yacht"? I know this advice is easier to give than to follow when you're worried about rent. The privilege of having capital to invest isn't lost on me. I've had periods in my life where I was cash-strapped, and I remember what it feels like when financial advice assumes a foundation you don't have.

If you're starting from zero, your path is harder - but not impossible. It's just different. And let me be specific about what that path actually looks like.

Free Tools Are Your Starting Point

The good news about AI is that the most powerful tools in human history are

available for free. Not watered-down versions - genuinely powerful systems that can teach you, help you build skills, and multiply your output. ChatGPT's free tier gives you access to GPT-4o. Claude has a free tier. Google's Gemini has a free tier. Microsoft Copilot is free. These are not toys - they are the same fundamental technology that companies are paying millions to deploy. Start using them. Every day. Not casually - systematically. Ask them to explain concepts in your field or area of interest. Use them to draft documents, analyze problems, and learn new skills. Treat them as a personal tutor, research assistant, and thinking partner all in one. The person who spends two hours a day learning with AI for a year will be dramatically more capable than someone who ignored these tools. That capability gap translates to economic value eventually.

You can develop these skills without spending a dollar

Prompt engineering. The skill of getting AI to do what you actually want. This sounds trivial, but it's not. The difference between a mediocre prompt and a great prompt is the difference between useless output and genuine value. You can learn this by experimenting, reading free guides, watching YouTube tutorials, and practicing. The best advice I can give: stop thinking that you’re talking to a machine. Start thinking like you’re talking to a human. The fundamental architecture of AI is the same as a human brain. It understands, reasons, and actions much closer to a human than a computer. Treat it as such. You’ll be mind-blown by the results.

Treat every prompt as the start of a conversation, not the start of a

command. AI-assisted writing. Whether it's marketing copy, business communications, reports, or creative content - learn to use AI as a collaborator in producing written work. This skill transfers across virtually every white-collar domain. Workflow automation. Learn to identify repetitive tasks and figure out how AI tools can automate or accelerate them. Start with your own life - your job, your side projects, your personal tasks. Then you become the person who can do this for employers. Data analysis and synthesis. AI can help you analyze documents, extract insights from large amounts of text, summarize research, and identify patterns. Get good at directing AI to do this kind of cognitive work effectively. None of this requires paid courses, expensive software, or formal credentials. It requires time and deliberate practice.

Side Projects That Build Credentials

You don't need capital to build a portfolio. You need initiative. Let me give

you concrete examples so this isn't abstract. A real estate agent in Phoenix used Claude to build a neighborhood comparison tool for her clients. She described what she wanted in plain English and iterated with the AI until it worked. She now sends every prospect a personalized report that used to take her two hours to compile manually. It takes her ten minutes. Her close rate went up. She didn't learn to code. She learned to describe what she needed. A laid-off paralegal spent three weeks using ChatGPT to build an automated contract-review workflow for small businesses. She'd feed in a contract, get a plain-English summary of the key terms, risks, and missing clauses. She posted about it on LinkedIn, got 50 DMs in a week, and turned it into a

consulting practice charging $500 per engagement. Her total investment was time and a free-tier AI account. A college student with no work experience used AI to analyze every earnings call transcript for a set of mid-cap companies, then published a weekly newsletter with his analysis. Within six months he had 2,000 subscribers and three job offers from hedge funds. His credential wasn't a degree or an internship - it was public evidence that he could extract insight from data. This is what "starting from zero" actually looks like in 2026. You don't need to code. You don't need capital. You need a problem worth solving and the willingness to sit down with an AI and iterate until you've solved it. Start a blog or newsletter where you analyze something in your domain using AI tools. Write about what you're learning. Document your experiments. This creates a public track record that employers can evaluate. Build small tools or workflows that solve real problems, even if it's just for yourself or friends. A Chrome extension. An automation script. A chatbot for a local business. These become portfolio pieces. Remember - you don't need to code anymore. You just need to start having a conversation with the AI. Offer to help small businesses or nonprofits implement AI tools for free or cheap. You get experience, they get value, and you build references. This is how you develop skills that larger employers will pay for. Contribute to open source projects that are building AI applications. The barrier to entry is learning, not money. And the credential of being a contributor is real. Check out OpenClaw - it'll blow your mind. The goal is to create tangible evidence that you understand how to deploy AI effectively. That evidence is worth more than certificates from expensive courses.

The Equity vs Cash decision

Here’s a unique angle that you can take if you have access to it with your

employer: when you have the choice, lean toward equity compensation over pure salary - if the company is positioned well for AI. I know this sounds contradictory. If you need cash, why take equity? Because equity is how you get on the capital ownership side of the barbell without having capital to invest. Early employees at successful AI-first companies can see returns that dwarf anything they could have saved from their salary. I'm not saying take below- market salary - I'm saying when you're negotiating, understand that equity in the right company is potentially worth more than a few extra thousand dollars in base pay. This means being picky about which companies you join. Prioritize companies that are genuinely AI-first, have viable business models, and could see significant growth. Taking equity in a dying company is worthless. Taking equity in a company positioned to benefit from AI transformation could change your trajectory.

Free Learning Resources

The internet is full of expensive AI courses taught by people who have never

actually deployed AI in production. The quality ones are extremely hard to find.

These are actually worth your time, and they're all free: YouTube. Channels covering practical AI deployment, workflow automation, and applied AI skills. Some creators are building in public and showing exactly what works. Company documentation. Anthropic, OpenAI, xAI, Google, Chinese models… all publish extensive documentation on how to use their systems

effectively. Reading primary sources beats reading someone's summary of the primary sources. If it’s too technical or hard to read, just ask AI to dumb it down for you. X. The practitioners are sharing what works and what doesn't in real time. Follow the builders, not the hype accounts. X is easily the best repository for up to date news on AI, and it’s not even close. Coursera, edX, and Khan Academy. Legitimate courses from real institutions, available for free if you don't need the certificate. The knowledge is the same whether you pay or not. Open source communities. Discord servers, GitHub discussions, and forums where people building with AI share knowledge freely. The information asymmetry that used to protect expertise is collapsing. Everything you need to know is available. The question is whether you're willing to do the work to learn it. Pick your favorite AI LLM. ChatGPT. Grok. Gemini. Claude. Deepseek. Whatever. Use it.

The Harder Truth

Even with all of this, the path from zero is harder than the path from some. That's just true. Having capital provides options that not having capital doesn't. Someone who can invest $50K in AI-first companies today has an advantage. Someone who can afford to take six months off to retrain has an advantage. Someone who has a financial cushion that allows risk-taking has an advantage. I'm not going to pretend otherwise. What I am saying is that the path exists. People have built significant careers and eventually significant wealth starting from nothing before. I’m a perfect example of this. I was worth negative $100k when I graduated college. My parents almost lost their house during the financial crisis in 2008.

The difference now is that AI tools make individual capability matter more

than ever. A single person who really understands how to deploy AI can be more productive than teams that don't. That's the leverage you have access to, and it's real. The practical next step is not to feel overwhelmed by everything I've described. It's to start with one hour tomorrow. Open ChatGPT or Claude. Use it to learn something in your field. Use it to do a task you were going to do anyway, but better. Use it to explore what's possible. Then do that again the next day. And the day after. Compound that over months and years. Capital is the fast path. But consistent skill development is still a path. And the tools to develop those skills have never been more accessible.

Building Your Own AI Capabilities

Beyond skills, let me talk about building actual AI capabilities you can use. Most people interact with AI as consumers. They use AI to answer questions, write emails, maybe help with coding. That's fine, but it's just the baby version of what AI can do. The real leverage comes from building workflows that incorporate AI into repeated processes. What tasks do you do repeatedly that could be partially automated with AI? Not fully automated - that's often not the right goal yet - but augmented. Can you create systems where AI does first drafts that you review and polish? Can you build processes where AI handles research and you handle synthesis? Can you develop workflows where AI generates options and you make selections? This is different from "using AI tools." It's building AI into how you operate. The people who figure this out become dramatically more productive. They can take on more work, deliver faster, and provide more value. That's worth something in a labor market, even as cognitive labor generally gets devalued.

And what I've found is that AI augmentation is a skill that transfers across

domains. Once you learn how to effectively incorporate AI into your workflow, you can apply that pattern to new tools as they emerge. You become someone who gets value out of AI rather than someone who gets replaced by it.

The Builder Roadmap: First Principles, Not Incremental

If you're actively building - starting companies, launching products, creating

at scale - your roadmap is different from both groups above. You're not just positioning to benefit from the rebalancing. You're actively creating it.

AI Is Your 10x Multiplier

Every hour you spend without AI leverage is an hour your competitors are

spending with AI leverage. This compounds. I'm talking about fundamentally rethinking how your work gets done. What tasks currently require human judgment that AI can do at 80% quality? What processes currently take weeks that AI can do in hours? What products currently require teams that AI can help you build solo? The builders who figure this out first win. Not just because they ship faster - though they do - but because they're learning how to operate in the new paradigm while competitors are still optimizing the old one.

Data Is Your Moat

Whatever you're building, think about the data layer. What data will your

product generate? How does that data create compounding advantage? Why can't someone else replicate it? The best AI businesses are not AI companies per se. They're companies that generate proprietary data through their core operations, then use AI to extract value from that data. The AI is a tool. The data is the moat.

If your business doesn't generate proprietary data, it's probably not defensible in the AI era. Someone else will use publicly available data plus better AI to do what you're doing cheaper and faster.

Scale Is Everything

The economics of AI favor scale dramatically. Training costs are largely fixed. Inference costs decrease with scale. Data advantages compound with user growth. This means going big is more important than going early. A well-executed later entrant with massive scale will crush an early entrant that stays small. This is why I think xAI will ultimately beat OpenAI & Anthropic despite OpenAI's head start - Elon knows how to scale. If you're building something, think about scale from day one. Not "maybe we'll scale later" but "how do we get to massive scale as quickly as possible?" The builders who capture scale capture the winner-take-all markets. First Principles, Not Incremental Thinking I've talked about first principles throughout this book. For builders, it's essential. Incremental thinking asks "how do we make this 10% better?" First principles thinking asks "if we were starting from scratch with current technology, what would we build?" The answers are often radically different. Incremental thinking led to better horse carriages. First principles thinking led to automobiles. Incremental thinking leads to ChatGPT plugins. First principles thinking leads to AI agents that replace entire workflows. Most builders think incrementally because it's safer. You can copy what worked before with modest improvements. But AI is moving so fast that incremental approaches get leapfrogged constantly. By the time you've made your thing 10% better, someone else has built something 10x different.

The builders who win in this era are the ones willing to throw out

assumptions and rebuild from first principles. What would a law firm look like if you designed it today? What would education look like? What would healthcare look like? Those are the questions worth asking.

Skills That Remain Human

One more thing before I close this chapter. Amid all this talk of AI replacing

cognitive labor, there are categories of work that remain distinctly human. Understanding these might shape your personal roadmap. Physical presence matters. Some jobs require being there - whether that's a nurse at a bedside, a plumber under a sink, or a salesperson in a boardroom. AI can advise, but it can't be present. Until Optimus-scale robotics changes this - which is coming but not immediate - physical presence jobs are more defensible. Novel judgment matters. AI is excellent at pattern matching and optimization within known parameters. It's less excellent at genuinely novel situations that require judgment without precedent - at least for now. The more your work involves navigating unprecedented territory, the more defensible it is. Frontier creativity matters. AI can generate content at scale. It can even generate good content. But the frontier of creativity - the truly novel ideas, the breakthrough innovations, the art that changes how we see things - that's still human territory. If your work is creative at the frontier rather than creative at scale, you have more runway. Trust and accountability matter. Some decisions require a human to be responsible. Not because AI couldn't make the decision, but because humans need another human to trust and hold accountable. This is regulatory in some cases, psychological in others. Either way, it creates sustained demand for human judgment in certain roles. Relationship building matters. Humans trust humans. Business deals, sales, partnerships, negotiations - these often require human-to-human connection

that AI can't replicate. The people who are excellent at building trust and relationships have more defensible skills than the people who are excellent at tasks that can be done remotely and asynchronously. Complex coordination matters. Managing multiple stakeholders, navigating organizational politics, aligning incentives across groups - these are human skills that require understanding human motivations and social dynamics. AI can analyze patterns, but it struggles to navigate the messy reality of human organizations. I don't list these to suggest they're safe harbors. AI is coming for everything eventually. And Optimus-scale robotics will eventually address the physical presence category too. But if you're thinking about where to position over the next five to ten years, these categories have longer timelines than pure cognitive labor.

The Uncomfortable Truth

I've tried to make this chapter practical. But let me end with the

uncomfortable truth that underlies everything. Not everyone will navigate this well. Some people will read this book, nod their heads, and do nothing. Some will try to adapt and find it harder than they expected. Some will make bets that don't pan out. This transformation will create winners and losers, and telling you how to be a winner doesn't mean everyone who tries will succeed. The statistics are brutal. If the middle 60% is getting crushed, that's roughly 200 million Americans. Not all of them will successfully reposition. Not all of them will accumulate capital in time. Not all of them will develop AI-adjacent skills fast enough. I'm not saying this to be fatalistic. I'm saying it because false optimism would be worse than honest assessment. The rebalancing is real. The disruption is coming. The window for repositioning is short.

What I can tell you is that trying is better than not trying. Understanding the

landscape is better than ignoring it. Moving urgently is better than moving gradually. And owning capital, however small the amount, is better than owning only labor. You can't control the macro forces reshaping the economy. But you can control how you respond to them. That's what this chapter is about. That's what the whole book has been about.

It Could Be Amazing

I want to end this chapter - and this roadmap - with a story that gives me

genuine hope about what comes next. Chess is one of the oldest games in human civilization. For centuries, it was considered the ultimate test of human intellect. Grandmasters were revered. The game was seen as a window into the deepest capabilities of the human mind. Then computers beat us. IBM's Deep Blue defeated Garry Kasparov in 1997. By the 2010s, AI chess engines like Stockfish could obliterate any human player on Earth without breaking a sweat. The gap between human and machine chess capability became a chasm. A grandmaster playing Stockfish is like a toddler playing LeBron James. So chess must be dead, right? AI mastered it. Humans became irrelevant. Game over. Except the opposite happened. Chess has never been more popular than it is right now. Chess.com has over 150 million members. Magnus Carlsen has become a global celebrity. Hikaru Nakamura streams to hundreds of thousands of viewers. The Queen's Gambit drove chess set sales through the roof. Prize pools are bigger than ever. More people are learning, playing, and watching chess than at any point in human history.

How is that possible? AI solved the game - and that made it better for

humans, not worse. Here is what actually happened: Those same engines that can destroy any human player became the most powerful teaching tools chess has ever seen. A club player can now analyze their games with an engine that plays at a level beyond any grandmaster who ever lived. They can see exactly where they went wrong, exactly what the optimal move was, and why. The learning curve that used to take decades can now be compressed into years. And for spectators, the AI evaluation bar transformed chess content. When you watch a grandmaster match with engine analysis running, you can see in real time whether a move was brilliant or a blunder. You do not need to be a grandmaster yourself to appreciate the drama. The AI made the human game accessible to everyone. Professional chess players are not poorer because of AI. They are wealthier. They are more famous. Their audience is larger. Their content reaches more people. The AI did not replace them - it amplified them. It made their talent legible to the masses in a way that was impossible before. Now think about what this means for the broader thesis of this book. AI will master cognitive tasks the way it mastered chess. That is inevitable. But mastering the task does not necessarily eliminate the human. It can amplify the human. It can make human expertise more accessible, more visible, more valuable to a broader audience. A doctor assisted by AI does not become irrelevant. They become more accurate, more efficient, and able to serve more patients. A teacher with AI tools does not become obsolete. They become capable of personalizing education for every student in ways that were physically impossible before. A creator with AI does not lose their voice. They amplify it to reach audiences they never could have reached alone. The chess model is not guaranteed. It requires that we build systems where AI augments rather than simply replaces. It requires that the economic gains

flow to the humans who use these tools, not just to the companies that build them. It requires active choices about how we structure this transition. But the chess model proves it is possible. AI mastery and human flourishing can coexist. They already do - in at least one domain. From a first principles perspective, there are no blockers to this dynamic playing out across every industry, physical or digital. That is the future worth fighting for. Not one where humans are sidelined, but one where humans are amplified. Where AI handles the drudge work so we can focus on the parts that make us human - the creativity, the connection, the meaning. And hopefully this book gets us going in that direction.