--- title: "Price Without Discovery: From the Signal to the Instruction" kind: "Research" topic: "McLuhan & media theory" date: "2026-09-01" url: "https://canopticon.com/p/price-without-discovery-from-the-signal-to-the-instruction" author: "Chris Telford" license: "CC BY 4.0" --- # Price Without Discovery: From the Signal to the Instruction ### A CANOPTICON thesis brief — lead essay plus two groundwork strands ## TL;DR - **The lead thesis is sound and publishable:** for two centuries the price was understood as an *emergent signal no one sets* (Hayek, 1945); algorithmic, dynamic, and surveillance pricing increasingly make the price *computed and pushed*. The sharpest framing — McLuhan's tetrad — holds up: surveillance pricing **retrieves the haggle**, but with the information asymmetry inverted (the seller knows everything about the buyer, the buyer knows nothing). The fixed price tag was a genuine 19th-century invention (Quaker merchants, A.T. Stewart, John Wanamaker) that abolished individualized bargaining; personalized algorithmic pricing brings the bazaar back through the back door. - **Strand One ("RealPage North") is already partly answered, and that is the story:** Canada's Competition Bureau ran a civil investigation (Jan–Nov 2025) into RealPage and Yardi, concluding Nov 10, 2025 that adoption was too low to prove harm — while a Canadian class action (Cynthia Black v. ~14 landlords, filed Dec 2024) proceeds and CBC documented YieldStar in use in Canada since 2017. Canadian usage is *confirmed but apparently shrinking under scrutiny*; the live investigative target is now the disclosure gap and which institutional landlords used it before retreating. - **Strand Two ("The Hayek Inversion") is most defensible stated narrowly:** big-firm computation (Walmart/Amazon) demonstrably solves **logistics**, which even Mises conceded planning could do; it does *not* dissolve the **capital-allocation** calculation problem, because internal transfer prices ultimately derive from real external factor markets. The genuinely new category to name is **administered pricing**: prices set by many private, opaque corporate algorithms — neither Hayekian discovery nor public planning, but an opaque third mode. ## Key Findings ### The fixed price is a recent invention, and that is the whole hinge of the argument The posted, non-negotiable price tag is neither ancient nor natural. Before the mid-19th century, most retail buying in America was done by haggling; merchants intentionally withheld a fixed price (if a merchant thought he could get $10 for an axe, he would not let you see a $7 price). The one-price system emerged from Philadelphia Quaker merchants — who, on a belief that all are equal before God, decided all should pay the same — was systematized commercially by A.T. Stewart in New York around 1846, and was popularized through the price tag by John Wanamaker, who opened his first store in 1861 and formalized fixed pricing and price tags at his Grand Depot department store (1875–1876). The department store and the fixed price were the same event: mass scale made person-by-person bargaining impossible, and the posted price was the technology that resolved it. This is precisely the figure that personalized algorithmic pricing now *obsolesces*. A useful honesty note for the essay: "administered prices" is not a neologism — economist Gardiner Means coined the term in the 1930s to describe prices set by oligopoly fiat rather than market clearing. The CANOPTICON contribution is to extend it from the boardroom to the individual consumer and to read it through McLuhan. ### Hayek's signal, stated at full strength Hayek's "The Use of Knowledge in Society" (*American Economic Review*, vol. 35, September 1945, pp. 519–530) argues that the economic data of a society "never exists in concentrated or integrated form, but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess," and that the price system is a "marvel" precisely because it aggregates this dispersed local knowledge with an economy of attention — no single mind sets or needs to know the whole. In his famous tin example, when the price of tin rises, users economize and seek substitutes "without knowing why," the information rippling through the system via price alone. The price is an *emergent* coordinating signal. The essay was written as a rebuttal to Oskar Lange's market socialism — which matters, because the thesis is in one sense Lange's revenge: computation now does reach into price-setting, just privately rather than publicly. ### The mechanism: medium-as-message — same number, different medium The thesis's strongest move is McLuhanite and does *not* depend on prices being higher. Even when an administered price equals the price a market would have discovered, its *medium* has changed. A discovered price is **feedback**: it carries dispersed information back to all participants symmetrically; everyone faces the same public number. An administered price is an **instruction**: it carries the seller's private model of *your* willingness to pay, computed from surveillance data and pushed at you alone. The content (the number) may be identical; the message (who knows what, who sets it, whether it is public) is inverted. The price stops being something the market *says* and becomes something the algorithm *tells you*. This is the medium-as-message argument in its purest economic form. ### The tetrad on the pricing algorithm - **Enhances:** the speed and granularity of price-setting; the seller's capacity to extract willingness-to-pay; margin. (In Dubé and Misra's field experiment, personalized pricing "improves the firm's expected posterior profits by an additional 19%, relative to the optimized uniform price, and by 86%, relative to the firm's unoptimized status quo price" — NBER Working Paper 23775, rev. June 2021.) - **Obsolesces:** the single posted price; the Wanamaker one-price settlement; the very idea of "the" price for a good. - **Retrieves:** the bazaar — individualized, person-by-person pricing — i.e., the haggle, but with the information asymmetry of the bazaar *reversed*. In the souk the buyer could read the seller's face and walk away; here the seller reads the buyer's data and the buyer sees nothing. - **Reverses into:** pushed to its limit, the "free-market price" flips into an administered instruction, and "the market set it" becomes a myth that conceals computation. The limit case is perfect (first-degree) price discrimination, where consumer surplus is driven to zero. ### Figure/ground reversal Classically the market is the **ground** and the price is the **figure** that emerges from it. Surveillance pricing reverses the pair: the *algorithm* becomes the new ground — the environment within which prices form — while "the market set it" recedes into background myth, a story told about an administered reality. McLuhan's instruction (study the ground, not the figure) is the method: do not ask what the number says; ask what the medium does. ### The evidence base is real, and 2025 was the inflection year - **FTC surveillance-pricing 6(b) study (Jan 17, 2025):** A staff perspective based on documents from intermediaries including Mastercard, Accenture, PROS, Bloomreach, Revionics, and McKinsey & Co. The FTC found these intermediaries "worked with at least 250 clients that sell goods or services ranging from grocery stores to apparel retailers," and that granular data — down to mouse movements on a webpage and items left in an abandoned cart — can be used to tailor prices. The Commission "voted 3-2" to release it, with Commissioners Andrew Ferguson (incoming chair) and Melissa Holyoak dissenting on timing; the accompanying request for comment was withdrawn days later. The findings are *preliminary staff observations*, not a finding of illegality. - **New York Algorithmic Pricing Disclosure Act:** Requires the consumer-facing disclosure "THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA." Passed May 2025; took effect July 8, 2025; enforcement was stayed pending a National Retail Federation First Amendment challenge and resumed November 10, 2025 after a court upheld the compelled disclosure as "plainly factual." Civil penalties up to $1,000 per violation, enforced by the NY Attorney General; first law of its kind. (The disclosure applies only where *personal* data sets the price — pricing varying solely on supply/demand or time of day is exempt. This statutory line is itself the thesis's distinction between dynamic and personalized pricing.) - **The 2025–2026 legislative wave:** California AB 325 (effective Jan 1, 2026) prohibits "common pricing algorithms" used in anticompetitive agreements; bills were introduced or proposed in Colorado (HB 1264, outright ban), Ohio (SB 136, disclosure), Vermont, Minnesota, New Mexico, and elsewhere; federal proposals include Senator Klobuchar's Preventing Algorithmic Collusion Act of 2025. A December 2025 White House executive order on AI directs DOJ to challenge some state AI laws, which could collide with these regimes. - **Adoption and scale:** The dynamic-pricing/yield-management software market was roughly $5.2B in 2024, projected toward ~$10.8B by 2034 (multiple market-research estimates). Penetration of true dynamic pricing is still early — McKinsey-cited material in the FTC record emphasizes margin upside ("two to seven percentage points of sustained margin improvement") rather than current penetration, and a commonly cited 15–20% retailer-adoption figure for 2024 lacks a clean primary source and should be presented as an industry estimate, not a hard number. On price-change velocity, the widely recycled claim that Amazon makes "2.5 million price changes a day" traces to a 2013 Profitero analysis (which contrasted Amazon's figure with Best Buy's ~53,000 and Walmart's ~55,000 changes that month) — it is real but a decade old and should be flagged as such. Walmart and Kroger are now rolling out electronic shelf labels; Walmart denies these constitute "dynamic pricing" while acknowledging an "end-to-end price markdown system" that adjusts online prices by predicted demand and price sensitivity. - **Algorithmic collusion — the deepest and most contested point:** In *simulation*, Calvano, Calzolari, Denicolò & Pastorello, "Artificial Intelligence, Algorithmic Pricing, and Collusion," *American Economic Review* 110(10) (2020): 3267–3297, found Q-learning algorithms "consistently learn to charge supracompetitive prices, without communicating with one another," sustained by reward-punishment strategies (a finite punishment phase, then gradual return to cooperation), robust to asymmetries and player count. This finding is **actively contested**: Asker, Fershtman & Pakes (*J. Econ. & Mgmt. Strategy*, 2024) show the outcome flips with the learning protocol (synchronous updating yields competitive prices, asynchronous can yield near-monopoly); Abada & Lambin (*Management Science*, 2023) argue the result "could originate in imperfect exploration rather than excessive algorithmic sophistication"; den Boer, Meylahn & Schinkel argue the collusive equilibria "can only be learned on time scales irrelevant to the firm's objective" and require synchronization so extreme it "suggests the need for an explicit cartel agreement." Calvano et al. rebutted (IJIO, 2023) that genuine collusion appears when algorithms are forward-looking with memory and explore randomly. The leading *real-market* evidence is Assad, Clark, Ershov & Xu, "Algorithmic Pricing and Competition: Empirical Evidence from the German Retail Gasoline Market," *Journal of Political Economy* 132(3) (2024): 723–771: adoption "increases margins by 9%, but only in non-monopoly markets," and margins "increase by 28% in markets where both [duopolists]" adopt, with the effect emerging only "about a year after market-wide adoption, suggesting that algorithms in this market learn tacitly-collusive strategies." The authors stress they observe outcomes, not proof of intent. **Present this as contested, not proven** — but note that the one strong real-market study points the same direction as the simulations. ### Strand One: RealPage North — confirmed, then contested - **The US template.** The DOJ Antitrust Division sued RealPage in August 2024; an amended complaint (January 2025) added six large landlords. The DOJ's framing is the thesis's spearhead: where collusion once required "a formal handshake in a clandestine meeting," now "algorithms are the new frontier." The proposed settlement (filed Nov 24, 2025) imposes no fine and no admission of wrongdoing but bars RealPage from using nonpublic, competitively sensitive data in model training and runtime, restricts it to data at least 12 months old, prohibits real-time lease data, bans sub-state geographic modeling, installs a court-appointed monitor for three years, and requires RealPage to cooperate against the remaining landlord defendants. Greystar — described by the nine-state coalition (e.g., Minnesota AG Ellison, Nov 19, 2025) as "the largest landlord in the United States," managing "nearly 950,000 rental units nationwide" — settled with the DOJ in August 2025 and agreed to a $7 million settlement in the parallel multi-state action. The case originated in ProPublica's 2022 investigation; prosecutors allege one landlord raised rents more than 25% within 11 months of adopting the software. The DOJ's own frame — "replacing competition with coordination" — is the cleanest available statement of the lead thesis in legal language. - **Canada — confirmed usage.** CBC's investigation (Angela Hennessy, Oct 31 and Dec 10, 2024) found YieldStar has been used in Canada since at least 2017. GWL Realty Advisors (a Canada Life division) confirmed it had used YieldStar and "terminated" it after an internal review; Tricon Residential said it stopped in summer 2024; Dream Unlimited and Choice Properties said a third-party property manager used it and were told to stop; CAPREIT says it has never used it and will seek dismissal. RealPage told CBC its "revenue management software has never served more than 1% of the rental market in Canada." So Canadian usage is *established by evidence*, not merely plausible by analogy — but it is also low and, since late 2024, retreating. - **Canada — the class action.** A proposed class action (lead plaintiff Cynthia Black, a GWLRA tenant in Toronto), filed December 2024, names roughly 14 landlords and managers (including RealPage, Westcorp, FirstService Residential, GWLRA, Dream Unlimited, and Associated Property Management) and seeks compensation for tenants who may have overpaid rent back to 2009. It is not yet certified. - **Canada — the regulator.** The Competition Bureau opened a civil inquiry into RealPage Canada and Yardi Canada (the latter's product is Revenue IQ, formerly RENTmaximizer; RealPage's Canadian products include YieldStar, AI Revenue Management/AIRM, and Lease Rent Options). It concluded Nov 10, 2025 that adoption was "not sufficiently pervasive" to support a finding of abuse of dominance or anticompetitive collaboration, noting that since late 2024 landlords "significantly reduced their use." The Bureau "remains concerned" and issued compliance guidance. (Threshold context the Bureau cited: a single firm with ≥50% share — or ≥65% for a group — generally prompts further investigation.) A separate Bureau investigation into Kalibrate Canada (retail-gasoline pricing data) is ongoing, signaling the Bureau treats algorithmic pricing as a cross-sector priority. - **The Canadian legal framework — the structural gap.** Section 45 (criminal conspiracy) requires proof of an agreement beyond a reasonable doubt; section 90.1 (civil) reaches anticompetitive agreements or arrangements but still demands an "agreement" — a "meeting of the minds." Neither reaches *tacit* algorithmic coordination, and conscious parallelism is not illegal in Canada. The Bureau's June 10, 2025 discussion paper notes that over 60 companies in Canada offer algorithmic-pricing services and explicitly flags tacit AI collusion as a recognized enforcement gap. This is the durable story even after the investigation closed: the law is built for human agreements, and the technology routes around it. - **The concentration that makes it matter.** REITs and other financialized landlords own roughly 25–30% of Canada's purpose-built rental stock (CMHC); the top 25 financial landlords held about 330,000 units (~20% of stock) around 2020, with continued concentration since. CAPREIT (about 45,400 Canadian suites, ~$14.5B in assets as of March 2026) and Starlight (about 66,000 units) are the largest. Vacancy decontrol — especially in Ontario — creates the unit-turnover incentive that revenue-management software exists to optimize. Algorithmic rent-setting is dangerous in proportion to concentration, and Canadian concentration is high and rising even as YieldStar specifically retreats. - **What an investigative instrument could feasibly detect from public data.** REIT annual reports and securities filings; vendor client lists and published case studies; job postings naming YieldStar / Revenue IQ / AIRM; property-management contracts surfaced in discovery; tenant-reported rent-increase patterns clustering across separately owned buildings in one submarket; FOI and class-action discovery. Honest flag: because of the post-2024 retreat, detection is now partly *forensic-historical* (who used it 2017–2024) plus a forward watch on successor products (AIRM/Revenue IQ) and on whether disclosure obligations are evaded. ### Strand Two: The Hayek Inversion — pressure-tested - **The original debate.** Mises (1920, "Economic Calculation in the Socialist Commonwealth") argued rational allocation is impossible without genuine money prices for capital goods, because absent factor markets such goods are "merely internal transfers," not "objects of exchange." Hayek (1945) reframed this as a knowledge problem — the dispersal of tacit, local information. Lange-Lerner market socialism answered with a central board setting simulated prices by "trial and error" toward equilibrium. - **The modern revival.** Phillips & Rozworski, *The People's Republic of Walmart* (Verso, 2019), argue Walmart and Amazon are de facto planned economies of country scale, allocating vast resources internally *without* an internal market, and so prove that large-scale planning is feasible — the only open question being whether it can be made democratic. - **The hard rebuttal (the defensible core).** What Walmart and Amazon do internally is **logistics optimization** — inventory, routing, scheduling, demand forecasting — exactly the class of problem Mises conceded planning could handle. What they do *not* do is dissolve **capital allocation**: they operate *inside* a market that supplies them real external prices for capital, labor, land, and acquisitions, and their internal "transfer prices" are calibrated to those external prices. Remove the surrounding market and the calculational anchor vanishes. The Mises Institute line (and adjacent Austrian critiques) is that Phillips & Rozworski conflate *economic planning* (which every firm does) with a *planned economy* (which abolishes the price system itself). This distinction is the spine of the strand and must not be blurred. - **The computational counter-counter.** Cottrell & Cockshott (*Towards a New Socialism*, 1993; "Calculation, complexity and planning: the socialist calculation debate once again," *Review of Political Economy*, 1993) argue modern computers can solve million-variable input-output systems using labor-time values as the numéraire, defeating the *feasibility/complexity* version of Mises's challenge. The honest rejoinder: this answers the **computation** problem but not the **valuation/knowledge** problem — labor-time values are not willingness-to-pay, and without factor markets you still cannot recover the opportunity cost of capital. Even a sympathetic Austrian reviewer (Brewster, *Quarterly Journal of Austrian Economics*, 2004) notes that C&C's retention of a consumer-goods market quietly reintroduces market valuation through the back door. - **The defensible synthesis — name the third category.** Prices today are increasingly **administered**: set by many private, competing, opaque corporate algorithms. This is *neither* Hayekian decentralized discovery (because the buyer is no longer a price-*taker* facing a single public number — the seller computes a private number for each buyer) *nor* socialist public planning (because it is private, profit-driven, and uncoordinated by any single transparent plan). What is genuinely new is that calculation has been *privatized and rendered opaque* — performed by thousands of firm-level models, each holding asymmetric information about individuals, with neither a public price nor a public plan. The Hayekian "marvel" of a shared signal and the socialist dream of a transparent plan are defeated *at once*: we get computation without either discovery or accountability. AI plausibly deepens this — it improves the computation and the willingness-to-pay estimation — but it does not resolve the valuation problem, so it does not vindicate central planning; it builds a third thing. ## Details ### The synthesis: one CANOPTICON worldview All three strands share a single structure — an **official story concealing an administered reality**: - "The market set this price" conceals a number computed for you and pushed at you. - "Competition keeps rents fair" conceals revenue-management software and a regulator that cannot reach tacit coordination. - "Markets versus planning" conceals a third, opaque mode that is neither. This pairs directly with CANOPTICON's wealth-measurement work: "the official statistics show this wealth" conceals an administered or invisible distribution beneath, exactly as "the market set this price" conceals an administered number. In both cases the *medium* — the official figure — is the message, and the figure functions to naturalize an engineered reality. The McLuhan method is identical across both: study the ground, not the figure; ask what the number *does*, not what it *says*. That is the throughline that makes the three strands one worldview rather than three articles. ### Publishable essay structure for the lead thesis 1. **Cold open:** a consumer sees a price; reframe immediately — it was computed *for them*, not discovered. 2. **The fixed price as invention** (Quakers → Stewart → Wanamaker): what the price tag abolished, and that it is only ~160 years old. 3. **Hayek's signal at full strength:** the price as emergent, the "marvel," the tin example; note the irony that the essay was an anti-Lange polemic. 4. **Medium-as-message:** same number, inverted medium — feedback versus instruction. 5. **The tetrad** (enhance / obsolesce / retrieve / reverse) **and the figure/ground reversal.** 6. **The evidence:** FTC 6(b), the NY disclosure law and the statutory dynamic-vs-personalized line, the 2025–26 legislative wave, adoption scale, the collusion literature (presented as contested). 7. **Counterarguments and honest responses** (below). 8. **Close:** the allocation economy — the same algorithmic logic that priced your *attention* now reaches into what you *pay*. ### Counterarguments and honest responses - **"It's just faster discovery, not its negation."** Partly true for *dynamic* pricing keyed to public supply and demand — a single public number moving fast (airlines, surge, ride-share). But *personalized/surveillance* pricing is categorically different: there is no longer a single public price to discover; the seller computes a private number from your data. New York's law draws exactly this line, exempting prices that vary solely on non-personal data. **Concede dynamic pricing is faster discovery; insist personalized pricing is something else.** - **"Markets always had price discrimination."** True — first-, second-, and third-degree discrimination (coupons, student discounts, airline fences) are old. What is new is *scale, granularity, and asymmetry*: machine learning approximates first-degree (perfect) discrimination at population scale, using surveillance data the buyer cannot see or contest. A difference in degree becomes a difference in kind. - **"It's efficiency-enhancing / welfare-improving."** Honestly contested, and the thesis should not deny it. Dubé & Misra found that "over 60% of consumers benefit from lower prices under personalization," and other work (Mercatus; airline studies) finds redistributive and access benefits for lower-willingness-to-pay buyers. But in the same Dubé–Misra experiment "total consumer surplus declines 23% under personalized pricing relative to uniform pricing" while firm profits rise, and perfect discrimination drives consumer surplus to zero. So total output/efficiency can rise while the *distribution* shifts hard toward sellers. **The honest position concedes ambiguous welfare and locates the objection in distribution, asymmetry, and autonomy — not in pure efficiency.** This is also the strategically strongest position, because it cannot be refuted by a study showing some consumers paid less. ### What each strand needs to stand alone - **Strand One** needs primary documents (the class-action statement of claim; the Bureau's now-public position statement and discussion paper; vendor case studies), a systematic scan of REIT disclosures and job postings, and tenant rent-pattern data. Frame it honestly as "confirmed but retreating": a forensic-historical investigation of 2017–2024 usage plus a forward watch on successor products and on disclosure evasion. - **Strand Two** needs a careful logistics-versus-capital-allocation section that does not overclaim, direct engagement with Cottrell–Cockshott *and* their critics, and a precise definition of "administered pricing" as the third category — including a clear statement of what AI does and does not change (it advances computation and willingness-to-pay estimation; it does not solve valuation without factor markets). ## Recommendations 1. **Lead with the tetrad and medium-as-message, not with "prices are higher."** The argument is strongest where it does not depend on prices rising. Concede that personalization sometimes lowers prices; locate the harm in inverted asymmetry and the loss of a shared public signal. This pre-empts the efficiency rebuttal and is the move a hostile economist cannot easily counter. 2. **Publish the lead essay now — it is fully sourced.** Use the eight-part structure. *Benchmark that would change the framing:* if adoption of *personalized* (as distinct from merely dynamic) pricing stalls below roughly a fifth of retail through 2027, soften "increasingly administered" to "emerging at the frontier." 3. **Run Strand One as a forensic-historical investigation, honestly framed.** The headline is *not* "secret Canadian rent cartel" — the Bureau closed that door for now. The headline is: *the software was here since 2017, it is retreating under scrutiny, the Competition Act cannot reach tacit algorithmic coordination, and the landlord concentration that makes such tools dangerous is intact and growing.* *Threshold to escalate to a harder claim:* evidence of a successor product (AIRM / Revenue IQ) deployed at scale, or class-action discovery showing live data-pooling among competitors. 4. **State Strand Two narrowly.** The defensible claims are (a) the logistics-versus-capital-allocation distinction and (b) the naming of "administered pricing" as a third mode. Do *not* claim AI "solves" the calculation debate in either direction. *Threshold that would change it:* a real economy demonstrating rational capital allocation without factor markets — none exists. 5. **Bind the three strands with the wealth-measurement pairing explicitly,** as the CANOPTICON throughline: the official figure as a medium that conceals an administered or invisible reality. This is what makes it a worldview, not a trilogy. ## Caveats - The FTC 6(b) study is a *preliminary staff perspective*, released on a 3–2 party-line vote and partly walked back (the comment request was withdrawn); it is not a finding of illegality, and the current FTC may not pursue it. Do not overstate it. - Algorithmic collusion is *genuinely contested*: robust in simulation (Calvano et al.), sharply disputed in interpretation (Asker–Fershtman–Pakes; Abada–Lambin; den Boer et al.), with one strong real-market study pointing toward harm (Assad–Clark–Ershov–Xu, German gasoline). Present it as an open debate, not a settled fact. - Canadian RealPage/Yardi usage is *confirmed but low and shrinking*; the Bureau found insufficient evidence of harm. Asserting an active Canadian "rent cartel" overstates the record. - "Administered pricing" as a named third category is an analytical move; flag it as CANOPTICON's framing — and credit Gardiner Means for the original term to avoid the appearance of inventing it from scratch. - Welfare effects of personalized pricing are ambiguous in the peer-reviewed literature; resist claiming it uniformly harms consumers. - The "2.5 million Amazon price changes a day" figure is a 2013 Profitero number, widely recycled; cite it with its date or find a current figure rather than implying it is contemporary. The "15–20% of retailers use dynamic pricing" figure lacks a clean primary source; present it as an industry estimate. - Walmart denies its shelf-label and markdown tools constitute "dynamic pricing"; quote the denial. *The People's Republic of Walmart* is an avowed polemic, and its sharpest critics (Mises Institute) are ideologically committed — weigh both sides rather than adopting either. - **Seductive dead ends to avoid explicitly:** (a) claiming the fixed price was invented by one man — it was a Quaker → Stewart → Wanamaker layering; (b) collapsing dynamic into personalized pricing — the NY statutory line is the whole distinction; (c) claiming computation "defeats Mises" — it defeats the *complexity* version, not the *valuation* version; (d) romanticizing "the market" as a lost Hayekian ideal — administered and monopoly prices long predate algorithms, as Gardiner Means showed in the 1930s.