English | 中文
Tony Seah
The companion piece, GPUs, Wombs & Money Printers, traced the likely impact of AI on the economy, geopolitics, demographics, finance, and the future of our species. One conclusion: the endgame of the AI era could be a quasi-utopia of radical material abundance and on-demand distribution. But that endgame is not arriving tomorrow. Between now and then lies a transition period of perhaps 10, 20, or even more years -- and for ordinary people, that transition is going to hurt.
This article is written for that transition.
It is not a pep talk. It is not a hollow call to "embrace change." It is an attempt at honest analysis: what you can do as an individual, what you cannot do but need to understand, and -- most importantly -- how to keep yourself from falling apart when everything becomes uncertain.
Before we start, I need to admit a few things:
First, much of this advice may turn out to be wrong. The pace of change AI is bringing exceeds any precedent in human history. A strategy that looks sound today could be obsolete in two years. This article is not scripture -- it is "the best framework we can think of right now," and reality may overturn it at any moment.
Second, not everyone has the luxury of choice. "Build passive income," "invest in yourself," "learn new skills" -- all of these assume you have the time, the energy, and a certain financial cushion. Telling someone who earns 3,000 yuan a month, works 12-hour days, and supports aging parents and young children to "build a personal brand" is a "let them eat cake" moment. This article will try to distinguish between "things an individual can do" and "things that require systemic solutions."
Third, some problems simply have no individual solution. Pension collapse, mass unemployment, financial system meltdowns -- you cannot save enough money to dodge these. For problems like these, I can only be honest: they require collective action and policy change. All an individual can do is soften the blow, not become immune.
Fourth, the endgame may be good, but the transition will be painful. If AI truly delivers radical material abundance, humanity may enter an age where no one needs to work to survive. But the road from here to there is littered with unemployment, devaluation, identity crises, and institutional gaps. What you need is not the patience to "wait for utopia" -- it is the ability to survive until utopia arrives.
Following the chain of reasoning from the companion piece to its logical end, we arrive at a seemingly paradoxical conclusion:
The combined productivity of AI + robotics may deliver radical material abundance. The production costs of food, clothing, housing, healthcare, and education could approach zero. Sam Altman predicts "massive deflation" -- AI usage costs falling tenfold every year [34]. Marx's vision of communism -- material abundance, distribution according to need, free human development -- may actually be realized, not through proletarian revolution, but through a fleet of machines that never need to eat.
Standing in 2050 and looking back, you might say: "That stretch was rough, but the outcome was not bad."
The problem is that you are not standing in 2050. You are standing in 2026.
Between now and that "abundant endgame" lies a transition of perhaps 10-20 years. During this period:
- The production side has already been revolutionized by AI -- efficiency skyrockets, costs plunge, GDP may keep growing
- But the distribution side is still stuck in the old system -- wages are the only income source, social insurance runs on payroll taxes, pensions assume people die around 80
The time lag between production and distribution is the ordinary person's nightmare.
GDP is growing, but you just got laid off. Prices are falling, but your mortgage hasn't. AI is making everything cheaper, but you can't afford even the cheap stuff -- because you have no income. This is the "ghost GDP" described in the companion piece: statistics paint a rosy picture while ordinary people see only ruin.
The core contradiction of the transition: technology advances far faster than institutions can adapt.
Self-driving cars could reach mass deployment within five years, but retraining taxi drivers takes ten. AI could replace most entry-level programming jobs within a year, but university curriculum reform takes five. Robots could take over warehouses within three years, but reforming the social safety net may require a generation of political struggle.
In the gap between technology arriving and institutions catching up, ordinary people are exposed. This article is about helping you get through that gap with as little damage as possible.
This is not the first time.
The British Industrial Revolution (1760-1840): The endgame was a several-fold increase in per capita income. But the transition -- roughly 60-80 years -- was filled with child labor, slums, cholera, and workers' uprisings. The average life expectancy of a Manchester worker once dropped to just 17. When Engels wrote The Condition of the Working Class in England, the Industrial Revolution had been underway for 80 years.
The American Gilded Age (1870-1900): The endgame was America becoming the world's largest economy. But the transition -- railroad barons monopolizing everything, workers toiling 16-hour days, no minimum wage, no social insurance, no workers' compensation. It took the New Deal of the 1930s to build a social safety net.
Every technological revolution follows this pattern: the endgame is prosperity, but the transition is blood and tears. The difference lies in how long the transition lasts, how painful it is, and whether institutions can keep pace.
What makes the AI revolution unique: the speed is an order of magnitude faster. The Industrial Revolution took 60 years to put textile workers out of jobs; AI may take 6 years to do the same to programmers. The transition is compressed -- meaning the pain is more concentrated and the window for institutional adaptation is shorter.
Most "survival guides for the AI era" put mental health last -- first they teach you skills, finance, and career planning, then tack on a "take care of your mental health" appendix.
That priority order is wrong.
If someone's mental state collapses, they will not learn new skills, manage their money, or build connections. They will lie in bed doomscrolling, or worse. Mental health is not a "soft" need -- it is the prerequisite for all action.
Viewed through Maslow's hierarchy of needs, AI does not strike from the bottom (can't afford food) -- it crushes downward from the top, simultaneously:
| Need Level | AI's Impact | Speed of Collapse |
|---|---|---|
| Self-actualization | "Everything I spent a lifetime learning is suddenly worthless" | Collapses first |
| Esteem | "I am no longer needed; I am surplus" | Follows closely |
| Belonging | "Social circle shrinks after job loss; AI companions replace real people" | Slow erosion |
| Safety | "Can't make mortgage payments; pension falls short" | Delayed onset |
| Physiological | "Can't afford food" | Last (unlikely in developed nations) |
Most people don't break down when they can't eat -- they break down when they feel useless.
Data from the Great Depression: Between 1928 and 1932, the U.S. suicide rate surged 22.8%, from 18.0 to 22.1 per 100,000 -- the highest in American history. The sharpest rise was not among the poorest but among working-age adults aged 25-64 -- the group with the strongest "career = identity" bond. They did not starve. There was no mass famine in 1930s America. They were stripped of purpose.
Imagine you are a 45-year-old senior developer. You spent 20 years mastering programming, from C++ to Java to Python, rising to the role of architect. Your professional identity, social standing, income, and self-confidence are all built on "I can write code."
Then Claude Code arrives. A fresh graduate uses it for a few weeks to deliver what would have taken you months. Your 20 years of experience become "legacy knowledge" in the face of AI.
This is not a "just learn a new tool" problem. This is identity-level destruction.
The same thing is happening to translators (10 years perfecting the craft, AI produces results in seconds), accountants (earned the CPA, AI processes books 100x faster), designers (years of visual intuition, AI generates better-looking work), lawyers (memorized thousands of cases, AI searches all case law without omission), and journalists (decades of editorial experience, AI mass-produces news copy).
Behind every devalued skill is a person's youth, sweat, and sense of self. Telling them to "just learn something new" is easy to say from the sidelines.
What's crueler still: learning something new may not even help. You spend a year acquiring a new skill, and AI may be able to do it six months later. Your learning speed can't outrun the rate of depreciation -- a predicament without historical precedent. In past technological revolutions, a horse-carriage driver could move to a factory; when factories automated, workers moved to the service sector. There was always a "next stop." This time, AI may replace all the stops simultaneously.
This is not a politically correct discussion -- it is a data-driven one.
Virtually every culture places a core expectation on men: economic provider. "Providing for the family" is not just a responsibility; it is the central pillar of male identity.
When that pillar is pulled away:
- The rate of depression among unemployed men is 1.5-2x that of women
- For every 1-percentage-point rise in male unemployment, the marriage rate drops 1.5-2%
- Thirty years of economic stagnation in Japan: the lifetime non-marriage rate for men soared from 5% to 28%
- In China, "owning a car and a home" is the price of entry to the marriage market -- AI-driven layoffs lock more men out
Women are equally affected by AI (customer service, admin, accounting, HR -- white-collar roles with high female representation are hit first), but men lose something additional: their core identity. Middle-aged men may be the highest-risk group for mental health in the AI era.
British psychologist Richard Wiseman spent ten years tracking hundreds of self-described "lucky" and "unlucky" people, publishing his findings in The Luck Factor. His central discovery: luck is not random -- "lucky people" and "unlucky people" exhibit entirely different psychological and behavioral patterns. And these patterns can be learned -- when he taught "unlucky" people to adopt "lucky" thinking, their lives genuinely improved.
This is highly relevant to surviving the AI era, because the years ahead are full of uncertainty -- and Wiseman's research is precisely about "how to find good outcomes amid uncertainty."
Four behavioral patterns of lucky people:
1. Create and seize chance opportunities -- expand your "luck surface area."
Lucky people have broader social networks, talk to people from different backgrounds, and are willing to try new things. Unlucky people stay in their comfort zone, associate with familiar faces, and reject new experiences.
In the AI era, your next opportunity will most likely not come from your existing circle. Actively expanding your exposure -- attending events you would normally skip, talking to people you would not normally talk to, trying things you would not normally try -- increases the probability of encountering good fortune.
2. Maintain a relaxed awareness -- anxiety is the enemy of opportunity.
Wiseman ran a classic experiment: he asked subjects to count the number of photos in a newspaper. On the second page, in large print, it said: "Stop counting -- there are 43 photos in this newspaper." The lucky people noticed it and finished in seconds. The unlucky people were so focused on counting photos that they missed it entirely.
Anxiety narrows your field of vision. Relaxation lets you see opportunities.
The biggest trap of the AI era is anxiety -- it locks your attention on "what you've lost" while blinding you to "what's emerging." Staying relaxed does not mean not caring -- it means paying attention to problems while preserving the bandwidth to notice new possibilities.
3. Positive self-expectation -- belief becomes self-fulfilling.
Lucky people believe good things will happen to them. This is not blind optimism; it is a mindset that fulfills itself. When you believe you can find a way out, you act more proactively, and that action itself increases the probability of finding one.
The concept of "reflexivity" from the companion piece applies perfectly here -- your beliefs about the future change your behavior, and your behavior changes your future.
4. Counterfactual thinking -- broke a leg? At least it wasn't both.
When bad things happen, lucky people automatically imagine "how it could have been worse" -- Got laid off? "At least it happened while I had savings, not when my mortgage payments were at their peak." Skills devalued? "At least I found out early, not five years from now."
This is not self-deception -- it is a validated psychological resilience mechanism. It does not change the facts, but it changes your capacity to bear them. Over the next few years, almost everyone will encounter "bad things." The difference is not whether you can avoid them (you probably can't), but your psychological elasticity after they happen.
Practical steps:
- Every week, have a conversation with at least one person you would not normally talk to (expand your luck surface area)
- When you feel anxious, deliberately slow down -- go for a walk, cook a meal, instead of scrolling your phone (cultivate relaxed awareness)
- Every day, write down one thing that "could have been worse" (train counterfactual thinking)
- Set a small, actionable positive expectation -- not "I'll earn a million this year" but "this week I'll finish X"
Anxiety has a formula:
Anxiety = Importance x Uncertainty
As long as something is unimportant or more certain, anxiety decreases. The problem in the AI era is that many things are simultaneously becoming "very important" and "very uncertain" -- Will I lose my job? Can I still make mortgage payments? Is the investment in my child's education worth it? Will my skills become obsolete?
Anxiety goes through the roof.
But the formula also gives you two clear levers: reduce the uncertainty or reduce the importance.
You cannot eliminate uncertainty (no one can predict the precise speed of AI development), but you can increase your tolerance for it:
- Break big questions into small ones. "What do I do for the next 20 years?" -- that question has 100% uncertainty and will crush anyone. But "what three things will I do in the next six months?" -- much less uncertain. Pull the focus of your anxiety from "the long-term big picture" back to "near-term actionable steps."
- Build a worst-case plan. The most tormenting thing about uncertainty is "not knowing what will happen." But if you have already thought through the worst case and have a contingency ("If I get laid off, I have X months of savings; Plan B is XX"), the destructive power of uncertainty drops dramatically -- because it becomes "certainty about something bad," and a certain bad outcome is far easier to handle than an uncertain one.
- Control your information intake. Reading ten "AI will replace everyone" articles a day only amplifies anxiety. This does not mean burying your head in the sand -- it means distinguishing between "useful information" and "anxiety-generating noise." Recommendation: catch up on AI developments once a week rather than being bombarded by push notifications every day.
- Wiseman's relaxed awareness: As mentioned earlier -- anxiety narrows your vision; relaxation lets you see opportunity. Deliberately practice "attention in a relaxed state" -- good ideas come more easily during a walk, while cooking, or while exercising than while anxiously scrolling your phone.
Many things making you anxious are not actually as important as you think they are. But you have never questioned their importance, because all of society keeps telling you "this matters."
Work. "Losing my job" is what most people worry about most. But ask yourself: what exactly are you anxious about? If it is "losing income" -- then the core issue is financial security, not "the job" itself. After cutting expenses and building a buffer, "losing this particular job" becomes less critical. If it is "losing identity" -- then the issue is that you have tied too much self-worth to your job title (see Strategy 1 in section 2.6: decouple identity).
Your child's grades. If you accept the idea that "winning exams doesn't matter in the AI era" (see Chapter 8), half your anxiety about grades evaporates instantly. What you are really anxious about is not the grades themselves, but the causal chain: "bad grades -> can't get into a good university -> can't get a good job -> life is ruined." When that chain itself is breaking, the "importance" of grades drops sharply.
Face. "What will people think if I get laid off?" "What will relatives say if I take a pay cut?" -- social comparison is a massive source of anxiety. But in an era of mass structural unemployment, getting laid off is not a reflection of personal ability; it is the tide. When the people around you are going through the same thing, "what others think of me" naturally fades -- because everyone is too busy dealing with their own situation.
Your home. "What if property prices drop?" -- If your home is where you live, what does the price have to do with your daily life? You still live in the same place. The price only "matters" when you plan to sell. Cross "property value" off your anxiety list (unless you actually need to sell), and your anxiety drops a notch.
One-line summary: Examine your anxiety list and ask yourself, "Does this actually matter, or am I being held hostage by social inertia?" You will find that many "important things" can actually be let go. Let go of one, and your anxiety drops by one degree.
Strategy 1: Decouple early -- detach your identity from your career
Do not wait until the day you lose your job to ask "Who am I?" Start now.
Exercise: Write down ten answers to "I am ____." If your first five are all work-related ("I am a programmer," "I am a director at XX Corp"), your identity structure is highly fragile.
Goal: Make at least half of your self-concept independent of work -- "I am a runner," "I am a father of two," "I am a member of the community basketball team." When work disappears, these identities remain.
Strategy 2: Build sources of fulfillment that AI cannot replace
AI can paint better pictures than you, write better articles, and write better code. But AI cannot replace your experience of the process.
Running is not about being faster than a car. Playing guitar is not about outperforming an AI musician. Cooking is not about being more efficient than a robot. The process is the purpose.
Find at least one thing you do purely because "the act of doing it feels good." In the AI era, this kind of intrinsic, process-oriented satisfaction is the only thing that cannot be replaced.
Suggestions:
- Physical activities (running, swimming, hiking, martial arts) -- AI has no body; physical sensation is yours alone
- Hands-on crafts (woodworking, pottery, cooking, gardening) -- the satisfaction of creating something tangible with your own hands
- Social activities (communities that meet regularly -- book clubs, sports teams, volunteer organizations)
- Creative pursuits -- but with process as the goal, not output (keeping a journal rather than chasing viral articles)
Strategy 3: Your social network is your "psychological ICU"
Isolation is the deadliest factor in a crisis.
Research consistently shows that social support is the strongest protective factor against suicide and depression. Not because friends can help you find a job (though they might), but because "someone cares about me" is itself a reason to keep going.
But the AI era carries a trend: social connection is shrinking. Remote work reduces informal interaction. AI companions make loneliness "bearable." Social media creates the illusion of connection while deepening isolation.
Actively maintaining real human social networks is one of the most important investments you can make in the AI era. You do not need many people -- 5-10 whom you trust deeply is enough. But these relationships require ongoing investment: meeting in person regularly, sharing experiences, showing up when the other person is struggling.
Specific suggestions:
- At least one face-to-face social interaction per week (not online)
- Join at least one community that meets regularly in person
- Proactively reach out to a friend you have not seen in over six months
- If you find yourself increasingly relying on AI chat to meet your social needs -- that is a warning sign
Strategy 4: Accept uncertainty as the new normal
The old life narrative was linear: school -> work -> retirement -> death. Every step had a "right answer."
The AI era has no right answers. You might change careers at 35, change again at 45, and discover at 55 that everything you learned before is useless.
Accepting this is itself a form of psychological fortification. If you still measure yourself against a linear narrative, every "deviation from the track" feels like failure. But if you accept that "there is no track," every turn is just a turn -- not a failure.
Recommended practices:
- Meditation/mindfulness -- training the ability to "accept the present without judgment," backed by solid neuroscience evidence
- Journaling -- writing out anxiety; research shows that putting worry into words reduces amygdala activation
- Deliberately exposing yourself to small uncertainties -- trying new foods, visiting unfamiliar places, talking to strangers
- Losing your job is not your fault. This is structural change. Just as a horse-carriage driver in 1900 did not lose his job because he was not working hard enough -- the automobile arrived.
- Seeking professional help is not weakness. If you have felt hopeless, sleepless, or unable to concentrate for more than two weeks straight -- see a doctor. It is as normal as seeing a doctor for a cold.
- Downshifting is not defeat. A former high earner delivering food is not failing -- it is being pragmatic. Being alive matters more than saving face.
- Do not carry it alone. Tell someone you trust what you are going through. Shame loses its power the moment you speak it.
- Crisis hotlines: China 24-hour mental health hotline 400-161-9995; Beijing Psychological Crisis Research and Intervention Center 010-82951332
If someone near you has recently lost a job, become withdrawn and quiet, said things like "I'm useless," or started drinking noticeably more -- reach out to them.
You do not need to say anything profound. "I've been thinking about you. Want to grab a meal?" That one sentence may be enough.
The most counterintuitive fact of the AI era: in a world of extreme technological advancement, the most primitive form of human connection -- "someone cares about me" -- may matter more than any technology.
In February 2026, a new term emerged on Wall Street: HALO -- Heavy Assets, Low Obsolescence.
Coined by Josh Brown of Ritholtz Wealth Management and subsequently adopted into Morgan Stanley's investment framework, the core logic is: tangible assets (factories, energy, logistics, land) are safer than intangible assets (software, brands, intellectual property) -- because AI can copy code but cannot copy an oil refinery.
Goldman Sachs data: since 2025, HALO-type companies have outperformed asset-light companies by 35%. The energy sector is up 25.3%, materials up 18.1%. Meanwhile, SaaS companies lost roughly $300 billion in market capitalization in a single trading day.
What this means for ordinary people: HALO is not just an investment strategy. It is a cognitive framework -- things that are hard to replicate digitally are what hold lasting value. Whether we are talking about companies, skills, or you yourself.
In 2025, Andrej Karpathy introduced "vibe coding" -- describe what you want, and AI generates the code.
Claude Code, Cursor, and Replit Agent already enable a mid-level developer to replicate the core functionality of a medium-sized SaaS product in a matter of weeks.
The core paradigm of software engineering has been upended. For the past 50 years, the industry pursued DRY (Don't Repeat Yourself), modularity, reusability, and scalability -- because writing code was expensive. But when AI can generate full features in minutes, the economic rationale for "reuse" vanishes. Code shifts from "asset" to "consumable" -- use it, discard it, regenerate when needed.
This is "disposable software." You no longer need to subscribe to a project management SaaS -- AI generates a custom version tailored to your needs in 10 minutes. Free, perfectly fitted, no monthly fee.
Which SaaS companies can survive? Only those whose moat lies not in "code" but in:
- Data: unique, non-replicable data (e.g., Bloomberg Terminal)
- Network effects: the more users, the more valuable (though AI agents are eroding this too)
- Regulatory barriers: legal requirements to use specific systems (healthcare, financial compliance)
Everything else -- project management, CRM, forms, basic analytics -- is a pseudo-asset.
The same logic applies to your career and income:
Ask yourself: Is my value "HALO-type" or "SaaS-type"?
- If your value lies in "I can do something AI can also do" -- you are SaaS, and your moat is being hollowed out
- If your value lies in "I possess something AI cannot replicate" (trust relationships, physical presence, unique experience, regulatory credentials) -- you are HALO
This framework is far more useful than "learn AI."
Characteristics: AI is a tool; humans are still in the loop. Most jobs still exist, but efficiency expectations are surging.
What you can do:
- Use AI to amplify your existing capabilities -- not "learn AI," but "use AI to do what you already do, faster and better"
- This is your highest-earning window -- save what you earn, reduce leverage
- Expand your social network -- in an economic crisis, connections are more valuable than resumes
- Cultivate at least one "HALO-type" skill or income source
Traps:
- "My job is fine" -- American homeowners in 2007 thought the same
- Going all-in on a specific AI tool -- tools change; judgment does not
- "I'm senior enough to be safe" -- seniority buys more time than being junior, but less than you think
Characteristics: Mass unemployment becomes visible; UBI or some alternative begins rolling out; consumption patterns shift dramatically.
Keys to survival:
- Low leverage -- no mortgage, no heavy debt is the biggest safety net
- Diversification -- not dependent on any single income source
- Health -- those who live to see age-reversal/AI medicine mature could gain decades of additional life
- Community -- the isolated are most vulnerable; those with mutual support networks are most resilient
An honest admission: For most ordinary people, outcomes in this phase depend on government redistribution policy, not individual effort. If UBI is not in place and distribution systems are not reformed, personal strategy has extremely limited impact. What you can do is make sure you hold on until the policy arrives.
Augmented humans emerge, species divergence begins, "work" may have ceased to exist.
This phase is too distant for specific strategies to matter much. The only certainty: being alive when it arrives is the prerequisite for everything.
Relatively safe ("atom world" assets):
- Real estate in core cities (location cannot be replicated)
- Energy and infrastructure investments
- Physical gold
- Your own health and skills (the ultimate "heavy asset")
High risk ("bit world" assets):
- Tech stocks (especially pure software companies)
- Anything tied to SaaS companies
- Commercial real estate dependent on white-collar spending
- Cryptocurrency (highly volatile, but may become the payment infrastructure of the AI economy)
Already depreciating:
- Real estate in third- and fourth-tier cities
- Investment in traditional degrees
- Business models built on "information brokerage"
In periods of extreme uncertainty, cash gives you time -- time to survive before you find your next income source. Recommendation: maintain 6-24 months of living expenses in liquid reserves.
Time for some cold water: most "passive income" sources are themselves under threat from AI.
- Online courses? AI can generate unlimited free courses
- E-commerce? AI agents make pricing so transparent that margins approach zero
- Content creation? AI can produce unlimited content
- Rent? Depends on the city
The only truly AI-resistant passive income comes from: returns on scarce, non-replicable resources in the "atom world." Prime urban locations, energy interests, trusted positions within specific communities.
For most ordinary people, rather than chasing "passive income," pursue "low burn rate" -- reduce fixed expenses so you can hold on even when income drops.
HALO-type (heavy skills, low obsolescence):
- Requiring physical presence: plumbing, welding, pipefitting, electrical work
- Requiring human trust: counseling, nursing, community organizing
- Requiring judgment and accountability: a surgeon making operative decisions, a lawyer arguing in court
- Requiring taste and aesthetics: haute cuisine, interior design
Light skills (high obsolescence risk):
- Pure information processing, pure coding ability, pure knowledge memorization, standardized process execution
The prompt engineering you learn in 2025 may be useless by 2027. Claude Code itself is upending the advice to "learn to code."
What you should learn is not the tool but the judgment -- "Is this AI output correct?" "When should I trust AI, and when shouldn't I?" "Which of these three options is best?"
- 2025-2028: Use AI tools to amplify your existing skills -- the highest-ROI window
- 2028-2032: Physical skills appreciate; judgment and trust relationships become core assets
- 2032-2040: Humanoid robots mature, closing the physical-skills window too; value may reside solely in "humanness" -- trust, companionship, emotional connection
The greatest danger is applying last phase's strategy to the next phase.
Advantage: Savings, networks, learning ability Risk: Overconfidence. "AI can't replace me" -- someone says this during every technological revolution. Recommendation: Use your advantages to build redundancy -- diversified income, assets in core cities, investment in health.
The most vulnerable group. Enough income to feel "fine," but not enough reserves to weather a shock.
What you can do: Deleverage (top priority), save money, spend your spare time learning skills that will not become obsolete, think through your Plan B. What you cannot do: Solve pension devaluation and mass unemployment on your own -- these require systemic solutions.
Paradoxically, safer than white-collar workers in the short term -- AI replaces cognitive labor first. Humanoid robots are still 5-10 years out.
What you can do: Use the window to learn trade skills that are in high demand and short supply. Protect your body. Build community mutual-aid networks. What you need (but cannot achieve alone): Government retraining programs, a social safety net that catches you, distribution reform.
Let me start with a fact that may be uncomfortable but must be faced: your child will almost certainly not beat AI in cognitive ability. No matter how good a university they attend, how trendy their major, how high their GPA -- on the dimensions of pure knowledge and pure skill, AI will be better.
But this is not bad news. This is liberation.
Think about it: you push your child to cram, attend tutoring classes, chase grades -- for what? So they can "get a good job and earn good money." But if the endgame of the AI era is radical material abundance and humans do not need to work to survive, the entire premise of that logic chain ceases to exist.
You are forcing your child to prepare for an exam that will never take place.
This does not mean education is unimportant. It means the purpose of education needs to be fundamentally redefined: not "beat everyone else," but "become a whole person capable of enjoying life."
The capacity for joy. This is not a platitude. In an era where material scarcity may vanish but meaning could become scarce, knowing what makes you happy and having the ability to pursue it is the most important survival skill. A child who plays guitar, loves hiking, and enjoys cooking may thrive far better in the AI era than one who only knows how to study for tests.
Curiosity. Not "studying for the exam" curiosity, but "this is fascinating and I want to figure it out" curiosity. The only "skill" that will never become obsolete in the AI era is the impulse to learn something new.
Social skills and empathy. Real human connection -- trust, friendship, love -- is the scarcest resource in the AI era. A child with three or five close friends will fare far better in the future than one with a 4.0 GPA but no friends.
Physical fitness. A healthy body is the prerequisite for all options. And bodily sensations -- the exhilaration of running, the feel of water against skin while swimming -- are things AI can never experience, nor replace your experience of.
Resilience. The future is not a straight line; it is a series of turns. Let children grow up accustomed to "fail and try again" rather than "one exam decides your life."
- Grade anxiety -- AI will outperform any human on exams; this arms race has already lost its meaning
- The linear path of "good university -> good job -> good life" -- this road is breaking apart
- Comparing your child to other people's children -- in the face of AI, everyone loses that competition
- The urge to control -- you cannot plan for a future you cannot imagine, but you can raise a person capable of handling any future
Your child may be the first generation in human history that does not need to work to survive.
What does that mean? It means they can spend their entire lives doing what they truly want to do. Painting, exploring, studying dinosaurs, learning to bake bread, making friends, falling in love, seeing the world -- not "endure decades of work to earn money and then maybe enjoy retirement," but living the life they want from the very start.
The ones who land on Mars and travel between stars will most likely be robots and AI, not your child. And that is fine. Your child does not need to land on Mars -- they need to be happy on Earth.
Be happy when there is reason to be happy. Enjoy what there is to enjoy. This is not giving up; it is the clearest-eyed outlook for the AI era.
The first eight chapters focused on defense. But an article that only talks about threats and not opportunities would be dishonest. Because while AI is dismantling the old world, it is simultaneously creating a wealth of things that were never before possible.
This may be the most direct benefit AI offers ordinary people.
AI is democratizing diagnosis. Getting a top doctor's opinion used to require booking a specialist months in advance or flying to a major city's best hospital. Now an AI-assisted diagnostic tool can deliver near-expert-level analysis in minutes.
Specific advances:
- AI has improved cancer detection accuracy by nearly 40%, enabling earlier discovery and intervention
- An AI blood test in clinical trials in the UK can detect 12 types of cancer from just 10 drops of blood with accuracy as high as 99%
- Mental health chatbots are already providing 24/7 emotional support -- for those who cannot get a counseling appointment, this is something where there was previously nothing
- AI medical tools support real-time multilingual translation -- rural elderly who do not speak Mandarin, immigrants who do not speak the local language, can access medical information without barriers for the first time
AI is also accelerating drug development. Traditionally, developing a new drug takes 15 years and $2 billion. AI is compressing this to 5 years or less:
- Insilico Medicine used an AI platform to design drug candidates targeting the cancer target KRAS (notoriously "undruggable"), screening 100 million molecules
- Google DeepMind used AI to discover a previously unknown protein interaction critical to cancer cell survival -- something traditional methods would have been virtually incapable of finding
- 2025 saw the largest number of AI-originated molecules entering clinical trials, concentrated in oncology, fibrosis, autoimmune diseases, and rare diseases
Rare disease patients may be the biggest beneficiaries. Roughly 300 million people worldwide suffer from rare diseases; most conditions attract no pharmaceutical investment because patient populations are too small and markets too limited. AI dramatically lowers R&D costs -- niche diseases that were previously not worth pursuing may now become viable.
This is a severely underestimated area.
For the visually impaired: Tools like Microsoft's Seeing AI use AI to describe the surrounding environment in real time -- identifying objects, reading text, recognizing faces. A blind person can "see" a restaurant menu, "read" the instructions on a medicine bottle, or "recognize" a friend walking toward them.
For the hearing impaired: AI real-time captioning lets deaf people "hear" conversations -- not just video-call subtitles, but live transcription of face-to-face interactions.
For those with limited mobility: AI-powered smart homes allow wheelchair users and bedridden elderly to control lights, air conditioning, door locks, and curtains by voice -- tasks that once required someone else's help can now be done independently.
For those with cognitive impairments/the elderly: AI companion robots can chat with dementia patients, conduct cognitive training, monitor mood changes, and remind them to take medication. South Korea has already deployed more than 5,400 AI companion robots across 115 local governments.
For these groups, AI is not "replacement" -- it is "empowerment." It grants them independence and dignity they never had before. This may be the purest benefit of AI, free of any negative controversy.
AI may be the greatest equalizer in education.
What used to be required to access top-tier educational resources? A house in the right school district (millions of yuan), international school tuition (hundreds of thousands per year), one-on-one tutoring (hundreds per hour). Poor children and rich children were never on the same track from the starting line.
AI is changing this.
A rural child can now learn the same material as an urban child using AI -- personalized, at their own pace. An AI tutor will not lose patience because you asked a "dumb question," will not give up on you for being a poor student, and will not stop teaching you because your family cannot afford the hourly rate.
Stanford's Data Ocean platform has already provided nearly 3,600 free AI medical education certifications across 93 countries -- for students in developing nations, this was previously unimaginable.
In Kenya and Nigeria, AI tools already support Swahili, Yoruba, and Amharic -- farmers can interact with AI by voice on basic phones to get crop disease diagnoses.
In the 2024-2025 school year, 60% of U.S. K-12 teachers were using AI tools, saving 6 hours per week. Those 6 hours can be redirected to things AI cannot do -- one-on-one attention for students, handling emotional issues, building personal trust.
This is AI's greatest gift to people with initiative.
Sam Altman predicts that the first "one-person billion-dollar company" will emerge between 2026 and 2028. Anthropic CEO Dario Amodei believes there is a 70-80% probability of this happening by 2026.
This is not hyperbole. Midjourney -- the AI image generation company -- generated $500 million in revenue in 2025 with approximately 107 employees, producing $4.7 million per person. What if you compressed that further to one person?
AI lets a single individual do what used to require a team:
- Programming: Claude Code lets one person run 4-10 AI coding agents simultaneously, equivalent to a small development team
- Design: AI generates drafts; you provide aesthetic judgment and make the final selection
- Marketing: AI writes copy, generates ad creatives, analyzes data
- Customer service: AI handles 90% of routine inquiries; you handle the 10% that are complex
- Finance/Legal: AI handles basic compliance; you make strategic decisions
What does this mean? It means entrepreneurship is no longer a game reserved for the wealthy. You do not need funding, employees, or an office. You need a good idea, basic judgment, and a computer with internet access.
The probability of success is still low -- most startups fail, and AI will not change that. But the cost of failure approaches zero. Try an idea; if it doesn't work, discard it and try another. A single attempt used to cost millions; now it costs a few weeks and a few hundred dollars in API fees.
Beyond these "big" benefits, AI is quietly improving everyday life in countless ways:
- Language is no longer a barrier: Real-time translation lets you communicate with anyone in any language. Traveling abroad, doing cross-border business, reading foreign materials -- AI means "not speaking the language" is no longer a constraint
- Self-driving reduces accidents: Expected to cut traffic fatalities by as much as 90%. Roughly 1.35 million people die in car accidents globally each year -- if that number drops to 135,000, that is 1.2 million lives saved every year
- Personal productivity: AI helps you sort email, manage your calendar, summarize long documents, and draft replies -- freeing you from repetitive mental labor
- Fraud and security: AI protects you without you even knowing -- detecting anomalous bank transactions, identifying phishing emails, preventing identity theft
- Independent living for the elderly: Smart homes + AI voice assistants allow the elderly to live independently at home for longer, reducing dependence on children or caretakers
Summing up the benefits above, the core cognitive shift is:
AI is not just something that steals your job. It is also:
- Your personal doctor (at the screening level)
- Your all-subject tutor (free, patient, 24/7)
- Your startup co-founder (no salary required, no equity demanded)
- Your accessibility assistant (if you have any physical limitations)
- Your translator, secretary, researcher, and designer
The same AI is a threat to one person ("it can do my job") and a superpower for another ("it can help me do ten times as much"). The difference lies not in AI itself, but in how you see it and how you use it.
One designer worries AI will replace them -- but another uses AI to generate 50 concepts a day, then applies their own aesthetic judgment to pick the best 3. The latter produces 10x more than the former.
One writer worries AI writes better than they do -- but another uses AI for research, organizing material, and generating first drafts, then injects their own perspective and personality. The latter produces in one week what used to take a month.
The question is not "can you do what AI does" but "what can you plus AI do that was previously impossible?"
Seize these opportunities. Use AI to do what you always wanted to do but couldn't -- learn a new skill, build a small product, help your community, improve your health. This is not the empty slogan of "embrace change." This is concrete action.
Humans can endure enormous suffering, provided they know it has an end.
The British endured the Blitz during World War II because Churchill said "We shall never surrender" -- not "everything will be fine," but "beyond the suffering lies victory."
Many who took their own lives during the Great Depression were not the poorest -- they were those who believed "things will never get better." Despair kills; poverty alone does not.
What should the narrative of the AI era be?
Not "the apocalypse is here." Not "everything will be fine." Rather: "The next few years will be hard, but the hard part has an end. Survive the transition, and what lies beyond is very likely a better world."
This is not blind optimism. It is a rational assessment based on technological trends:
- The cost of AI is falling exponentially -- Sam Altman predicts a tenfold drop per year
- The cost of material production will approach zero
- The day when humans no longer need to work for food, clothing, housing, or basic healthcare is technologically achievable
- AI is already making healthcare, education, and accessibility services better, cheaper, and more widely available
The only question is how long the transition lasts. It could be 5 years; it could be 20. But the direction is clear.
Knowing the storm will pass and believing the storm is eternal produce entirely different behaviors. The former makes you repair the ship, stock provisions, support each other, and even savor moments of calm amid the rain. The latter makes you give up.
This is not optimism. This is strategy.
As analyzed in the companion piece: age-reversal technology and AI medicine may achieve breakthroughs between 2035 and 2045. Those who are alive at that point could gain decades of additional life. The return on your current health investments may be the highest in history.
Things that cost nothing but pay off enormously:
- 150 minutes of exercise per week
- Adequate sleep (7-8 hours)
- Weight management
- Regular health checkups
- Maintaining social connections -- the health damage of loneliness is equivalent to smoking 15 cigarettes a day
The core conclusion from section 5.5 of the companion piece:
Do not try to produce more humans to compete with machines. Instead, ensure that the wealth machines create flows to people.
The answer on the production side is already clear -- AI and robotics. The distribution system on the consumption side is the problem you need to solve. Before the first wave of mass AI layoffs arrives (estimated 2028-2030), distribution systems need to be in place. Otherwise, "ghost GDP" will become a real social crisis.
Let me lay out the core logic of this article one final time:
The endgame of the AI era is very likely a good one. Radical material abundance, humanity liberated from labor -- these are the logical endpoints of technological trends. Your children may be the first generation in human history that does not need to work to survive.
But the endgame is not arriving tomorrow. In between lies a transition that could last 10-20 years -- technology has already been revolutionized, but institutions have not caught up. During this gap, ordinary people are exposed.
This article is written for the transition. It cannot make you immune -- mass structural change is not something an individual can withstand. What it can do is help you hold on until the good days arrive:
- Mental resilience first -- decouple identity, train elasticity, expand your luck surface area, stay relaxed
- Know your assets -- the HALO framework: things in the atom world are more resilient than things in the bit world
- Respond in phases -- different windows call for different strategies; the greatest danger is using an outdated strategy
- Reduce leverage -- low debt is the biggest safety net
- Maintain social ties -- "someone cares about me" may be the most important resource during the transition
- Stay healthy -- living to 2035 may be the best investment of your lifetime
- Let your children be -- raising them to be happy matters far more than drilling them for an exam that will never exist
- Go on offense with AI -- it is not just a threat; it is a superpower. Use it to do what you could not do before
One last thing.
The storm will pass. Not because I am an optimist, but because technological trends point in that direction.
But before the storm passes, you need to be alive.
So: hold on to what you can, let go of what you should. Reduce risk, but do not stop taking action. When anxiety hits, go for a walk, cook a meal, see a friend. When life feels meaningless, do something that makes you happy -- not for anyone else, for yourself.
Pessimists are right; optimists move forward.
Not because optimists fail to see the danger -- but because they see it and choose to act anyway.
Stay alive. Cross the storm. Then see what lies on the other side.
This article is the companion piece to "GPUs, Wombs & Money Printers: A Scenario Analysis of Human Society in the AI Era." The companion piece analyzes "what will happen"; this article discusses "what you can do."
Read the companion piece: GitHub | PDF (Chinese)
Written by Tony Seah
Sources:
- Richard Wiseman - The Luck Factor (original paper)
- The new anti-AI trade sweeping Wall Street: 'HALO' - CNBC
- The SaaSpocalypse: AI Agents, Vibe Coding, and the Changing Economics of SaaS
- Vibe Coding & Disposable Software 2026
- Disposable Software: Why Code is No Longer an Asset
- Citrini Research: The 2028 Global Intelligence Crisis