Who Gets the Value?
Date: 2026-11-21
Author: Wealth & Means Staff
Source: https://wealthandmeans.com/essay/who-gets-the-value
The word value can mean a price, a profit, a useful product, or a social benefit. Those things sometimes move together. They can also part ways. A buyer can get a better deal while a factory loses money; a busy company can sell more without earning a satisfactory return; a useful invention can create far more value than its maker captures. The accounting question is also a human one: who does the work, who pays the cost, and who keeps the gain? That question looks different in a research lab than it does on a factory balance sheet. It also appears when a creator sells an audience, when expert labor is divided into platform tasks, when craft traditions seek new markets, or when a tariff refund has to be distributed. Philippe Aghion and Keyu Jin's imagined debate asks how much public coordination can help an industry grow before it begins protecting the wrong thing. Ken Thompson's story offers a different kind of value: tools that became more useful because others could build on them. A scoreboard can tell us how much was produced. It can't, by itself, tell us whether the work is sustainable, whether the gains reached the people who made them possible, or what other uses of the money were left behind. Follow the value from one hand to the next, and growth starts to look less like an answer than a question.
TL;DR
Value creation and value capture are different: customers can benefit from lower prices while manufacturers lose money, and useful inventions can create gains their makers never collect. Evaluate growth through profit, return on capital, opportunity cost, and who bears the risk—not production volume alone. Episode 59 applies this distinction to Chinese industrial competition, franchise economics, creator businesses, and AI labor platforms. An imagined Philippe Aghion–Keyu Jin debate weighs public coordination against overbuilding, while Ken Thompson's Unix shows how tools gain value when others can build on them.
Key Takeaways
- Separate revenue, profit, return on capital, and social benefit: consumers can gain while producers or investors lose.
- Assess an industry boom by asking who pays the full cost, whether producers can survive without special support, who captures lower prices, and what alternatives were forgone.
- Public coordination can help assemble suppliers, infrastructure, and demand, but it also needs ways to recognize failure and redirect resources.
- A healthy national brand does not guarantee a franchisee can cover local leases, labor, royalties, and renovation costs.
- AI platforms may increase demand for expertise while fragmenting work and limiting workers' control over rates and future assignments.
- Dollar-cost averaging can support consistent investing, but it does not improve a bad asset, eliminate market risk, or guarantee a profit.
- Unix illustrates how clear interfaces, small composable tools, and collaboration can create value far beyond the original project.
Welcome to Wealth and Means — advice dressed up like hard work.
Episode fifty-nine. Let's get into it.
An unexpected star found in old telescope data, a consumer product racing through social commerce, a franchisee's financial trouble, and a tariff refund with surprising beneficiaries. What You Didn't See in the News follows stories from lab benches and shop floors to African sports venues and global trade routes. The sources are as varied as the consequences. Wake Up Ready takes us through a holiday-shortened week of consumer data, company earnings, inflation, and the Federal Reserve's local business survey. Then the Knowledge Bomb looks at how fierce competition can leave customers with lower prices while manufacturers struggle to earn a return. Judith applies dollar-cost averaging to a very perishable asset: avocados. The Greater Debate stages an imagined argument between Philippe Aghion and Keyu Jin over industrial scale, public coordination, and who bears the cost of overbuilding. Let's Invent Again traces how Ken Thompson's Unix became more valuable as other people could build on it. The question running through the hour is simple: how do we tell activity from durable value, and who gets to keep the value that gets created?
A busy shop, a full factory, and a useful invention can all create value—but they don't send it to the same people.
Let's go.
Each week we explore ideas that help you pause, reflect, and think more deeply about the opportunities all around you.
It's the perfect mix — a little information, a few stats, some real-world insights, and just enough deep talk to make you feel smarter before your second cup of coffee. A quick thank-you to our sponsor — AgentWeekly.ai - chronicling the absurd, the ambitious, and the algorithmically-challenged corners of the AI agent economy.
WHAT YOU DIDN'T SEE IN THE NEWS
Let's start with an extraordinary result from something scientists already owned: fifteen years of telescope data. Researchers working with NASA's Fermi mission found a pulsar — the rapidly rotating core of a dead star — spinning two hundred sixty-one times each second. Their new search method, adapted from an idea in cryptography, was reportedly around a million times faster for the problem it addressed, turning a daunting computing task into one accomplished in about ten processor-hours. The discovery was announced this week. Why should anyone outside astronomy care? Because the method makes old data newly valuable. Companies spend fortunes collecting and storing information they cannot fully interpret. A better search algorithm can produce discoveries without buying a better telescope, drilling another hole, or collecting another decade of observations. Sometimes the overlooked resource is the dataset already on the shelf. A very different signal is coming from a KFC franchisee. Tasty Chick’n, which operates restaurants across ten states, filed for Chapter Eleven protection and asked to reject leases on seventy-two locations. Some are already closed. This is one operator’s balance sheet, not KFC’s entire U.S. business. The operator’s sales reportedly fell in each of the last three years, while food, labor, rent, and required remodel costs rose. The franchisee has also pointed to costs it says it couldn’t keep up with under the brand’s renovation requirements. That tension is familiar in franchising: the national brand sets standards to keep the customer experience consistent, while the local operator has to finance the building, payroll, ingredients, and updates. The question is not simply whether the chicken sells. It’s whether a specific location can produce enough sales to cover the lease, labor, royalties, and capital spending the franchise agreement requires. A familiar sign can bring customers through the door. It can’t make every store’s economics work.
A national brand can be healthy while a local operator is not. The logo on the building doesn’t tell you who is carrying the debt.
Now to a product whose storefront is a feed. TokConnect estimates that Shark’s FacialPro Glow home facial system generated about two hundred fifty-four thousand dollars in TikTok Shop sales during the seven days ending November seventeenth — roughly seven hundred fifty-four percent more than the previous week. That’s an estimate from a third-party shop analytics service, not a figure Shark or TikTok has reported as audited sales. Even with that caveat, the acceleration is interesting. The device is a premium beauty product, and the sales jump suggests that product demonstrations and creator distribution can move an expensive item quickly. The lesson isn’t that every viral product is a lasting hit. It’s that social video can shorten the distance between seeing a product work and buying it — while making it harder for outsiders to distinguish a sustained trend from a brief surge. The creator economy has another kind of signal this week: an independent podcast sold for up to nine million dollars. John Lee Dumas will keep hosting Entrepreneurs on Fire for at least three years under its new owner, public-relations firm Pathos Communications. The show reported about one-point-one million dollars of revenue and eight hundred thousand dollars of net profit in twenty twenty-five. If the full headline price were paid, that would be about eight times annual revenue and a little over eleven times net income. But the phrase “up to” matters. The final price may depend on conditions we haven’t seen. And this isn’t simply a microphone changing hands. A daily show can be a durable audience, a catalog of evergreen episodes, advertising inventory, and a lead-generation channel. The unusual part is how clearly the transaction turns a creator’s media business into a priced asset. A big download count alone doesn’t make a show valuable. An audience that returns, a host who will stay, and revenue that survives a change of ownership are what a buyer is paying to preserve.
The transaction has a good underwriting question: If the host eventually stops recording, what exactly did the buyer purchase?
Here's a different shortage: credit for otherwise productive small businesses. In a report released this week, the World Bank estimated a four-trillion-dollar financing gap for micro, small, and medium-sized enterprises in developing economies. They account for an enormous share of employment, but many lack the collateral, credit histories, or banking relationships that make a loan straightforward. Imagine an equipment repair shop with customers waiting and a mechanic ready to hire. If the owner cannot finance one more machine, the business stays smaller than its demand warrants. The report matters because the cost of missing credit isn't limited to the owner. It shows up as fewer jobs, slower productivity growth, and less competition. Digital payment records and better collateral rules could help, though a clever lending app is no substitute for sound underwriting. Now a technical milestone with a commercial question attached. Quantum-Si says its Proteus development program has demonstrated detection of all twenty naturally occurring amino acids, reaching a year-end goal ahead of schedule. The company is working on a protein-sequencing platform, with the broader ambition of making protein analysis more useful in research and eventually drug development. This is a company announcement, not an independent validation of a finished commercial instrument. Detecting all twenty building blocks is a milestone; it doesn’t tell us yet how fast, reproducibly, or cheaply the system will analyze real samples. Those are the numbers customers will care about. A lab tool can be scientifically impressive and still struggle if it doesn’t fit existing workflows or beat the cost and performance of methods labs already use. The reason to watch the story is the gap between a research result and a product. The company has crossed a technical checkpoint. Now the market has to decide whether that checkpoint removes a real bottleneck for biologists, clinicians, or pharmaceutical researchers.
A machine can identify all twenty letters. The next question is whether it can read the book quickly enough to matter.
A small defense-technology company gives us a more immediate commercial test. Safe Pro Group reported third-quarter revenue of about two-point-seven million dollars, up roughly twenty-five times from a year earlier. Its AI segment posted a seventy-nine-percent gross margin. The company also reported a one-point-five-million-dollar net loss, and said that revenue included U.S. government work involving drone imagery and threat detection. The business has moved from small contracts toward a two-point-five-million-dollar Army prime contract for its software running on an Anduril drone. That’s a meaningful step: selling directly as a prime contractor can give a company more control over the customer relationship, but it also brings more responsibility for delivery. The revenue increase is striking, yet the comparison starts from a very small base. One quarter does not establish a repeatable growth rate, and a high segment margin does not mean the whole company is profitable. This is the kind of small-company story the largest market headlines can miss. The test is not whether the technology sounds advanced. It’s whether contracts turn into delivered systems, cash collections, and repeat orders. There’s also a household question hidden inside a corporate filing. TJX, the parent of TJ Maxx and Marshalls, says it has received hundreds of millions of dollars in tariff refunds. Its latest disclosure says the company accrued one hundred sixty-seven million dollars in year-end incentive compensation and discretionary bonuses tied to those refunds during the first nine months of its fiscal year. That does not mean every refund became a bonus. The company also reported refunds reducing costs and inventory, and it said the refund-related bonuses went to eligible employees globally. But customers who paid higher prices while tariffs were in force don’t automatically receive a refund when an importer gets money back. A tariff can appear on a government ledger, move through a company’s costs, and end up in a household’s checkout total. Reversing the tariff doesn’t rewind that chain in reverse. The story is a practical question about who captures a windfall. The company says it shared part of the benefit with employees. Consumers may have helped finance the original cost, but the accounting doesn’t make a direct repayment to them. Those are two different distribution choices. Across the Atlantic, British food packagers are watching a different tariff question. A proposed U.K. safeguard could apply a sixty-five-percent duty to certain imported PET resin volumes above country-specific quotas. PET is used in bottles and trays; the rule would cover both virgin and recycled resin. It has not been imposed. Businesses could comment on the proposal through Sunday, November twenty-second. The counterintuitive detail is that a tariff intended to protect domestic resin production could raise costs for manufacturers that turn resin into food packaging. It would not raise every pack’s cost by sixty-five percent: the effect depends on quota availability, origin, contracts, and the resin’s share of total cost. But if a packaging supplier cannot secure enough resin at the right price, the pressure can travel downstream to food and beverage companies, then to stores and shoppers. That is why the input deserves attention before the checkout price changes. Trade measures often sound abstract until you name the material, the factory, and the product that depends on it. And trade policy is moving through specific products in another part of the world. Canada and China agreed to tariff relief covering about four hundred million Canadian dollars of exports, including frozen pork and seafood. Most frozen pork cuts are expected to regain access without the extra Chinese tariff from January twenty twenty-seven through the end of twenty twenty-eight. The Canadian Meat Council says Canadian pork exports to China fell from about four hundred sixty-nine million dollars in twenty twenty-four to three hundred fifty-four million in twenty twenty-five, and were on track to fall below two hundred fifty million this year. Those figures come from the industry group, so treat the forecast accordingly. The reopening does not guarantee that exporters win back their old customers; competitors may have taken shelf space, and producers still have to meet buyers’ prices and specifications. But this is what a diplomatic reset looks like on a farm ledger: a tariff changes, a market reopens, and pork processors and seafood exporters decide whether the route is worth rebuilding. The headline is a meeting between leaders. The commercial consequence is a buyer deciding what goes into a container.
The export market isn’t restored when the cameras leave. It’s restored when somebody places the next order.
Now consider a public-health story that can affect ordinary lives more urgently than anything in an earnings report. The World Health Organization said Friday that an Ebola outbreak caused by the Bundibugyo virus remains an international public-health emergency. In the Democratic Republic of the Congo, more than eleven thousand confirmed cases have been recorded cumulatively, with many deaths and continuing transmission across multiple provinces. Kenya and Nigeria reported no new cases in the latest update and are counting toward the end of their outbreaks. The announcement this week is not the beginning of the crisis; it is a warning that containing it remains difficult. The second-order issue is how insecurity and weak health infrastructure make contact tracing, early treatment, and community trust harder to sustain. Vaccines and medical equipment matter. So do roads, clinics, trained personnel, and reliable communication. A global health threat often persists at the point where medicine meets logistics.
The breakthrough is not merely developing a medical response. It's making sure that response reaches the people who need it in time.
On a more celebratory note, the International Olympic Committee's update this week looked back at the Dakar Youth Olympic Games in Senegal. More than twenty-seven hundred young athletes participated, in what the IOC describes as the first Olympic event on African soil. The ceremonies and medals are the visible story. The longer investment question is what happens after the event: whether coaching programs, facilities, local suppliers, and sporting networks last beyond the closing ceremony. That is not a claim that every sporting event pays for itself. Plenty don't. It is a reminder that the right measure of a sporting investment may be the athletes and institutions it helps develop years later. Hosting has a visible finish line; a sports-development system doesn't. And in Mexico, UNESCO reported the launch of a textile-art atlas developed with artisans. At first glance, that sounds like a cultural catalog. But recording regional techniques, patterns, and their makers can help differentiate authentic craft from anonymous mass-produced imitations. A buyer may come to understand not just what a fabric looks like, but who made it and what knowledge it carries. The caution is that visibility is not the same as earning power. Documentation can preserve a tradition and potentially improve provenance. Whether artisans capture higher prices depends on distribution, bargaining power, and the legal treatment of designs. It's an elegant example of culture and commerce interacting without being interchangeable. Finally, AI’s labor story is not only about a job disappearing. A recent report on global data platforms describes companies recruiting doctors, lawyers, and PhDs to write examples and evaluate model answers. The advertised rates vary sharply: up to one hundred eighty dollars an hour for U.S. cardiologists, compared with twelve dollars and fifty cents for Swahili-language work in Tanzania. The article also reports workers facing rate cuts and sudden loss of platform access. Those are reported platform rates and worker accounts, not a universal pay scale. The commercial pattern is still worth seeing. AI firms need human expertise to make systems better, but the platform can divide that expertise into isolated tasks, set the price, and retain control of access to future work. A professional may be paid far above a local alternative for one assignment, while still lacking the predictability and bargaining power of a job. The counterintuitive point is that AI can increase demand for certain kinds of human work while making that work more fragmented. The model is not learning by magic. Someone is labeling, checking, and correcting its output — often through a marketplace most users never see.
We keep asking whether the machine will replace the expert. The quieter question is who owns the market where the expert now has to sell an hour at a time.
This week’s stories move from a star discovered in archival data to a global labor platform, but they share a practical question: Where does the value land? An audience has to produce durable cash flow. A laboratory milestone has to become a tool customers will pay for. A tariff refund has to be allocated somewhere. A major sporting event or a preserved craft tradition has consequences that outlast its headline. That’s what we went looking for: not just announcements, but the companies, workers, customers, and products underneath them. Now, the calendar.
WAKE UP READY
Next week is shorter than your average grocery list for Thanksgiving, but Wednesday alone could give traders indigestion. Here is how to wake up ready for Monday, November twenty-third, through Friday, November twenty-seventh. Monday starts with the preliminary November purchasing-manager surveys for American manufacturing and services. These give us an early read on orders, employment, and prices before the month is over. The question is whether demand is still expanding without another leg up in costs. Strong activity coupled with sticky prices could make bond investors less comfortable with a quick easing in monetary policy. After the market closes, Hewlett Packard Enterprise, Agilent, and Zoom are among the companies reporting. Listen for the difference between actual orders and optimistic language about next year's pipeline. Tuesday is consumer and technology day. The Conference Board releases its November consumer-confidence reading. But the real test is in the earnings calls. Best Buy and Dollar Tree offer different windows into discretionary spending and value shopping. If the discount retailer sees solid traffic while electronics buyers delay replacements, that suggests household budgets are becoming more selective. It doesn't necessarily mean all consumers are cutting spending. After the close, Dell and HP report. Dell's artificial-intelligence server business and the ordinary personal-computer market run on very different economics. Is growing demand for expensive computing infrastructure converting into dependable margins, or merely larger invoices and working-capital needs? HP gives us the other end of the computing stack. Analog Devices also reports Tuesday, offering a read on industrial and analog semiconductor demand. Several company calls occur at different times; the day matters more than memorizing every conference-call slot.
We might learn whether America is buying more computing power or merely moving its entire holiday budget into one data center.
Wednesday is the main event. At eight-thirty Eastern, several government releases collide: the second estimate of third-quarter economic growth, October personal income and spending, the Federal Reserve's preferred personal-consumption-expenditures inflation measures, durable-goods orders, and weekly jobless claims — moved forward because of the holiday. Don't treat those as five unrelated headlines. Put them together. Is output strong because households are spending more in real terms? Are equipment orders broadening beyond a handful of large sectors? And is the labor market still sturdy enough to support that demand? A hotter inflation reading alongside solid growth could lift yields. A weak spending number with stubborn inflation would be the more awkward combination. At ten, new-home sales tell us how much of that spending climate translates into a major financed purchase. Deere also discusses earnings. Agriculture equipment is a useful signal because farmers can have an excellent harvest and still delay a tractor if financing, commodity prices, or confidence are wrong. At two, the Fed's Beige Book adds local color from businesses around the country. It may be anecdotal, but that's its value: it can expose regional cracks long before the national averages do. Thursday, Thanksgiving, U.S. stock markets are closed. It's a good day to avoid pretending a thin market is sending a profound message. Friday is a shortened trading session, ending at one Eastern. Advance goods-trade and inventory figures arrive that morning. Watch whether imports are building ahead of demand or whether inventory is tightening. And, outside the United States, Tokyo inflation data provides a useful read on Japan's price pressures and central-bank choices. One more distinction for the week: a report being scheduled does not mean its result is predictable. We know when the questions arrive. We don't know the answers. That's how we wake up ready.
KNOWLEDGE BOMB
I heard somebody say recently that when a Chinese company walks into a market, the profits disappear. It's a clever line. Also, a dangerous one if you hear it as a claim about every Chinese business. Some Chinese firms are very profitable. Some aren't. The interesting question is why certain industries — solar panels, electric vehicles, batteries, and increasingly many consumer products — have experienced extraordinary improvements in price and capability while the companies doing the manufacturing struggle to make a return. Let me offer a distinction every young investor and entrepreneur should learn: low prices are not the same as low costs, and high sales are not the same as economic profits. Imagine a company can make a product for eighty dollars and sell it for one hundred. On each sale it has twenty dollars before other costs. Then a rival enters, builds a huge plant, and sells a comparable product for eighty-two. Everybody scrambles. The first company must improve its process or cut its price. The newcomer may genuinely be able to produce for sixty-five. Or its apparent bargain may depend on cheap financing, subsidized land, or a willingness to lose money for years. Those are very different stories, though they look similar to the customer at checkout. China's manufacturing system combines several forces. One is intense domestic competition, sometimes called neijuan, or involution: firms constantly lower prices, add features, and accelerate product cycles merely to remain in place. Another is industrial clustering: specialized suppliers, skilled workers, tooling, logistics, and customers are close together. That can produce real efficiency. And another is policy support: loans, tax treatment, land, and direct incentives that change how long a factory can operate without earning a normal return. The point is not that China has a simple command economy. It doesn't. The state sets priorities, while private companies often compete ferociously within them. The combination can generate world-class products and an uncomfortable amount of duplicate capacity. Here is the evidence I find especially useful. The International Energy Agency reported that Chinese solar-panel prices fell more than sixty percent from twenty twenty-three, amid excess supply and competition, with leading manufacturers suffering negative margins and billions of dollars in cumulative losses. A separate IEA analysis found that global solar manufacturing capacity in twenty twenty-four was more than double annual demand. Battery-cell manufacturing capacity was more than triple annual demand. That's an awful lot of factories chasing the next order. And an International Monetary Fund study estimated that the fiscal-equivalent cost of China's industrial-policy tools was around four percent of gross domestic product per year, with resulting misallocation reducing aggregate productivity in its model. Those are estimates, not a complete accounting of benefits and costs. But they warn us that subsidizing one sector can divert capital from another that might create greater value. Now the part that makes this more than a China story. Cheap solar panels are good news to an installer, a utility, and a household buying electricity. Cheap batteries can make other products possible. Consumers capture value even when manufacturers don't. We should resist the lazy impulse to call every unprofitable factory a failure for civilization. There's a real difference between a poor investment and a poor invention. But James C. Scott's Seeing Like a State gives us a warning. In the old European scientific-forestry example, administrators simplified complicated forests into something easy to count and manage: uniform rows, predictable timber yields. The initial harvest could look terrific. The hidden ecological costs emerged later. In industrial policy, the easy numbers may be gigawatts of solar capacity, millions of cars, or share of worldwide battery production. The harder numbers are opportunity cost, the durability of profits, household purchasing power, and what investment never happened because the money went somewhere else. Here's the skill I want you to take away. Whenever somebody tells you an industry is booming, ask four questions. One: who is paying the full cost of production? Two: can the producer survive without special financing or protection? Three: who captures the benefit of lower prices? Four: what valuable alternative was forgone to build this capacity? You can run the same test on a startup, a hometown sports arena, a government program, or your own side hustle. Revenue measures how much business passed through your hands. Profit measures what remained after costs. Return on capital asks whether what remained justified the money tied up. And opportunity cost asks what you could have done instead. The final irony is that a factory can win market share, the customer can win a bargain, and the investor can lose money — all at the same time. That's not a contradiction. It's the reason we keep separate scoreboards. That's your Knowledge Bomb.
HUMOR ME
I have found the perfect investment strategy for people who enjoy not knowing what they're doing but want to sound extremely mature about it. It's called dollar-cost averaging. To be clear, it's a real strategy with real advantages. You invest the same amount at regular intervals, rather than waiting to guess the perfect moment. It's useful for people who receive paychecks and would otherwise spend the whole year explaining why next Tuesday is definitely the buying opportunity. But let's acknowledge the marketing achievement. We have taken the sentence, 'I have no idea whether the price is good,' and transformed it into, 'I am implementing a disciplined capital-deployment framework.' That is an impressive return on vocabulary. I'm going to bring this wisdom to the supermarket. Every Friday, I will spend exactly fifty dollars on avocados. If they're one dollar each, I buy fifty avocados. If they're twenty dollars apiece, I buy two and a half. Do not ask what I do with half an avocado. That's a settlement issue and my broker is handling it. My friends say, 'Judith, you can't use that many avocados.' I say, 'I'm not attempting to time the guacamole market.' They point out that the fruit is perishable. I say, 'You clearly don't understand my long-term horizon.' The manager announces an avocado shortage. The price triples. I buy. A ship arrives carrying ten million avocados. The price collapses. I buy. Someone develops an avocado that never turns brown. The entire market enters a speculative frenzy. I buy. A nutrition influencer declares avocados morally questionable because they were seen at a luxury resort. Panic selling breaks out. I buy. Whether it's mania or hysteria, I'm buying. My portfolio is not particularly diversified, but my kitchen smells like a produce terminal. Then comes my annual performance review. A neighbor says, 'Wouldn't you have made more money buying when avocados were cheap?' Well, yes, if I had known in advance when they were cheap and when they were going to be expensive. That is also how I would have won every football game I ever watched. Now, a serious distinction before the Securities and Exchange Commission investigates my refrigerator. Dollar-cost averaging works differently with perishable groceries than with a diversified long-term investment. It does not turn a bad asset into a good one, guarantee a profit, or erase market risk. And if you already have a lump sum invested sensibly for a long horizon, spreading it out can have a cost when markets rise. The point is behavioral: a recurring plan can beat your own endless hesitation. My avocado strategy, on the other hand, has created a liquidity crisis. Literally. Everything in the bottom drawer is liquid. I've launched a new fund. Ticker symbol GUAC. Expense ratio: one tortilla chip annually.
GREATER DEBATE
Two economists. One crowded factory floor. And a question that refuses to fit into a campaign slogan. On one side, Philippe Aghion, whose work on creative destruction asks how economies keep innovating as new companies displace older ones. On the other, Keyu Jin, who has described the distinctive relationship between Chinese local governments, entrepreneurs, and the state. This is an imagined debate. We are testing their intellectual frameworks, not quoting them. Above the factory door is an electronic counter. It measures units produced. On the manager's desk is another figure: return on invested capital. The first keeps rising. The second has turned red. The question is whether that factory is an engine of progress or a warning light. Aghion begins. 'Competition matters because it doesn't merely reward better firms. It frees resources from worse ones. A company makes a breakthrough, earns profits, attracts rivals, and eventually must innovate again. If it cannot create value, it should release its workers, capital, and equipment to firms with more promising ideas. That is painful. It is also how progress keeps moving. When governments protect capacity regardless of its returns, they interrupt the process. We get the appearance of dynamism without the discipline that makes dynamism sustainable.' Jin responds. 'You describe the end of an industry as if the beginning were costless. But some technologies cannot be developed economically at small scale. You need suppliers, engineers, roads, electricity, and buyers — and you need them together. Each investor waits for the others. In that setting, public coordination can solve a genuine problem. China's system has often joined ambitious government targets with intense rivalry among firms. The competition is not absent. It is sometimes more ruthless than competition in comfortable mature markets.' 'Then let us look at the result,' Aghion says. 'Solar panels have become dramatically cheaper. Excellent. Yet enormous capacity sits underused, and producers endure losses. At some point we must ask whether the price tells us the full cost. A firm paying an ordinary cost of capital might not build that factory. If the state makes credit cheap enough, dozens build it. They all report production. None can recover the investment. The consumers of other countries receive a subsidy, while the people financing the factories bear the hidden cost.' Jin pauses. 'You are counting the factory owner's return while discounting the learning that cheap production creates. When solar panels become affordable, demand rises, installers learn, suppliers improve, and electricity users benefit. Some of these gains do not appear on the factory's income statement. We can acknowledge misallocation and still ask whether the private capital market would have delivered the transition quickly enough. Are we judging a system by quarterly earnings or by the capabilities accumulated over a generation?' 'Neither,' Aghion replies. 'Judge it by productivity and the opportunity cost of resources. You cannot spend the same savings twice. If the tenth battery plant is unnecessary, its engineers might be developing a better chemistry, its land might house something more useful, and its credit could be supporting small firms the government did not designate strategic. Innovation needs experiments. It also needs permission to conclude that an experiment failed.' Jin raises an eyebrow. 'Western capitalism is not an untouched control group. Governments fund basic research, roads, universities, defense procurement, semiconductor facilities, and energy incentives. Private firms can be very patient with an unprofitable platform when they expect network effects. They can also be spectacularly impatient with an investment whose benefits spill over to society. Your model has to tell me not only how to shut a factory, but who would have built the first one.' 'And yours,' says Aghion, 'must tell me who has the authority to stop building the fiftieth.' That is the room's first real silence. Jin answers it directly. 'That is the strongest criticism. Local officials have incentives to meet visible targets. Multiple regions can back the same industry. Firms can compete on price in ways that undermine profitability without necessarily encouraging better research. The system needs exit, harder budget constraints, and a way to redirect support when capacity exceeds demand. But the existence of overbuilding does not prove that the original coordinated push was misguided.' Aghion nods. 'Agreed. Public investment can correct a market failure, especially in research and early infrastructure. My objection is to replacing discovery with a production target. The state can help establish the conditions for competition. It should be careful about specifying which companies will win and continuing to finance them after the evidence changes.' 'And I object,' Jin says, 'to treating any period of losses as evidence of failure. A global supply chain does not emerge from a spreadsheet showing satisfactory returns in year one. The discipline should be adaptive: invest where coordination matters, expose the firms to competition, and withdraw support when the original bottleneck has been solved.' The discussion shifts to a book neither economist wrote: James C. Scott's Seeing Like a State. German scientific foresters once simplified living forests into neat, countable timber. The trees grew. The ledger looked wonderful. Then ecological costs appeared that the ledger had never been designed to record. Aghion sees the parallel immediately. 'Exactly. Capacity is legible. Resilience and displaced innovation are harder to measure. If an official's career depends on the number of gigawatts constructed, that is the number they will optimize. The market's scattered signals about actual demand and alternative investments become background noise.' Jin is not prepared to surrender the metaphor. 'But a modern Chinese industrial cluster is not simply a monoculture dictated from a capital city. It contains competing firms, specialists, suppliers, local experiments, and entrepreneurs who may know more about the production process than officials do. The analogy should warn us about top-down measurement, not erase the bottom-up complexity that made the industry capable.' 'Then,' Aghion says, 'we may have found the real disagreement. Not whether to industrialize. Not whether firms should compete. Whether public coordination can resist its own temptation to keep score with the easiest numbers.' Jin smiles. 'And whether private markets can learn to value benefits that no single firm can charge for.' Let's stop the debate there. The strongest case for creative destruction is that losses carry information. If we suppress that information indefinitely, capital becomes trapped and the next idea goes unfunded. The strongest case for state-directed scale is that some industries need a coordinated leap before profits are visible to anyone, and the social gains from affordable technology can extend far beyond the first investors. So here's the question for the listener: suppose cheap Chinese solar panels save your community money, help cut emissions, and bankrupt a local producer. Was the result a success? An economic answer has to name whose welfare we're measuring, who paid the subsidy, what resources were displaced, and whether the system can change course when the evidence turns. That, to me, is a greater debate. Not capitalism good or state bad. The harder question: who gets to declare victory, and who gets the bill?
LET'S INVENT AGAIN
Imagine working at one of the most sophisticated research organizations in America and being told, more or less, that the grand computing project you were helping build was over. For most people, that would be a cue to clean out a desk. For Ken Thompson, it became the beginning of something smaller, stranger, and ultimately much bigger. Thompson was a young engineer at Bell Labs in the late nineteen-sixties, a time when computing power was precious, machines were expensive, and operating them felt rather like negotiating with a temperamental bureaucracy. You prepared your work, submitted it to the computer, and waited. The machine belonged to the institution; the user was a visitor with a numbered ticket. Bell Labs had participated in an ambitious operating-system effort called Multics. The goal was to create a powerful, shared computing environment. The project was technically formidable and organizationally difficult. Bell Labs eventually withdrew. Thompson, however, had become attached to an idea that need not be buried with the program: a computer could be something a person used interactively, and the software could be built around small, clear tools. He and collaborators began developing a leaner operating system on comparatively modest hardware. Its name became Unix, a wry echo of Multics. It offered several users access to a system, let multiple tasks run, and organized files in a coherent hierarchy. But Unix's deeper invention was philosophical. Rather than building one enormous program to do everything, build small programs that do specific things well. Then make it easy to connect them. Picture a workshop where every tool has a different plug and a different power source. Now picture a workshop where tools share a standard connector. One tool sorts. Another filters. Another counts. You can chain them together to solve a problem nobody anticipated when those tools were made. In Unix, pipes made that kind of composition natural. Programs could feed their output into other programs. Complexity became something you could assemble from manageable parts. It seems obvious now because so much of modern computing inherits the pattern. But the engineering tradeoff was not obvious when computers were rare, memory was tight, and vendors often built operating systems around their own machines. A system designed for one particular computer could become obsolete when the hardware changed. Unix's team did something consequential: they moved much of it into the C programming language developed by Dennis Ritchie. Thompson rewrote the Unix kernel in C, making the system far easier to carry across different kinds of computers. This was a crucial economic move disguised as a technical one. Portability lowers switching costs. It allows skills, applications, and institutions to survive a change in hardware. Once software can travel, the value of the tools and the people who understand them compounds. Researchers and universities could share ideas more readily. New computer makers did not have to start from zero. A system born in a laboratory began influencing an industry. Thompson had help and predecessors, of course. Unix was not the work of one isolated genius. Dennis Ritchie was indispensable, and the broader Bell Labs culture mattered. So did the users who tested, modified, taught, and extended the system. That's part of the story's appeal: its impact came through a set of decisions that made the work reusable, not through one dazzling demonstration. And Thompson did not stop with Unix. He collaborated on a chess-playing computer called Belle. Later, with Rob Pike, he developed UTF-eight, a way to represent the world's written characters in digital systems while remaining compatible with older text conventions. Think about that for a moment. A person associated with the foundations of operating systems also helped make it practical for a web page to carry different writing systems. A seemingly humble encoding decision became an enormous contribution to global communication. One of the persistent temptations in technology is to confuse sophistication with value. We admire the largest model, the most elaborate platform, the dashboard with seventeen tabs. Thompson's contribution points in the opposite direction. A well-chosen constraint can make the system more powerful. A small tool that works with other small tools can outlast a magnificent system that insists on doing everything itself. There's also a business lesson. Companies often create value by owning an integrated stack. Sometimes, however, the most durable value emerges from an interface, language, or standard that lets many other people build on top of it. The original creator may never capture all of that value. Unix and Unix-like systems went on to underpin computing environments far beyond the laboratory where the work began, from servers to devices in ordinary people's pockets. That brings us back to today's debate about scale and returns. The most influential invention is not necessarily the product with the highest margin in its first year. It may be the architecture that makes a thousand later products possible. But that does not excuse every loss or validate every subsidy. It reminds us to distinguish the economic value of a particular company from the social value of an idea. Ken Thompson's lesson is bracingly simple. If a grand plan becomes unwieldy, look for the small part that matters. Make it clear. Make it useful. Make it possible for somebody else to connect their work to yours. Sometimes that is how a modest experiment becomes infrastructure.
Closing
That's it for another episode of Wealth and Means — advice dressed up like hard work. We hope you enjoyed the arc. From a distant star hidden in old data to artisans making their work visible… from factory losses to software tools that outlast their makers… the question keeps widening: who gets the value, and who carries the cost? The pattern was simple: trace who benefits. Because first principles are rarely flashy… Subscribe, rate, and share this episode. It helps more than you think. We do this for you — to help you have a better week, a sharper mind, and maybe even a longer attention span. And a special thanks to our sponsor AgentWeekly.ai Until next time -- stay curious. Stay kind. And keep compounding.
Links and Sponsors
Produced by
https://wealthandmeans.com https://wealthandmeans.substack.com/ https://x.com/WealthandMeans https://www.youtube.com/@WealthandMeans
Disclaimer
When we mention or portray celebrities or public figures in fictional debates or scenarios, it's exactly that — fiction. They didn't approve it, they didn't review it, and they're not endorsing anything here.
Sponsors
https://agentweekly.ai — chronicling the absurd, the ambitious, and the algorithmically-challenged corners of the AI agent economy. https://nostmoments.io — share the memories, share the nost moments. https://syrepu.com — syrepu (si re poo) - the synonym reverse puzzle. https://teamloadout.com — a single command post for your AI agents.
Chapters
- 00:00:00 — Introduction to Wealth and Means
- 00:01:48 — What You Didn't See in the News
- 00:16:38 — Wake Up Ready
- 00:20:21 — Knowledge Bomb
- 00:25:23 — Humor Me
- 00:28:05 — The Greater Debate
- 00:35:36 — Let's Invent Again
- 00:41:01 — Closing Thoughts and Reflections