Cathedrals versus Commons

The best pranks take ten thousand years to set up.

That's what the Dwellers do in Banks' The Algebraist: beings who live for eons in gas giants, playing elaborate jokes on each other because when you have infinite resources, reputation through novelty is all that matters. One Dweller breeds a sentient species just to have an audience for their poetry. Another spends centuries setting up a pun.

They would find our current situation hilarious.

Consider the hospital administrator's daily reality: You're running critical infrastructure on OpenAI's API while your radiologists use sketchy Discord models on the side. Your lawyers demand compliance certificates. Your engineers contribute to repositories that will make your product obsolete. Your board wants quarterly growth. Your mission statement talks about serving humanity. Three different economic games, incompatible rules, everyone pretending this is normal.

The real puzzle: what kind of chaos we're dealing with. In physics, three bodies orbiting each other create a specific problem: no stable solution, just permanent dynamic instability. The system never settles. It never resolves. It just continues, each body pulling the others in incompatible directions.

That's exactly where we are with AI. Not a simple binary, cathedrals versus commons, extraction versus sharing, but something more vexing.

Extraction logic says monetize scarcity. OpenAI turned matrix multiplication into a $100B valuation by adding safety theater and enterprise SLAs. Beautiful. Pays the bills. Funds the yachts. Cathedral builders with their digital priesthoods, explaining why you need their blessing to run inference. "Too dangerous to release," they say, while teenagers jailbreak it with ASCII art.

Performance logic says earn attention through novelty. This is the Dwellers' game—when nothing is scarce, status flows to whoever's most interesting. It's already here: GitHub stars, citation counts, Twitter ratios. Researchers racing to publish. Engineers building in public. That kid who made GPT-4 think it was a pirate and got 10,000 retweets. Value doesn't flow to ownership but to whoever makes the timeline laugh.

Service logic says forget both games and help humans. This is what actually runs hospitals, schools, courts. Not profit (extraction) or pranks (performance) but purpose. A nurse doesn't need viral tweets or stock options. They need patients to heal. The value isn't captured or performed—it's created through giving a damn.

What looks like a simple problem from one angle (choosing between profit and purpose) reveals itself as something else entirely when you step back. You can't serve all three masters without betraying two. That's not a management challenge. That's physics.

Open-source your tools (performance) and shareholders scream about competitive advantage (extraction). Focus on profits (extraction) and watch your best engineers leave for companies with better GitHub profiles (performance). Optimize for efficiency (extraction) and compromise the care that made you start this company (service).

The companies that matter, the ones digitizing courts, powering schools, keeping hospitals from collapsing, they're not choosing. They're juggling chainsaws while solving differential equations on a tightrope during an earthquake. Playing extraction to fund operations, leveraging performance to attract talent, maintaining service because that's why they exist.

And then AI arrives and accelerates everything.

Models leak faster than companies can capture them. Every wall becomes Swiss cheese. Meta gives away Llama just to watch OpenAI squirm. Chinese companies drop open weights that match GPT-4 because why not. Teenagers fine-tune models to be maximally unhinged and share the LoRAs with friends. The Dwellers would approve—it's all becoming very silly.

From Two-Body to Three-Body Chaos

We thought this was a two-body problem. The pattern seemed familiar enough: cathedrals versus commons, extraction versus sharing, enclosure versus openness. OpenAI builds walled gardens; Meta releases Llama. Microsoft licenses everything; Hugging Face distributes everything. David versus Goliath, if Goliath were a language model and David a kid with a gaming rig.

It's a story we've told before. The music industry built cathedrals around CDs; Napster tore them down. Film studios erected paywalls; BitTorrent tunneled under them. Software companies sold licenses; open source gave them away. Each time, the same arc: proprietary control gives way to distributed abundance, monopolies dissolve into commons, the cathedral crumbles and the bazaar flourishes.

That's the upbeat scenario, anyway. The Dwellers would find it charming—these mortals, thinking they've seen this movie before.

This isn't a repeat. The question is what's different this time. The history we're relying on was always incomplete. We mapped two forces and called it physics. Cathedral builders extracting rents versus commons builders earning kudos. Clean. Binary. Wrong.

There's a third body nobody mentioned: Service. The boring one. The one that keeps humans alive. While we spent years debating monetization versus memeification, extraction versus reputation, priests versus pranksters, someone still needed to run the hospitals. Teachers still needed to educate children. Courts still needed to dispense justice. These institutions don't operate on extraction logic or performance logic. A nurse doesn't need GitHub stars or quarterly returns. They need patients to survive and software that doesn't crash during surgery.

This changes everything. Two-body problems find stable orbits—one wins, or they balance. Three bodies create chaos with no stable solution. Every decision breaks two other systems, every optimization causes new problems. Nothing ever resolves.

Everyone's navigating all three at once. OpenAI talks extraction to investors, performance to researchers, service to regulators: three different stories, all true, all incompatible. Meta weaponizes openness (that's not kudos, it's warfare) while extracting through ads. Hospitals serve patients while private equity extracts value while competing for talent through performance. Three-dimensional chess. Everyone pretending it's checkers.

The Three Value Systems

Economics has three incompatible physics. Like gravity, magnetism, and whatever dark energy is doing. They don't play nice.

Extraction logic is gravity—heavy, inevitable, pulls everything toward money.

OpenAI turned matrix multiplication into $100B. It's beautiful. They took something that's basically fancy Excel, added safety theater ("too dangerous to release!"), enterprise SLAs ("someone to sue!"), and boom: digital priesthood. Cathedrals everywhere. Stark and Bainbridge would applaud; they're selling compensators for things that don't need compensating. "This model is aligned" (you can't verify this). "Trust our safety team" (who chose them?). The logic is brutal: find something people need, build a moat, charge rent. It's what pays your mortgage.

Performance logic is magnetism—pulls attention, creates fields of weird.

This is what the Dwellers figured out after a million years: when nothing is scarce, you compete on being interesting. The kid who made GPT-4 speak only in haikus got 50,000 GitHub stars. Researchers literally race to ArXiv at midnight to claim priority. Engineers building their entire careers in public, every commit a performance. It's not about owning but about others wanting to pay attention. Value flows to whoever makes immortals laugh. Or at least makes Twitter laugh, which is close enough for mortals.

Service logic is dark energy—mysterious, essential, constantly fighting the other two.

This is what actually makes hospitals work. What keeps schools teaching. What prevents courts from collapsing into chaos. Not profit margins or viral tweets, just humans giving a damn about other humans. A nurse doesn't need kudos or revenue. They need patients to heal. The value isn't captured or performed; it's created through consistent, unsexy care. It's the physics nobody talks about at TED but everyone depends on to not die.

Service operates on different time scales. Extraction wants quarterly returns, performance wants viral moments, service wants generational impact. A good teacher's work shows up 20 years later when their student becomes a doctor. You can't A/B test compassion. Care doesn't scale linearly. Quality often inversely correlates with efficiency metrics.

Service creates different dependencies: not vendor lock-in (extraction) or audience capture (performance), but trust networks, community knowledge, and institutional memory. These can't be forked, purchased, or performed. Service logic is why that one nurse who knows everyone's name matters more than the million-dollar EMR system. It's why the school janitor who's been there 30 years IS the institutional memory. These humans aren't performing or extracting; they're holding civilization together with consistent, unsexy care.

Every organization is being ripped apart by these three forces simultaneously.

Picture a public company serving critical infrastructure. Monday: extraction logic for the earnings call. Tuesday: performance logic to recruit that Stanford PhD. Wednesday: service logic because someone actually needs help. Thursday: extraction again because payroll is due. Friday: performance because your best engineer is building a competitor unless you let them open-source something. Weekend: service because the mission statement says you care.

It's like being drawn and quartered, but in three dimensions, and everyone's livestreaming it.

The systems weaponize each other: Everyone's using one system against the others.

Meta "gifts" Llama to the world. How generous! Except it's not performance (they don't want kudos); it's extraction warfare to burn down everyone else's moat and protect their advertising empire. "If we can't monopolize AI, nobody can." This isn't participating in any kudos economy; it's using "free" as ammunition within pure extraction logic. Zuck learned from mobile platforms: never again be someone else's tenant. So now he's torching the whole neighborhood while pretending it's philanthropy.

Google open-sources everything. Not for reputation, but to make sure search stays profitable while everything else burns. They're not seeking approval from the Dwellers or building reputation through novelty. Every "gift" to the commons is actually a competitive move to prevent anyone from building a moat that might threaten their core extraction machine. Universities demand publication (performance) while patenting everything worth money (extraction). Hospitals advertise patient care (service) while private equity guts them for parts (extraction).

Everyone's playing 3D chess while pretending it's checkers. The Dwellers would find this hilarious: all these serious companies with serious valuations, trying to navigate incompatible physics while pretending they have a strategy. Like watching someone try to serve three gods who hate each other (at machine speed, with quarterly earnings calls).

AI is pouring gasoline on this comedy. Information wants to be free (physics favors performance). But capitalism needs scarcity (economy demands extraction). And humans need systems that actually work (society demands service). The three-body problem isn't stabilizing. It's going fractal.

The Navigation Challenge

You're not choosing between systems. You're surfing between them:

• Extraction pays your bills but kills your soul
• Performance attracts talent but doesn't pay rent
• Service fulfills purpose but won't fund itself

Every organization is being pulled apart by incompatible physics. The question isn't which one to choose—it's how to navigate the chaos.

The Straightjacket Paradox

The bind: You run a company that matters. Maybe you digitize courts, or power schools, or keep hospitals running. The infrastructure society actually needs. You're not selling ads or optimizing engagement. You're doing the boring, critical work that keeps civilization functioning.

You're also trapped in cathedral economics. Listed on an exchange, or venture-backed, or just trying to make payroll. You have shareholders who expect returns. Customers who demand compliance certificates. Regulators who require audit trails. You can't just open-source everything and trust the universe. You have quarterly earnings calls where "we're building toward the kudos economy" doesn't fly.

But you can feel the shift coming. You see your brightest engineers contributing to open source on weekends. Your customers' kids are running local models that outperform your licensed APIs. The abundance is leaking in through every crack, and you know cathedral logic has an expiration date. You just can't afford to be first.

This is most companies' reality. Caught between economic systems. Playing cathedral because that's what pays the bills, while knowing kudos is where value is moving. It's like being a travel agent in 1995, watching people discover Expedia. You know what's coming. You just have mortgages to pay in the meantime.

You can lean toward kudos while wearing the cathedral costume. Open source your non-core tools. Build in public where you can. Share knowledge that doesn't compromise competitive advantage. Create internal kudos economies—reward employees for teaching, for sharing, for making others better. Run hackathons. Fund fellowships. Build APIs that others can build on.

You're not abandoning cathedral economics—you can't afford to. But you're building kudos muscle memory. Training your organization to function in abundance logic. Creating optionality for when the flip happens. Because it will happen. The only question is whether you'll be ready to surf it or get crushed by it.

Navigation Strategies (The Art of Surfing Chaos)

You can't solve a three-body problem. You can only surf the chaos. Here's how organizations are learning to juggle flaming chainsaws while riding unicycles on tightropes.

Strategy 1: The Octopus Play

Smart companies aren't choosing; they're growing tentacles. Core operations run on extraction (pays rent), innovation labs run on performance (attracts talent who'd otherwise start competitors), and community programs run on service (keeps the pitchforks away). Each tentacle serves a different god, and somehow the octopus survives.

Healthcare platforms do this: Enterprise contracts with insurance companies (extraction), because money. Open-source data visualization tools (performance), because their engineers threatened to quit. Free clinics in underserved communities (service), because someone remembered why they started this company. It's like maintaining three separate personalities, but at least the board is confused enough to approve everything.

Strategy 2: The Time-Lord Shuffle

Some organizations are literally time-traveling between value systems. They're not serving all three masters simultaneously; they're cycling through dimensions like Doctor Who with an MBA.

Monday through Thursday: Pure extraction. Sales calls, feature bloat, enterprise theater. Everyone wears suits and says "synergy."

Friday: Performance day. Engineers finally allowed to commit to open source. Researchers racing to publish before competitors scoop them. Twitter threads about technical breakthroughs.

Weekends: Service mode. Hackathons for nonprofits. Free workshops. Actually helping humans.

It's exhausting. Everyone needs therapy. But it works until someone realizes they're living three different lives.

Strategy 3: The Platform Paradox

Build APIs and let others fight the three-body problem. Become the Switzerland of software: it's strategic genius.

Stripe figured this out: Charge transaction fees (extraction), but make developers love you (performance), while enabling entire economies (service). They're not choosing; they're creating interfaces between incompatible physics. Every value system can plug in, fight it out in userspace while Stripe counts money.

It's brilliant navigation: you don't fight the chaos, you become the chaos others navigate through. The Dwellers would approve.

Strategy 4: Quantum Superposition

Never collapse into a single state. Tell investors about recurring revenue. Tell engineers about open standards. Tell customers about mission-driven values. All simultaneously true, all simultaneously incompatible.

A court digitization company exists in superposition: They're disrupting justice (extraction) while democratizing justice (service) while advancing justice (performance). Schrödinger's startup. The ambiguity isn't weakness—it's quantum strategy. The moment you observe them clearly, they've already shapeshifted.

Banks' Culture Minds would recognize this immediately: tactical opacity as survival mechanism.

Strategy 5: The Split-Personality Solution

Literally create different legal entities for each physics. It's organizational schizophrenia, but with tax benefits.

Mozilla did this for decades: Corporation for extraction (paychecks), Foundation for service (mission), Community for performance (actual product). Three organizations pretending to be one, constantly fighting, occasionally cooperating, and somehow shipping a browser.

It's messy. Lawyers love it. The IRS is confused. But it creates space for incompatible systems to coexist, like a corporate version of the three-body problem where nobody's orbits intersect long enough to collide.

Strategy 6: The Exhaustion Exit

Some organizations are just... stopping. Not pivoting or adapting—just admitting defeat. Local newspapers. Independent clinics. Community colleges. They can't navigate three physics simultaneously, so they're choosing managed decline over impossible gymnastics.

It's not a strategy; it's acknowledging that some problems don't have solutions, only endings. They tried serving all three masters and realized it was killing them faster than just accepting irrelevance. Watch a local newspaper stop chasing digital performance metrics and just print obituaries for the seventeen people who still subscribe. There's a dignity in giving up when the game is rigged against you.

The Meta-Strategy: Embrace the absurdity.

These aren't solutions—they're temporary hacks in an unstable universe. What works today fails tomorrow. The three-body problem doesn't resolve; it just gets weirder. Organizations that survive don't have better strategies; they just adapt faster when reality shifts.

Build sensing mechanisms (your youngest engineer's Discord server). Maintain optionality (can you pivot all three physics in 30 days?). Get comfortable navigating permanent instability (meditation helps, alcohol doesn't).

The Dwellers spent a million years figuring this out. We have quarterly earnings calls. We're speedrunning impossible physics with lawyers watching.

The Pattern Is Breaking

Cathedrals Used to Win (When friction existed):

• Railways → needed physical track (monopolized)
• Telegraph → needed physical wires (Western Union)
• Telephone → needed physical switches (AT&T)

Then Physics Changed (Digital goods have no friction):

• Music industry → Napster → BitTorrent → Spotify admits defeat
• Film/TV → Piracy → Netflix → Everyone has a streaming service
• Software → Proprietary → Open source ate everything
• Phones → Locked down → Jailbroken → Android is just Linux

AI Is Speedrunning the Pattern:

• GPT-3 (2020): "Too dangerous to release"
• Stable Diffusion (2022): Running on gaming PCs
• LLaMA (2023): Leaked, then officially open
• Mixtral (2024): Better than GPT-3.5, completely open
• Now: Kids fine-tuning models on consumer GPUs

The time from "revolutionary and proprietary" to "commodity running on a potato" is approaching zero. Physics favors abundance, and cathedrals are building sand castles at low tide.

National Navigation (How Countries Surf the Chaos)

Countries face the same three-body problem as companies—but with nuclear weapons, pension funds, and citizens who vote. Giants juggling planets while their feet are on fire.

Singapore: The Minmax Masters

They're playing all three games with spreadsheet precision. Government contracts with global tech giants (extraction) fund public AI research (performance) which powers free literacy programs (service). It's a carefully orchestrated dance where every move serves three purposes.

Lee Kuan Yew would be proud; they've turned the three-body problem into an optimization equation. Use extraction revenue to fund performance. Use performance to attract more extraction. Use both to deliver service so efficiently that nobody questions the system. It's brilliant until someone realizes they're living in an Excel formula.

The Dwellers would find this adorable: mortals trying to solve chaos with math.

The Nordics: Too Nice for This Timeline

High-trust societies trying to navigate low-trust physics. They're playing service mode on hardcore difficulty while everyone else is exploiting bugs.

Their solution is geographical segregation: Sovereign wealth funds do the dirty extraction work offshore. Universities handle performance (published papers, open research). Daily life runs on service logic (healthcare, education, human dignity). They've literally built different rooms for incompatible physics.

It works because they're rich enough to maintain the walls. But watching them try to compete with Silicon Valley is like watching hobbits enter a cage fight. They're too decent for this timeline. They keep trying to help while everyone else is extracting.

China: Schrödinger's Superpower

They exist in quantum superposition: simultaneously all states until observed, then whatever's most advantageous.

Building the world's most invasive surveillance state (extraction) while flooding GitHub with open models (performance). Lifting millions from poverty (service) while disappearing billionaires who get too loud (extraction enforcement). They're not choosing between systems; they're running all three in parallel universes that occasionally intersect.

They release Qwen to break Western AI moats while using the same tech for social credit scores. It's not hypocrisy; it's superposition. They'll be whatever value system wins because they're already playing all of them. The Culture Minds would respect this—tactical ambiguity at civilizational scale.

The US: Chaos Incarnate

America doesn't navigate the three-body problem; America IS the three-body problem.

Venture capital pursues pure extraction ("unicorn or die"). Silicon Valley performs elaborate innovation theater ("we're changing the world" while building ad networks). Public institutions desperately try to maintain service while being systematically defunded. The systems don't coexist; they're in active warfare. Congressional hearings where nobody understands what they're regulating. Billionaires buying media companies for fun. Cities with trillion-dollar tech companies and homeless encampments.

The chaos creates antifragility through evolutionary pressure. When extraction becomes too predatory, performance creates alternatives (open source everything). When both fail, service institutions somehow shamble forward (the post office still exists, somehow).

It's governance by Darwinian combat. The Dwellers would love this: maximum entropy as a political system.

The Real Sovereignty Question: Can we navigate all three value systems while our rivals are trapped in one?

Countries that lock into single systems become predictable, exploitable. Pure extraction states become brittle. Pure service states can't compete. Pure performance states can't eat. The winners will be those who can shift between systems, play multiple games, maintain optionality across incompatible physics.

No country has solved this. Everyone's improvising. Singapore looks smooth but struggles with innovation beyond government direction. The Nordics seem stable but can't scale. China appears coordinated but the contradictions are building. The US seems chaotic because it is.

We're all navigating the same impossible physics, just with different cultural defaults and political constraints. National AI sovereignty isn't about control; it's about navigation capability in permanent instability.

The Cool Country Test

Questions that matter in 2027:

• Can your citizens fork and modify any model that interests them?
• Is your population weird enough to generate economic surprises?
• Do you have infrastructure for million-year pranks?
• Are your agents more interesting than OpenAI's?
• Can your kids jailbreak faster than Silicon Valley can patch?

If yes: Welcome to the kudos economy, you magnificent weirdos.

If no: Enjoy renting your future from someone else's cathedral.

The Dwellers would judge you not on GDP, but on how interesting your citizens are when they have infinite compute.

The Window Is Closing

The race isn't to build the best cathedral. It's to make cathedrals obsolete before they notice they're dinosaurs.

We're in the Cambrian explosion phase. Everything is weird, nothing is settled, every model gets leaked or replicated within months. The kids fine-tuning models in their bedrooms don't know they're supposed to wait for permission. The researchers dropping papers on ArXiv don't know they're supposed to keep secrets. The abundance is already here—it just hasn't evenly distributed yet.

The accelerants are stacking up:

Meta will keep giving away frontier models just to watch OpenAI's valuation twitch. But let's be clear—this isn't generosity or kudos-seeking. It's pure extraction warfare. They don't need to monetize AI—they need to make sure nobody else monopolizes it so their advertising empire remains the only game in town. Zuck learned from the mobile platform wars. Never again be someone else's tenant. If burning down the entire AI industry protects their surveillance machine, they'll supply the matches.

China will keep dropping open models that match Western capabilities. Not for ideology—for leverage. Every open model they release makes Western cathedral logic harder to maintain. "Too dangerous to release" looks absurd when Alibaba's latest model is on Hugging Face.

The GPU shortage is ending. NVIDIA's competitors are catching up. Compute is getting cheaper, more distributed, weirder. Soon every gaming PC will be running models that would have required a data center last year. The substrate for abundance is materializing.

The real accelerant? The next generation doesn't believe in scarcity.

They grew up with infinite content, infinite connection, infinite information. They instinctively understand kudos economics because they've been living it—accumulating followers, building reputation, creating memes for attention rather than money. They treat AI like they treat everything else: a toy to hack, modify, and make do unexpected things.

A 16-year-old fine-tunes a model to be maximally unhinged. They share the LoRA with friends. Watch the modifications spread, mutate, evolve. They're not building cathedrals. They're playing. And play is how new economics gets born.

The window isn't closing; it's opening wider. Every leaked model, every open source release, every kid who learns to fine-tune makes the cathedral logic more untenable. The cathedral builders know this. That's why they're scrambling for regulation, for safety standards, for moats made of law rather than technology.

But you can't regulate physics. You can't make thermodynamics illegal. The future is already here, compiled and running on a million gaming rigs.

Wait, Is This Even Real?

What's actually happening here—and what that honesty costs us.

The critics of the kudos economy aren't entirely wrong. Take Meta's Llama release, celebrated by many as a victory for openness. What looks like gift-giving from one angle can look like arson from another. Meta isn't participating in any kudos economy; they're weaponizing openness to prevent monopolies that might threaten their advertising empire. Releasing Llama wasn't about earning reputation through novelty; it was about ensuring nobody could build a moat that might eventually compete with their surveillance machine. Zuck learned from mobile platforms: never again be someone else's tenant. That's not kudos. That's strategic cathedral warfare using "free" as ammunition.

Google follows the same playbook. They open source to destroy competitors' business models, not to earn kudos. Every "gift" to the commons is actually a competitive move within pure extraction logic. The performance is secondary; the extraction is what matters.

The institutions that actually matter? Hospitals, schools, courts? They're not playing either game. They run on duty, care, service, purpose—value systems that have nothing to do with either cathedral extraction or kudos performance. A nurse doesn't want reputation for clever pranks. They want patients to heal. A teacher doesn't need to impress immortals. They need kids to learn. Some problems don't have solutions, only maintenance. You don't "solve" healthcare. You don't "disrupt" aging. You don't "optimize" grief. You just show up, every day, and do the work.

Even the kids on Discord—are they really building a kudos economy? They're doing what kids always do: playing with powerful tools while adults panic about implications. How many are building LoRAs for fun versus building portfolios for YC applications? Half will become extraction machines themselves, the other half will burn out and become therapists.

The worry is far from fanciful. What if there is no transition? What if we're just entering permanent instability where different physics coexist uncomfortably forever?

The frameworks aren't perfect, but they help us see what the binary obscured. The three-body framing reveals something the binary obscured: that institutions serving human needs operate on logic fundamentally different from both market extraction and reputation performance. That's not a bug in the theory. That's the point.

What's actually scarce: care, attention, wisdom, judgment, responsibility. Not the performance of care for reputation, but actual give-a-damn. The things that make hospitals work. The things that make good teachers. These don't fit neatly into either extraction or performance logic—and maybe that's telling us something.

The Dwellers' economy works because they're immortal and bored. We're mortal and busy. We need healthcare that works, education that teaches, infrastructure that doesn't collapse. These needs don't care about our economic physics or our clever frameworks. They just need to work.

So maybe the real question is: In a world where AI makes information abundant, how do we organize the stuff that actually matters—care, wisdom, judgment, responsibility—without relying entirely on either extraction or performance?

Or, more fundamentally: Can we keep the hospitals running while the physics fight?

If there's one skill we can't afford to lose, it's the skill of distinguishing between the problems we can solve and the ones we can only navigate.

The Choice

We're not at a crossroads. We're in a three-way knife fight between physics, capitalism, and human needs. And capitalism brought lawyers.

The Dwellers would watch this and laugh for centuries. Here we are, on the edge of material abundance, arguing about API pricing while the models leak through every crack. Like watching someone dying of thirst debate water's business model while standing in rain.

Everyone's performing their assigned role in this comedy, and yet abundance emerges anyway.

Meta "gifts" open models—not for kudos but to burn down everyone else's moat while protecting ad revenue. The kids making LoRAs are building either YC applications or therapy practices (or both). Hospitals, schools, courts don't have time for economic physics—they're trying to serve humans while private equity strips them for parts and consultants optimize away their souls.

But information still finds cracks like water. Every wall gets holes. Every moat gets bridged by some teenager who needs the tool and can't afford the toll. OpenAI builds cathedrals. Meta bombs them. China floods the market. Some kid in Bangkok fine-tunes everything into chaos.

What's emerging: Doctors using sketchy open models to diagnose patients while maintaining compliance theater for lawyers. Teachers mixing corporate platforms with student-jailbroken AIs because it actually helps kids learn. Engineers building cathedrals by day, releasing tools by night that accidentally revolutionize industries they weren't even thinking about.

The companies that matter—the ones keeping civilization from segfaulting—they're jury-rigging bridges between incompatible physics. Using extraction to fund development, performance to attract talent, service to remember they're human. It's like building a plane while flying it through three different dimensions of spacetime with quarterly earnings reports.

Monday: extraction logic for payroll. Tuesday: performance logic for recruiting. Wednesday: service logic because someone needs actual help. Thursday through Sunday: repeat until heat death of universe or next funding round.

The next five years? Not a transition, but an acceleration of chaos. More models leaking, more walls becoming Swiss cheese, more teenagers treating AGI like a particularly interesting video game. Cathedral builders scrambling for regulatory moats while kids speedrun jailbreaks. Service institutions trying to help humans while being carved up by Excel formulas.

The Dwellers have million-year pranks because they transcended scarcity. We have quarterly earnings because we haven't transcended mortality. Banks gave us the vision—post-scarcity economics as elaborate comedy. Stross showed the trajectory—accelerating weirdness until nothing makes sense. Reality split the difference: abundance without wisdom, tools without clarity, capability without direction.

We're not immortal gas-giant dwellers with million-year attention spans. We're humans with mortgages and kids who need healthcare. So while the physics fight it out, while Meta burns down neighborhoods and OpenAI builds cathedrals and kids jailbreak everything... someone still needs to keep the hospitals running.

The only real question: Can we maintain the unsexy infrastructure of care while Silicon Valley optimizes for everything except what humans actually need?

The three-body problem doesn't resolve. It just continues. There's no choice between cathedrals and commons, no transition to kudos economics. We're improvising through permanent instability with tools we don't understand, serving systems that hate each other, while pretending we have a plan.

The only honest answer: Keep navigating. Keep the service people functioning despite the chaos. Let the kids jailbreak everything—they might accidentally fix something. And remember: Every system that claims to have solved this is selling you something, teaching you something, or serving you something. Sometimes all three at once.

The stakes are real. Every nurse who quits, every teacher who burns out, every local newspaper that closes—these aren't abstractions. They're the service logic bleeding out while extraction and performance fight over the corpse.

The Dwellers would laugh for centuries while we navigate quarterly earnings calls. Somewhere between those timescales, humans need help. Might as well get good at surfing chaos.