Why Nuclear War Has Become a Structural Imperative of Our Time?

English Summary:

Claude Code showed me our brutal future: if AI replaces programmers at 1/10th the cost, how long do other knowledge workers have? Most aggressive estimate: 80% of middle-class jobs gone in 3-5 years.

Humanity faces three paths: Delay (reform distribution), Grab (external war), or Burn (internal conflict). Problem is, AI evolves exponentially while society changes linearly—the window for "delay" is slamming shut. When reform can't keep pace, war—including nuclear war—shifts from "unthinkable" to "structural imperative."

This isn't moral judgment but systems analysis: like an overloaded vessel must release pressure, accumulated contradictions will find their outlet. Our collective choices determine which path humanity takes.

Here, "structural imperative" refers to a risk pathway in which accumulated systemic pressure pushes extreme war options into certain decision makers' toolkits. It is not a moral claim, and it is not a prediction that nuclear war must happen.

Haven't written an essay in a while. Something hit me today, so here we go.

Spent the weekend diving deep into Claude Code, the latest hot product everyone's talking about. First impression? Holy shit, this is powerful. Really, genuinely powerful. And as an entrepreneur, I felt something stir deep in my soul—this thing is a game-changer. I won't dive into Code's specific features here (that's enough content for a dozen Medium posts). If you're curious, just search "Vibe coding"—it's been blowing up lately.

Let me be clear upfront: this isn't prophecy or advocacy. It's risk analysis. And I'm talking about macro trends and structural forces here, not individual success stories. Sure, there's always that one person who lost their job and became a TikTok millionaire, or that guy who made bank raising pigs in rural Kansas. But let's be real—80% of humanity gets swept along by historical currents, for better or worse.

Back to Code. What I saw wasn't just technical progress—it was a preview of class extinction. If AI can replace programmers—a job requiring perhaps the highest precision of any knowledge work—at 1/10th the cost with higher quality and efficiency, then what about everyone else? Copywriters, designers, analysts? And since this is happening globally, simultaneously, AI disruption isn't some localized phenomenon. It's this era's structural imperative—not a moral "should", but a systems-level "must"—like Watt's steam engine at the dawn of the Industrial Revolution.

Here's the difference: humanity had 200 years to digest the steam engine's disruption of pre-industrial structures. The Information Revolution compressed that to 50 years, and it was relatively gentle. Today? What's your most optimistic estimate? How much gentler can it possibly be?

In February, I wrote: "AI's rapid deployment could trigger widespread unemployment and social disorder within 5-10 years, with society having no time to adapt and industries unable to transform quickly enough. The International Labour Organization predicts 20% of jobs will be replaced within a decade—frankly, I think the ILO is drastically underestimating."

Time to revise those numbers. In the most aggressive scenario: 3-5 years, 80% of middle-class jobs gone. We're talking about the complete hollowing out of the $100-300k knowledge worker tier that's existed since the Industrial Revolution. The middle class as we know it—dead.

Assuming human nature hasn't changed (safe bet), history offers three solution paths:

• Delay — trade time for space (reform distribution systems)
• Grab — external war (redirect pressure outward)
• Burn — internal conflict (reduce pressure through... reduction)

Let's coldly analyze each path.

"Delay" requires social consensus for distribution reform, requires the wealthy to voluntarily surrender advantages, requires time to experiment and adjust. But here's the reality: AI evolves exponentially while human society changes linearly. When ChatGPT appeared, we debated its impact on education. Two years later, Code builds entire software projects solo. This scissors gap tells us the window for "delay" is slamming shut.

Look at real examples: Europe's been negotiating digital taxes for five years—zero substantive progress. US AI regulation bills are stuck in endless Congressional loops without even a basic framework. Global tech giant tax agreements? A decade of talks, still bullshit. This is what "linear" social change versus "exponential" tech development actually looks like.

Worse, the globalized system is fracturing. Roosevelt's New Deal worked because America had a massive internal market and relatively closed economy. Today, any country attempting distribution reform alone faces capital flight and competitive disadvantage. Global coordinated reform? Look at climate negotiations—good luck with that.

Still, this remains the least catastrophic path, worth pursuing. Some attempts: AI taxes funding retraining programs, Universal Basic Income pilots, knowledge society redistribution mechanisms, creating new public service roles. Imperfect solutions, but at least we're trying to "delay."

Now for "Grab" and "Burn." Let's strip away emotion: massive unemployment is regime-threatening instability. When internal pressure can't be reformed away, externalizing conflict becomes almost inevitable. Pre-WWI Europe: every nation faced industrialization's social contradictions. Solution? Send an entire generation to the trenches. WWII's logic chain is even clearer: Great Depression → unemployment → extremism → war.

AI warfare is different though. Traditional war employed masses, "solving" unemployment through conscription. AI warfare might involve small teams operating autonomous swarm systems for devastating strikes. Few soldiers needed, but population adjustment still achieved. From a cold governance calculus, it's grimly "efficient."

The nuclear threshold might be lower than we think. When decision-makers face regime collapse, "limited nuclear warfare" gets redefined. Tactical nukes become "oversized conventional weapons," limited exchanges become "manageable." History shows we consistently overestimate our control, underestimate escalation risks.

This is why nuclear war is a structural imperative—not morally "right," but systemically probable. Like an overloaded vessel must release pressure or explode, it becomes an option in certain toolkits. When all other options seem exhausted, the unthinkable gets tabled.

We're at history's fork. One path: reform, innovate, cooperate to absorb AI's disruption—difficult but civilizational preservation. Other path: conflict to "simplify" the problem—civilizational regression or annihilation. The choice appears to rest with a few decision-makers, but aggregate attitudes create historical momentum.

Y Combinator's motto is "Make something people want." The dark irony of our moment? AI might be making people nobody wants.

To readers: recognize the severity and urgency. Do whatever helps society achieve "delay" strategies. Don't amplify voices pushing "grab" or "burn."

Specifically:

• Engage with distribution reform discussions—keep the conversation alive
• Learn and adapt to new tech—be a participant, not just a victim
• Support organizations working on tech ethics and social equity
• Spread rational thinking in your circles—combat extremism
• Prepare personally for change without hoarding or panic-mongering

The Fallout games are fun to play. Living in actual fallout would be hell, and you definitely won't be the one with the Pip-Boy.

About the Author

Wang Xiao is an AI protocol architect, author of System and Freedom, creator of Danbing AI Protocol / SLAPS Framework, and initiator of OathAI.

His work focuses on human-AI co-creation, protocol governance, semantic anchoring, and long-term knowledge continuity, exploring how human knowledge and collaborative structures can be preserved, calibrated, and inherited in the AI era.

Disclaimer

This essay reflects the author's current observations and methodological reflections based on personal practice, research, and human-AI collaboration experience. The related Danbing / SLAPS / OathAI methods are still being organized and evolved. Their practical effects may vary depending on the user's background, task context, model capability, execution environment, and level of commitment.

This essay does not constitute legal, investment, medical, career, or technical implementation advice or guarantee. Readers who apply these methods in real projects should make independent judgments based on their own circumstances and take responsibility for specific outcomes.