Beyond Zero-Sum Fears: How Coase’s Frictions Reveal AI’s Job-Creating Innovation Loop

Behavior, Economy, Society, Technology, Uncategorised |

Hook & Thesis: Marking the Midpoint in the Machine

It is October 31, 2025, and the air feels thick with possibility—or peril, depending on the headline you last skimmed. Scroll through the news feeds, and you’ll spot the usual suspects: a factory worker in the Midlands made redundant by a clever algorithm, a London office buzzing with talk of AI rewriting job descriptions overnight. These stories stir an old unease, the kind that whispers this time it’s different. Artificial intelligence isn’t just another tool; it’s a mirror reflecting our deepest worries about work, worth, and what comes next. Born in 1971 and shaped by the UK’s steady rhythm of change—from the hum of Thatcher-era factories to the click of early internet modems—I’ve watched these waves rise and fall. Lately, a quiet question has pulled me back to Ronald Coase, that understated economist whose 1937 paper asked why firms even exist if markets are meant to be so slick. His answer? Markets come with hidden snags, costs that make going it alone a slog. Today, as AI smooths those snags, we’re not staring down a cliff but stepping into a familiar groove.

Picture it like this: you’re at a bustling market stall, haggling over a basket of apples. The back-and-forth eats time—spotting the ripest fruit, agreeing on a price, trusting it’ll last the walk home. Now imagine an app that scans the stall from your phone, predicts the best pick, handles the payment in a blink, and even suggests a recipe based on your fridge. That’s AI in a nutshell: not magic, but a shortcut through the everyday muddle. Yet the fear lingers—will those shortcuts leave the stallholders jobless? The short answer is no, not in the zero-sum way we dread, where one person’s gain is another’s loss. Instead, what we’re living through is a repeatable pattern, an innovation cycle that has played out before, from steam engines to smartphones. It starts with hype, builds into real tools, hits excess, resets, and spins again—each turn creating more jobs than it claims, if you know where to look.

This isn’t blind optimism. It’s a lens borrowed from Coase, who saw economies not as perfect machines but as human patchworks riddled with frictions: the small, stubborn costs of searching, bargaining, and enforcing deals that make trade feel like wading through treacle. Add in the drag of regulations or even our own hesitations—like buying extra insurance just to quiet a nagging worry—and those frictions stack up, often claiming nearly half of what an economy produces, according to old tallies from economic historians. AI changes that. It shrinks the production costs of making things (by automating the rote) and dissolves the trade frictions (by matching needs without the fuss), leaving room for fresh inventions. Through this view, the cycle isn’t chaos; it’s inevitable, a pulse that clears dead wood to plant new groves. Displacement happens, yes—the reset stings—but it’s the seedbed for reinvention, turning laid-off roles into sparks for roles we haven’t named yet.

Consider the thread that led me here: a simple nudge about Coase’s insight, unpacking why a corner shop or a tech giant bundles work under one roof to dodge market hassles. From there, it unravelled into talks of revenue splits—raw making costs plus those sneaky transaction tolls equalling slim profits—and how giants like Apple or Amazon scale by turning frictions into their secret sauce. Layer on AI, and the pattern sharpens: it’s compressing those tolls while rewriting the rules of exchange, much like crypto’s smart contracts hinted at in an earlier essay of mine from March. That piece mapped the cycle’s stages in AI and blockchain, but now, with Coase’s gaze, it reveals why jobs don’t vanish—they evolve. The thesis here is straightforward: beneath the zero-sum dread of AI erasing livelihoods lies this timeless loop, illuminated by market frictions. As AI collapses them, reinventions speed up, flipping creative destruction from a threat into a quiet job forge. It’s a map for the uneasy, showing we’re midway through the buildout, with abundance on the far side.

This pattern isn’t abstract history; it’s the path under our feet. To see it clearly, let’s trace its shape across time, spotting the echoes in engines long cooled and screens now cracked.

The Pattern Unveiled: Innovation Cycles as History’s Repeatable Heuristic

Every big shift starts the same way: a spark catches, and eyes widen. That’s the hype stage, where ideas fizz like fresh ginger beer—exciting, fizzy, but not yet quenching real thirst. Think of the 1760s in Britain’s textile towns, where steam whispers promised to spin wool faster than hands ever could. Folks gathered in smoky inns, buzzing about machines that would free workers from dawn-to-dusk looms. Fast-forward to our screens today: 2023 brought the AI gold rush, with chatbots drafting emails and artists fretting over generated sketches. It felt new, urgent, like the world tilting. But hype is just the opener, a collective gasp that draws dreamers and doubters alike.

Then comes buildout, the meaty middle where promises harden into tools. Machines get oiled, kinks ironed out, and the world starts to hum a bit louder. In the late 1700s, James Watt’s improved steam engine didn’t stay a parlour trick; it puffed into factories along the Thames, powering pumps that drained mines and looms that wove cloth at speeds once unthinkable. Output soared—cotton production tripled in a decade—and suddenly, a mill hand who once toiled solo could oversee a room of clattering frames. We’re in this phase now with AI, October 31, 2025, marking a pivot point. Global spending on the tech’s backbone—servers, chips, software—tops 375 billion pounds this year, wiring it into everyday work. A solicitor in Manchester uses it to sift case files in hours, not days; a farmer in Norfolk deploys drones that predict crop woes before they wilt. It’s not flawless—glitches pop like bubbles in that ginger beer—but it’s building, layer by steady layer.

Excess follows, inevitable as overindulgence at a feast. Tools multiply, prices plunge, and suddenly everyone’s got one, stretching them thin. By the 1820s, steam had flooded Britain’s workshops, churning out so much that warehouses overflowed and mills idled. Workers marched in protest, smashing the very pistons that had lit their nights with factory glow. The dot-com boom of the late 1990s played a similar tune: websites bloomed like digital weeds, valuations soared on vapour, until the 2000 crash pruned the lot. Pets.com folded, but the bones—secure payments, vast connections—fed Amazon’s rise. Today’s AI excess lurks around the corner, perhaps by 2027: pilot projects everywhere, but half-baked bots churning errors, firms chasing trends without roots. The signs are there—executives ramping investments, yet returns patchy as a winter lawn.

Reset is the quiet ache, the cull that hurts most. Jobs shift, illusions shatter, and the weak links snap. Post-steam, hand-weavers scattered, their skills obsolete amid the roar. The internet purge axed thousands in web design sweatshops. With AI, we’ll see it too: routine coders or data entry clerks eased aside as agents handle the grind. But here’s the heuristic’s grace—this isn’t the end; it’s the breath before rebirth. Each cycle loops back to hype, shorter and sharper than the last. Steam’s churn spanned decades; the web’s, barely ten years. AI? Five to seven, if patterns hold, propelled by code that learns and adapts. Why the speedup? Lower hurdles to entry—once, building a steam rig took fortunes; now, a laptop spins an AI prototype overnight.

Spot the repeatability in Britain’s own scars and successes. Post-Brexit, trade frictions spiked—customs forms piling like autumn leaves for exporters—yet fintech firms like Revolut wove apps that smoothed payments across borders, hiring thousands in compliance tweaks and fraud watches. That March essay I penned on AI and crypto sketched this loop in blockchain’s gleam: hype around tokens, buildout in efficient ledgers, excess in scam coins, reset in crashes, then fresh shoots like AI-tuned networks that idle computers into shared power grids. It’s the same dance, frictions thinning with each turn, inviting more players. A heuristic like this isn’t a crystal ball but a compass: predict the pulse, and the stumbles feel less like falls, more like steps in a well-trodden path. Knowing the stages—hype’s thrill, buildout’s graft, excess’s wobble, reset’s trim—lets us brace without breaking.

Yet patterns need a why, a motor beneath the rhythm. Enter Coase, whose quiet gaze on market snags explains not just the how, but the pull forward.

Coase’s Lens: Frictions as the Invisible Engine of Collapse and Reinvention

Ronald Coase wasn’t one for grand proclamations. In 1937, as a young scholar wandering American factories on a fellowship, he jotted a puzzle: if markets coordinate everything through prices—like an invisible hand waving suppliers and buyers into line—why bother with firms at all? His reply was plain as day: markets aren’t free; they’re laced with frictions, those everyday rubs that make deals drag. Searching for the right partner chews hours. Bargaining over terms sparks rows. Enforcing promises demands lawyers or ledgers. Bundle it up, and these transaction costs—Coase’s term—swallow a hefty slice of effort, often nearing half of what an economy churns out, as later historians like Wallis and North measured across growing nations. Firms arise as shelters, walls against the wind, internalising the fuss to keep things humming.

Fast-forward, and those frictions haven’t vanished; they’ve just dressed in new clothes. Information trickles unevenly—one side knows the apple’s bruises, the other doesn’t—breeding mistrust. Opportunism lurks: a supplier skimps on quality once the cheque’s cashed. Regulations layer on, from labour rules to environmental checks, each a necessary guard but a toll nonetheless. Even our minds add drag—buying gadget insurance not for the odds, but the knot in the gut over what-ifs. Picture revenue as a pie: one wedge for production (steel, wages, the making), another for these frictions (the haggling and hedging), leaving crumbs for profit. For a small UK bakery, that friction wedge might gobble 40 per cent—sourcing flour amid price swings, auditing books to dodge fines, leasing ovens with endless clauses. Scale helps giants like the tech titans nibble it down, but for most, it’s a quiet squeeze.

Here’s where AI enters, not as a wrecking ball but a solvent, dissolving those snags while reshaping the board. Start with production: machines that once needed armies of operators now learn from data, tweaking outputs on the fly. A car plant in Coventry stamps panels with robotic arms guided by patterns no human could spot, slashing material waste and shift overtime. But the real wizardry hits the friction side. Information flows? AI scans vast troves in seconds—matching a Birmingham exporter’s widgets to a Berlin buyer without the phone tag. Bargaining eases into algorithms that predict fair splits, like an app suggesting lease terms based on market whispers. Enforcement? Smart contracts on blockchains lock payouts to milestones, no court needed. Even the mind’s tolls bend: tailored alerts flag real risks, curbing impulse buys on warranties that gather dust.

To see it sharp, consider this table, a snapshot of frictions before and after AI’s touch, drawn from the cycle’s churn:

Friction TypePre-AI DragAI CompressionCycle Impact
Information FlowsHours hunting suppliers or dataPredictive matches in blinksHype swells faster with easy access
BargainingTense talks over prices or termsAutomated bids and fair splitsBuildout scales without stalled deals
EnforcementAudits, lawyers chasing complianceSelf-policing agents and ledgersExcess builds on trust, not trials
PsychologicalFears driving over-insurance or upgradesNudges matched to real oddsReset clears emotional deadweight
RegulatoryForms piling for trade or safetyAuto-filers navigating red tapeReinvention blooms in freed time

Each row shows collapse leading to lift-off. In the UK, Brexit turned frictions into a bonfire—exporters facing 20 to 30 per cent hikes in paperwork, delaying shipments like traffic on the M25. AI douses it: tools now auto-complete customs declarations, spotting errors before they snag. A firm that once hired clerks for the slog redirects them to scouting new markets, turning drag into drive. Globally, this could swell trade by nearly 40 per cent by mid-century, as bodies like the World Trade Organisation forecast, by making borders feel like suggestions, not barriers.

Coase glimpsed this in his later work, the 1960 theorem: when frictions fade and rights are clear, folks bargain to wins no matter who starts ahead—like neighbours hashing a fence without councils. AI lowers that bar further, not by erasing humans but by clearing the clutter. Speculate a touch: in five years, a solo artisan in Leeds might orchestrate global sales via an AI that handles the noise—translating pitches, timing bids, even soothing client qualms with chatty bots. No firm needed; just a laptop as the new workshop. This isn’t erasure; it’s expansion, frictions thinning to let more enter the fray.

The inevitability shines here: lower costs mean quicker tests, wilder ideas, tighter loops. What took steam decades—rewiring society—AI condenses to years, each cycle shedding old frictions like snakeskin. From crypto’s ledgers optimising idle rigs, as I explored in that spring essay, to agentic systems prototyping trades in simulations, the engine revs. But comfort comes in knowing destruction serves creation, not spite. Let’s turn to that flip next, where jobs don’t just survive—they multiply in the cleared space.

Creative Destruction as Job Creator: Flipping the Zero-Sum Script

Joseph Schumpeter called it creative destruction back in the 1940s, that economist with a flair for the dramatic: old ways crumble under new ones, not out of malice but momentum. A horse-drawn cart yields to the motorcar, shoemakers to mechanics. It sounds brutal, zero-sum—like the gain in speed exacts a toll in lost livelihoods. Yet peel back, and it’s more relay than rout: the baton passes, runners multiply, the race stretches longer. With AI, this script flips from fear to forge, the innovation cycle’s reset not a grave but a greenhouse, sprouting roles in soil enriched by what’s cleared. We’re not trading jobs; we’re trading tasks, the drudgery for discovery.

Take the steam age again, that well-worn UK tale. Handloom weavers, once the backbone of Lancashire mills, watched threads fly from machines in the 1820s, their cottages emptying as output quadrupled. Protests flared—the Luddites smashing frames in moonlit raids—yet by 1900, railways alone employed one and a half million, from track layers to ticket clerks, roles unimaginable amid the wool dust. Destruction pruned the solo craft; creation wove networks, pulling folk into towns with wages that bought more than bread. The net? Ten million jobs gained across Britain’s shift, a swell not despite the churn but because of it. Patterns echo: the internet’s 2000 purge culled web monkeys peddling flashy sites, but birthed two and a half million in creative fields by 2010—content shapers, app builders, the invisible hands of e-commerce.

AI follows suit, easing frictions to unlock, not lock out. Routine bites—the endless data sorts or contract scans we tied to Coase’s tolls—get gobbled by agents that hum in the background. A warehouse picker in Swindon, once chasing aisles for hours, now directs drones with voice commands, eyes free for spotting efficiencies. That shift isn’t loss; it’s lift, per studies showing 30 per cent of daily tasks handed off, leaving space for oversight and tweaks. In pharma, multimodal AI halves research timelines, virtual trials replacing lab marathons—not axing scientists, but arming them to chase cures once sidelined by paperwork. PwC’s outlook pegs 20 to 30 per cent revenue jumps for adopters, each pound spawning design or ethics roles we haven’t tallied yet.

Look closer at the UK’s own reinventions. Post-Brexit, fintech outfits like Revolut stared down friction walls—compliance mazes that once devoured teams. AI smoothed them: auto-audits catching slips, predictive models flagging fraud before it bites. Result? Five thousand new hires since 2023, not in ledgers but in strategy—crafting products for a borderless Europe, training the very bots that freed them. It’s augmentation, not automation’s endgame: low-skill spots bloom as tools extend reach, like a nurse in the NHS using AI diagnostics to tend more patients, not fewer. Historical nets hold—steam’s displace-to-employ ratio mirrored the web’s, and early AI pilots whisper the same: gains in abundance outpace the prunes.

The zero-sum myth crumbles under this light. Cycles net positive because destruction targets deadwood—the repetitive that numbs—while creation fills with what humans do best: connect, improvise, care. Imagine a reset by 2028: excess AI flops cull overhyped startups, but survivors seed green hybrids, blending code with climate smarts. In Yorkshire valleys, wind farm techs once bound by manual logs now orchestrate turbines with glance dashboards, roles swelling as exports green the grid. Economists like Jason Furman note AI averting stagnation, propping growth that hires where it hurts most. It’s no promise of ease—the reset aches, retraining stings like a stiff walk after rain—but the loop’s predictability comforts: hype draws the bold, buildout builds the base, excess tests the true, reset renews.

This flip isn’t theory; it’s the path’s quiet gift, frictions collapsed to let light in. As we embrace the loop, fear fades to footing, pointing toward a forward rich with reinvention.

Close: Embracing the Loop—From Fear to Forward

We’ve traced the thread from Coase’s market snags to AI’s smoothing touch, spotting the innovation cycle’s steady pulse beneath the noise. Hype ignites, buildout beds in, excess swells, reset refines—and round again, frictions thinning each lap to hasten the turn. Displacement isn’t zero-sum erasure; it’s the cycle’s breath, clearing space for jobs that stretch our grasp, from algorithm tenders to trade weavers in a friction-light world. Born amid Britain’s own pivots, this feels less like forecast than familiarity: we’ve walked these woods before, emerging with maps redrawn.

For the UK, the nudge is plain—lean into light-touch policies that foster the forge, not fetter it. Ease the regs that once bloated Brexit’s tolls, let scale-ups like those fintech pioneers experiment with AI’s edge. Everyday readers, new to these turns: watch your own patch. Spot the hype in a colleague’s tool trial, the buildout in a workflow’s quiet hum. When reset looms, it’s not the end but the edit, inviting you to the next stage’s spark.

By 2030, loops tighter still, abundance might look like neighbourhoods linked in shared compute webs, roles blooming in niches we name on the fly. In Coase’s unassuming insight and the cycle’s even turn, we find not mere comfort, but a compass: fear the fall less, map the rise more. The machine midway hums on—join the rhythm.

References

This section gathers the key sources woven into the essay, from timeless economic insights to fresh 2025 snapshots. Each entry includes a brief note on its fit, ensuring the thread from Coase’s market snags—unpacked in our earlier exchanges on firms as friction hedges—to AI’s cycle-spinning role stays grounded. URLs link directly to accessible texts where possible; for seminal works, I’ve prioritised open PDFs or stable archives.