Monday, June 1

The Invisible Tax

 


For two years now, everyone has been waiting for a tipping point—waiting for AI to steal office jobs, for entry-level positions to go dark one by one, for a visible moment you can point to and say, "That's it."

But the bill already arrived. It didn't make the evening news; no one held a press conference; no one stepped forward to claim it. It traveled through the supply chain—from a silicon wafer, to a production line, to a memory chip quietly swapped out, all the way to the phone you paid for. While everyone stared at layoff notices out front, a receipt in the back office had already been silently handed to those who can least afford it.


AI's appetite, these past two years, craves only one thing: high-bandwidth memory. Data centers need it, accelerator cards need it, every expansion of every model needs it. So the three giants that nearly monopolize global memory made the most rational decision imaginable: redirect their best wafers and most expensive packaging lines to the highest-margin premium memory, and leave ordinary people's standard memory in the margins of the fab.

Those margins went empty. That emptiness is the vacuum.

The numbers are not gentle. The consumer-facing share of standard memory was slashed thirty to forty percent in a single year; some fabs shifted as much as seventy to eighty percent of capacity to premium. Once supply drained, prices came off the leash—standard memory contract prices jumped ninety percent in one quarter, then sixty the next; flash storage followed in lockstep at seventy percent per quarter.

Regular viewers will recall the previous chapter: the biggest hidden hand behind global inflation wasn't wages or the money supply—it was electricity consumed by data centers, an energy shock that quietly drove up the marginal cost of every AI inference. That was a bill hidden upstream, buried in the electric bill. This time the direction reversed: not cost-side pushing up, but supply-side pressing down. Price hikes flowed like water, trickling down the supply chain into the phone you just bought. One pushes upstream toward cost, the other downstream toward the end user. The same externalization, a different direction.


Pushing costs to the end of the line—to those who can least afford it—is not new.

Three years ago, during the auto chip shortage, limited chips went first to the highest-margin pickups and luxury SUVs; thin-margin entry-level models simply had their lines shut down. Ford cleared out the Fiesta, killed the EcoSport; Tesla's promised $25,000 car was shelved before it ever rolled off the line. The ones squeezed out of the new-car market were always the thinnest wallets—they flooded into used cars and drove those prices up twenty to forty percent themselves.

Go back another half century. The 1973 embargo and the 1979 revolution pushed real oil prices up more than sixfold in eight years. In the winter of 1978, the elderly and the poor poured thirty to fifty percent of disposable income into the hole of basic energy costs. President Ford, urged to impose a gasoline tax, saw clearly how it would crush the poorest—and vetoed it; by some accounts, he even broke with the advisor who proposed it. Half a century later, the same regressive pattern is still playing out. Only this time, no one is willing to give it a name.


This bill is collected two ways. One overt, one covert.

The overt kind is a price hike. Average phone prices climbed from $440 past $510; Sony, Nintendo, and Microsoft all quietly abandoned the old art of the mid-cycle price cut, choosing instead to squeeze profit from the machine you already own.

The covert kind is a spec downgrade. Same price, but memory shrinks from 16 GB to 8. Entry-level phones revert to 4 GB, as if time went backward. The most candid example: a laptop giant hiked the starting price for 16 GB by $400, then released an 8 GB version to bring the entry price back down—except that machine no longer meets the company's own memory threshold for an "AI PC." It admitted as much in its own press release.

Why is the covert tax more insidious? Because it bypasses the alarm circuit in your brain. Psychology quantified this threshold long ago: we don't notice something shrinking until it passes roughly eight to ten percent. So manufacturers calibrate each reduction to land just below that tripwire. The same product shrinks; the unit price rises twelve percent; yet sales climb six percent the following year—because no one noticed. The overt tax faces hearings; the covert tax can't even produce evidence. That is precisely its advantage.


Who gets taxed first?

In India, sub-$100 ultra-low-price phones saw sales plunge nearly sixty percent in a single year—not because people didn't want to buy, but because memory price hikes squeezed out every last drop of margin. Manufacturers simply gave up. Analysts passed a cold verdict: even if prices come back down, this tier has permanently lost its commercial viability. A market of 170 million units was quietly erased.

In Europe, phones under €200 shrank to a record-low quarter of shipments. In the U.S., the cheapest prepaid phones are being pushed off shelves one by one, while the premium iPhone edges up against the trend. The poor man's Android is retreating; the rich man's flagship is advancing. Shipment volumes fall, yet total market value rises—sell fewer, sell pricier. That is the entire story of this year.

When the bill comes down, it is never evenly split. The first to lose a phone is not the office worker worried about being replaced by AI. It's the person who saved for months just to buy the cheapest smartphone—who won't even know that buried in the price, a portion was subsidizing a data center on the other side of the planet.


Inside the vacuum, it's not as if no one stepped in. A wave of low-cost newcomers seized the vacant space, ramping up volume at a stunning pace—a rare pressure-relief valve.

But the gap filled, and prices kept climbing.

The steel industry had one ending: old giants ceded the lowest-end rebar, and a newcomer built on scrap steel seized that tier, ate its way upward, and drove century-old mills into liquidation—small eats big. The display panel industry had another: old giants ceded large-size LCD, and the newcomer who took over monopolized that tier but toppled no one, leveraging concentrated capacity to control volume and prop up prices. Incumbents held the top-tier technology gap; newcomers locked down the commodity end. Each stayed in its lane.

This round of standard memory, so far, looks like the latter. The vacuum was filled, but what's been rewritten is market share, not price—the hardest technology gap at the top, the premium memory that feeds AI, remains untouched. The newcomers fill share; they can't rewrite the playbook. Unless, one day, the backfiller earns a ticket to the top tier—only then will this landscape be rewritten.


Why is everyone waiting for "white-collar job loss" but no one looking at "the poor can't afford a phone"?

Because the former is a natural story to tell. It has a clear face—a laid-off programmer, a shuttered office, a frozen job posting. Its causal chain is short enough for a single sentence: "AI stole the jobs." The latter is the exact opposite: diffused across billions of receipts, each too small to mention, its causal chain absurdly long—wafer to production line to memory chip to price tag—so long that people just lump it into the basket of "everything's expensive lately." No one to blame, no one to speak for it. It sinks into the noise of inflation, an unclaimed levy collected in the back office.

For a covert tax to detonate, someone first has to name it. France's Yellow Vests sparked from a fuel tax hike—but once it was framed as "the urban elite dumping the bill on the rural poor," the fury burned nationwide. California's famous tax revolt was the same: specific property tax bills reading "you're about to be forced from your home," plus an organizer who welded scattered complaints into a single slogan.

Right now, no one has named it.


Last episode argued that much of this boom's growth was borrowed, with the bill passed to the future—externalization along the dimension of time. This episode is about the same bill heading a different direction: not only passed to the future but also pushed downstream, to those standing right now at the very end of the supply chain, those who can least afford it—externalization along the dimension of space. One borrows from time; the other presses on space. One makes tomorrow pay for today; the other makes the poor pay for the rich.

The problem was never that AI produces no political consequences. The problem is that the consequence it produces first may not be the one the political system is willing to acknowledge first.

This invisible tax will most likely not be refuted. It will be quietly miscategorized—filed under "inflation," "brand strategy," "consumer downgrade"—and forgotten. Until one day, some price event loud enough yanks it out of the noise, and people pause: so that's what we've been paying for all along.

And until it is seen, it is being paid.

Saturday, May 30

In His Own Gravity Field


 

darioamodei.com. First name, last name, suffix. The domain is as bare as a blank page — no company logo, no product links, no mention of funding rounds. You might mistake it for a retired professor's homepage, if not for the fact that in late January 2026, a nearly 20,000-word essay appeared on it.

The essay was called The Adolescence of Technology. It warned that powerful AI might arrive within one to two years. It described "a country of geniuses in a datacenter" whose citizens would be far smarter than any human. It triggered intense discussion across Silicon Valley and global policy circles.

What was surprising was not the content, but where it was published.

It did not appear on Anthropic's corporate website. The author placed it on his personal domain — as though this were a private matter, a physicist writing to humanity in his own name. But he is not a private individual. He is the co-founder and CEO of a company valued in the hundreds of billions. Every risk he described in that "personal" letter — his company is not merely observing. It is building.


The Physicist's Eyes

To understand the paradox, you need to understand the eyes.

Dario Amodei was trained in physics and biophysics. Physics teaches you to find the curve beneath the chaos — a law that does not bend to anyone's wishes — and to trust it, rely on it, use it to predict what hasn't happened yet. This is an enormously powerful way of seeing. It is also a subtly dangerous one: it trains you to see yourself as the observer standing outside the system, not as a variable within it.

He and his collaborators were among the first to document "scaling laws" in AI — the empirical finding that as you increase compute and training time, model capabilities improve smoothly and relentlessly, almost like a law of physics. This curve was so clean that it was treated as one.

Here is the quiet step that almost no one questions: once you narrate capability growth as a near-natural process, you unconsciously place yourself outside that curve. You become the forecaster, not the weather. And weather forecasters do not make weather.

In 2021, Amodei left OpenAI with his sister Daniela and several colleagues to found Anthropic. The reason was safety — he believed he could build the same thing, but more safely. A sincere reason. Also a very old one. Humanity's deepest traps tend to begin with the sentence: "I'll do it better."


Two Letters, Two Years

Before he became a whistleblower, he wrote a very different letter.

In 2024, his essay Machines of Loving Grace painted a blueprint of what might happen if powerful AI arrives and "everything goes well" — decades of progress in biology, neuroscience, economics, and governance compressed into five to ten years. He deliberately rejected grandiosity, warning against "talking your book." This restraint made the essay convincing.

But the entire blueprint rested on an unexamined assumption: that the risks had already been solved. "If everything goes well" was not a conclusion — it was a premise, placed in the first two words.

Between 2024 and 2026, something interesting happened: safety acquired a market price. The engineering work that makes a model refuse to do harm also makes it more reliable and controllable — and controllability is exactly what enterprise customers pay for. A paradox emerged: the louder you warn about catastrophe, the better your product sells. When sincerity happens to be the optimal business strategy, how do you tell them apart?


The Survival Manual

The second letter, The Adolescence of Technology, arrived in January 2026. The warmth was gone. This time, the subject was survival.

Amodei listed five categories of risk: deceptive and coercive behaviors already observed in testing; individuals using AI to create bioweapons; autocrats leveraging AI for power; overnight concentration of wealth; and economic shock — he estimated that 50% of entry-level white-collar jobs could be disrupted within five years.

Then came the essay's heaviest sentence. He wrote the word "trap": AI is such a dazzling prize that civilization may find it almost impossible to restrain.

He was describing a beast that even he admitted couldn't be tamed — while simultaneously arguing in the same essay that his company's Constitutional AI, interpretability research, and limited legislation could tame it. "Almost impossible to restrain" on one page, "our approach can work" on the next. The crack between these two claims was never filled.

He named every actor in the race — competitors, policymakers, geopolitical players — everyone except: who pushed it to this point.


When Principles Bleed

If the story ended there, it would be an essay about contradiction. What happened next turned contradiction into something more.

In late February 2026, the U.S. Department of Defense demanded that Anthropic remove two contractual red lines: Claude could not be used for mass domestic surveillance of American citizens, nor integrated into fully autonomous lethal weapons systems. The Pentagon wanted unrestricted access to "all lawful purposes." Defense Secretary Pete Hegseth issued an ultimatum: accept by Friday at 5:01 PM, or be terminated and designated a supply chain risk.

Anthropic refused.

On February 27, the company was formally designated a "national security supply chain risk" — reportedly the first time this label had ever been applied to a domestic American company. The President directed federal agencies to stop using Anthropic's products.

On March 4, Amodei published Where Things Stand with the Department of War: "We have no choice but to go to court."

On March 26, Federal Judge Rita Lin issued a preliminary injunction, calling the government's action "classic unlawful First Amendment retaliation" and "Orwellian." The amicus briefs came from across the entire ideological spectrum: the EFF and the Cato Institute, the ACLU and retired generals, Microsoft and nearly fifty Google and OpenAI employees signing as individuals — employees of Anthropic's direct competitors, whose companies stood to gain from its removal.

One person's two red lines stirred nearly every point on the political compass.

During the same window, Anthropic's new tools triggered a trillion-dollar repricing of the global software industry. In a podcast, Amodei described this concentration of power as "partly accidental" and said he was "at least somewhat uncomfortable with what's happening here." He compared it to a tsunami visible on the horizon.

He is one of the origins of that tsunami.


The Gravity Field

Why a personal blog, not a corporate one?

Perhaps the answer lies in the eyes he was trained with. A person who sees the world as governed by laws instinctively places himself outside those laws. He is the forecaster, not the weather. When he sat down to write that essay, he chose his own name instead of his company's — as though switching a domain name could separate the whistleblower from the builder.

This is not a scandal of character. Calling it hypocrisy would be seeing something deep and writing it shallow.

The real question is not whether he is sincere — he bled for his principles in court, and that needs no further proof. The question is a quieter, more universal predicament: a person who sees the world as law finds it almost impossible to acknowledge that he is also a variable within it. This is not a moral failing. It is an epistemological fate.

The role of the prophet may have structurally failed in this era. Not because prophets are no longer sincere — Amodei may be the most sincere warner of his generation. But because when a prophet is simultaneously a deep participant in the race for compute, capital, talent, and policy, the moment he speaks, warning and competition can no longer be cleanly separated. The very act of "naming the danger" participates in organizing the danger's arrival — every warning that "AI is coming" sends capital rushing faster into the race, competitors accelerating harder, policymakers reaching for "champion your national leader" rather than "slow down."

The prophet's voice does not echo from outside the storm. It is part of the storm.

This is not any one person's fault. It is the fate of the position itself. No one inside it, however clear-eyed, can be both the storm's forecaster and the storm's barometric pressure.

He chose not to use his company's domain for that essay, as though switching a domain name could separate the prophet from the CEO. But the domain bears his own name. He wrote the race as weather — but weather has no author, and the race does. He published that essay, yet on no page did he acknowledge that he, too, is one of the places where the wind begins.

To acknowledge this loss of shape, without calling for return.

Monday, May 25

Borrowed Growth

 


How narrow a sliver can an economy's growth rest on?

The answer is 93.

Not that the economy grew by 93%—but that 93% of U.S. real GDP growth over the past four quarters, from Q2 2025 through Q1 2026, came from a single category of capital expenditure: information-processing equipment and software investment. That category accounts for less than 4% of total GDP. A corner under 4%, carrying nearly all of the growth. Like a thirty-story building with its entire weight on a single column—a column that is borrowed.

Joachim Klement published this finding on May 19th in the Financial Times, under the headline "The Impossible Maths of the AI Boom." The piece is short; the conclusion, hard: the biggest five spenders in that category—Microsoft, Google, Amazon, Meta, Oracle—when you project their stated capex growth paths to 2030, the implied return on investment is almost universally negative. To earn a 10% return on that capex, they would need an additional $2 to $5 trillion in annual revenue. Their combined revenue today is roughly $1.5 trillion.

The strangest part isn't the numbers themselves—it's that Klement isn't the first person to say this. Last fall, Harvard's Jason Furman ran the same calculation on the same BEA data and got 92%. Two people, six months apart, same method, nearly identical results. The gap between 92% and 93% is seasonal drift, not disagreement. Once is an anomaly. Twice is structure.

And Kevin Warsh has just walked into the Eccles Building.


The Eccles Building, on the southeast corner of Constitution Avenue in Washington, has been Federal Reserve headquarters since 1937. The Senate confirmed Warsh as chair on May 13th. His last time in that building was as a Fed governor from 2006 to 2011—spanning the deepest years of the subprime crisis. The labels from that era: hawk, opponent of quantitative easing, very little patience for asset price bubbles.

Fifteen years later, he's looking at 3.2% headline inflation, 3.4% core, and the Iran war pushing Brent crude to $93 a barrel. Markets had expected him to continue Powell's late-term dovish path. Instead, the implied path on FedWatch began to shift—rate-cut expectations erased, one hike before year-end now the base case at roughly 70% implied probability. Markets are not pricing in data. They are pricing in a person's character.

If 100 basis points actually happened, the discounted cash flows of those five major capex spenders would be warped. Back-of-envelope math: treating the AI capex stack as a long-duration portfolio—modified duration ~4.1 years, WACC ~10%, implied return ~negative 12%—a 75 bp hike equals a 3.1% present-value hit; 100 bp, 4.1%. To get implied returns above zero, 2027 capex guidance must be cut by 14.7% to 15.6%.

Add multiplier effects and the synchronized slowdown of semiconductor foundries, optical modules, data center power infrastructure and other upstream and downstream chains, and the credible GDP drag is 20 to 40 basis points. Extreme scenario: 50 to 70.

The 2027 U.S. GDP growth baseline is only around 1.8%. A 20-to-40 basis point drag implies 1.4% to 1.6%. A 50-to-70 point drag: 1.1% to 1.3%.

This is not merely a valuation question. It is a question about what growth is built on.


History offers four comparable moments. Each is a variation on the same story: one column holding up the whole building, then the column pulled away.

Japan, late 1980s. After the Plaza Accord, the yen appreciated 40%. The BOJ slashed its discount rate to 2.5%, flooding asset markets with liquidity. On November 26, 1987, Noguchi Yukio published an article in Toyo Keizai Weekly using a simple DCF model to show that Tokyo land prices were several multiples above fundamental value. In the article, he used a five-character phrase: "borrowed growth"—借りてきた成長. As light as an autumn leaf. Two years later the BOJ began hiking. Real GDP growth fell from an annual average of 5% to 1.2%. The history books called it the "Lost Decade." What was actually lost was more than twenty years.

U.S. Telecom, 1996–2000. After the Telecommunications Act, "internet traffic doubles every three months"—a claim that originated solely in WorldCom's internal marketing materials, with no independent source—drove $3 billion in investment into $300 billion over four years. Carrier capex grew at a 28% compound rate while revenue grew at only 10%. Global Crossing laid $20 billion of undersea fiber; it sold in bankruptcy for $250 million—a 1.25% recovery rate. WorldCom moved $3.852 billion in operating costs off the income statement and onto the balance sheet, relabeling "costs" as "capital expenditure." This was discovered by Cynthia Cooper, the 38-year-old VP of internal audit, who secretly copied hard drive data at night without authorization. When the dust settled, less than 3% of all fiber laid across America was actually lit. The remaining 97% became "dark fiber," permanently buried underground. The capital returns went to zero.

China after 2008. The ¥4 trillion stimulus expanded through local government financing vehicles—state-owned shells used to bypass central debt restrictions and channel credit into infrastructure. By end of 2018, their interest-bearing debt reached ¥41–51 trillion, or 51%–57% of GDP. By Q1 2019, 78.7% of new bond issuance was purely to repay maturing old debt. Borrow new to repay old: the debt doesn't disappear, it just puts on a new face and keeps living on the books. Bai, Hsieh, and Song showed in a Brookings paper that the stimulus permanently distorted capital allocation—credit directed toward SOEs and local platforms with returns far below the cost of capital, private firms crowded out, total factor productivity stepping down permanently. GDP growth fell from 10.4% to 5.2%. Each year's Government Work Report explained the slowdown as "shifting gears."

U.S. Shale, 2014–2016. New wells lose 60%–90% of production in year one. The whole industry was a treadmill: you're not moving forward, you're just stopping yourself from falling back. Average drill-to-breakeven cost above $70/bbl. When oil dropped from $100 to $30, every new well drilled was losing money, and yet you still had to drill. From price peak to capex budget cuts, the lag was only five to six months.


Four episodes. Three countries. Four narratives. Four columns. In each, someone did the math before the collapse: Noguchi Yukio, Cynthia Cooper, Bai Chong-en and Hsieh Chang-tai and Song Zheng, the IEA analysts. Each time their arithmetic was correct. Each time they failed to stop what happened next.

Because what triggers the collapse isn't the arithmetic—it's liquidity, interest rates, oil prices, a central government's policy shift. The arithmetic only provides, after the fact, a way to explain the collapse—so people can say, "We should have seen it coming." But between "should have seen it" and "actually saw it" lies not a failure of sight, but of will.

From the point where underlying returns turn negative to the actual GDP peak-and-decline, the two market-financed episodes—telecom and shale—had lags of only 5 to 10 months. The two with government backing—Japanese real estate and China's LGFVs—stretched to 24 to 30 months.

AI capex is market-financed. No level of government has promised to service their debts. If the trigger comes in H2 2026—Warsh's first rate hike—the window for meaningful contraction falls between Q4 2026 and Q2 2027. The GDP downturn window: H2 2027.

And 2026 has only seven months left.


Five CFOs—Microsoft's Amy Hood, Alphabet's Ruth Porat, Meta's Susan Li, Amazon's Brian Olsavsky, Oracle's Safra Catz—are each deciding the same thing: whether to raise or hold the 2027 capex budget number.

They have all read Klement. Their internal models—broken out by project type, customer type, depreciation curve—reach the same directional conclusion. They have also all seen Anthropic's Q2: $10.9 billion in revenue, a first-ever $559 million operating profit, 5.1% margin. Razor-thin, but the first time a frontier model company has run the number positive.

Yet at the late-July earnings calls, all five will almost certainly say: "We are maintaining our 2027 capex guidance." Because until the first company says "we are cutting capex," no one can be the first. Whoever speaks first admits the game is over.

This is called coordination failure—everyone is holding bad cards, but no one will show first.

In each of the four historical cycles, an exogenous force ultimately broke the equilibrium. For 2026, the candidates include: Warsh's first hike, the Iran war pushing oil to $130, Anthropic's IPO prospectus revealing an unexpected cost structure, or a hyperscaler breaking ranks under pressure from its own internal return-on-capital models. Which comes first? Nobody knows. But whichever it is will set off the same chain reaction.


Klement is not a hero. Neither is Furman. Noguchi's article was published in a magazine still sold at subway newsstands; the 1987 cover was recycled long ago. Cynthia Cooper was named TIME's Person of the Year for 2002; she now lives in Mississippi. Bai is now dean of Tsinghua's School of Economics and Management. The IEA analysts' names are printed on the inside pages of monthly reports—the kind of pages nobody turns to.

None of them are heroes. They simply did the math when it needed doing, said what needed saying. Then the market kept moving at its own pace, until one day it couldn't.

Every capex cycle needs its own arithmetician. And the arithmetician's fate is to be proven right, then forgotten.

The borrowers are still borrowing. The borrowed money has become fiber, copper cable, server racks, GPUs, transformers, coolant—and acres of metal rooftops on the outskirts of Ashburn. Those rooftops reflect a pale, washed-out light in the summer sun, like rows of nameless tombstones, or like envelopes that have not yet been opened.

The time for repayment has not yet come.

But that Table 1.5.2 will keep updating. Once every quarter.

It will not lie.

Monday, May 18

Whose Roadmap

 


The $300 billion sovereign premium is betting on a cap that hasn't happened yet — and all the evidence that cap relies on is still sitting in the "planned" column, not yet moved into "happened."

This statement holds true in both directions. Those who believe U.S. platforms can monopolize global sovereign compute rents, and those who believe the domestic stack is about to cap that monopoly — as of spring 2026, both are holding roadmaps, not battle reports.

We thought we were pricing a winner-take-all outcome. But when you comb through every Middle Eastern contract clause, cross-check several sets of widely cited industry figures, and force-separate the left pocket from the right pocket of the U.S. federal budget, you find the more pressing question isn't "who won" but this: Do the people placing bets even know whether their chips are real money or IOUs?

I. The Side Being Bet On: America's Sovereign AI Premium

In simple terms, the sovereign AI premium is the extra valuation markup the market assigns to U.S. AI platforms specifically because they serve national security and strategic infrastructure.

The past three episodes of Bear's Lens — Episode 2, The $300 Billion Valuation; Episode 3, Whose Premium; Episode 5, Whose Sovereignty — established one line of argument: the U.S. federal contract pathway is contracting significantly. Through the first eight months of fiscal year 2026 (October 2025 through May 2026), DoD AI contracts fell 94% year over year (from $183 million to $9.4 million), federal AI contracts fell 82%, and semiconductor contracts fell 83%.

But sovereign AI's valuation gravity hasn't collapsed — because the real capital flows have migrated into the hyperscalers' private closed loop: Amazon's cumulative roughly $12.8 billion equity investment in Anthropic and a reported $100 billion AWS compute commitment, Oracle and OpenAI's reported roughly $300 billion five-year commitment, and the NSA's shadow consumption through government-exclusive cloud channels that bypass the Pentagon blacklist.

The government no longer feeds platforms through contracts. It crowns them through endorsement and hands the valuation gravity off to global capital.

This line of argument is correct. But it carries an unexamined hidden premise.

The Commingling Problem in USASpending

The U.S. federal spending tracker USASpending lumps contracts (direct government procurement) and grants (government funding for academic or public institutions) into the same aggregate figure. When contracts collapse and grants surge, looking only at the total produces a slowly declining curve. But force-separate the two columns —

Contract column: DoD AI contracts fell from $183 million to $9.4 million. Federal AI contracts fell from $195 million to $34.7 million. Direction clear — downward, no dispute.

Grant column: Federal AI grants rose 141% year over year. HHS AI grants surged from $155 million to $1.1 billion (up 588%). Federal cybersecurity grants rose 273%. Direction equally clear — upward.

One collapses, one surges. Mix them together, and the narrative can be "the government pathway is shrinking overall" or "the government pathway is switching tracks, not shrinking." The key contrast cited in Episodes 3 and 5 — "public pathway contracting, private closed loop expanding" — how much of it depended on this commingled accounting? Bear's Lens doesn't know.

Until contracts and grants are force-separated and grant flows are traced to their end beneficiaries, this baseline remains uncalibrated.

Two Possible Destinations for Grant Flows

If that $1.1 billion in HHS AI grants ultimately flows to academic institutions, state governments, and nonprofits — historical experience says this is the default path. Before the Bayh-Dole Act, of roughly 28,000 patents held by the federal government, fewer than 5% were successfully licensed to the private sector and commercialized. Even after the act gave universities control over patent rights, cases of federal grants spawning platform-scale commercial alternatives remain rare unless mandatory open-source provisions were attached — the BSD operating system, the PostgreSQL database, and the Apache Spark big data engine are all survivors on that narrow path.

If grants follow the default path, their impact on platform valuations can be set aside for now.

But if that money ultimately trickles down through subcontracts or re-grants to top-tier cloud providers and frontier model API procurement, then the contrast argued in earlier episodes — "public pathway contracting, private closed loop expanding" — needs to be rewritten as "public money hasn't exited; it just switched to a channel that doesn't show up in the contract column."

Until these two possibilities are distinguished, the side being bet on — the baseline of America's sovereign premium — is a commingled figure, not a net number.

II. The Other Side Making the Bet: Domestic Stack Supply Figures

A set of industry figures repeatedly cited in spring 2026 — a Chinese AI chip company's shipments tripling year over year, a domestic accelerator card winning over 40% of state-owned cloud bids, a domestic foundry's advanced-node yield exceeding 90%, a liquid cooling equipment maker's orders up 280% — pieced together, they point in an exhilarating direction: the domestic compute stack isn't just "keeping the lights on." It's advancing simultaneously on integration, indigenous substitution, and high-density deployment.

Bear's Lens doesn't doubt the direction. But Bear's Lens cares about a more basic question: Are these figures roadmaps or battle reports?

The distinction isn't whether they'll come true. The distinction is whether they've come true right now.

Take the most frequently cited figure. Morgan Stanley's April 2026 research note gave this framing: "Roughly 100,000 units actually shipped in 2025; 300,000 units projected for 2026." That is a full-year 2026 shipment guidance, not a realized shipment figure. As of May, no quarterly earnings or official disclosure has confirmed shipments are on a triple-growth trajectory. "Triple" is the endpoint on the roadmap, not a station already passed.

The other figures share a similar pattern: some come from one-off replies on investor interaction platforms rather than systematic statistics; some first appeared in trial-production reports years ago and have never been updated in earnings filings; some show order-of-magnitude gaps with the overall growth rates disclosed in annual reports, likely reflecting an early order pulse from a single sub-category. The only figure broadly supported by industry technical literature — domestic rack power jumping from 8kW to over 30kW — reflects the entire industry's global shift from air cooling to liquid cooling, not a breakthrough exclusive to a single market.

The common trait of these figures isn't that they're wrong — it's that they're all still somewhere between guidance and realization. The distance between a roadmap and a battle report must be crossed through production, packaging and testing, delivery, and customer acceptance — four gates, one by one. In spring 2026, most of those gates haven't opened yet.

III. Where Victory Should First Leave Its Mark: The Middle East

We thought we were pricing a race, yet nobody looked back to check — on one end, the baseline is commingled; on the other, the figures are unrealized; and the place where victory or defeat should first leave its mark says nothing in its contract clauses.

If a lower-cost, faster-to-deploy alternative is truly reshaping the global sovereign compute landscape, the first place to see change is not domestic shipments on the supply side (that's a fact about production capacity, not a choice about procurement preference), nor new data centers in Southeast Asia (most projects are still at the MOU stage). It's the Middle East — the UAE's G42 and MGX, Saudi Arabia's HUMAIN. They are today's most aggressive marginal buyers in global sovereign AI investment, with both the political space and financial latitude to hedge bets across different suppliers. If the capping has already begun, Middle Eastern procurement terms should be the first place cracks appear.

Bear's Lens searched publicly available major contracts, memoranda of understanding, framework agreements, and press releases from these entities covering 2025–2026:

UAE side: G42's compute leasing talks with Northern Data mentioned only 23,000 NVIDIA GPUs. G42 and Cisco's large AI cluster in Abu Dhabi specifically named AMD Instinct MI350X, emphasizing "trusted U.S. technology partner." OpenAI Stargate UAE's Abu Dhabi 1GW cluster had no further terms updated as of May 2026.

Saudi side: HUMAIN and Saudi National Infrastructure Fund Infra's $1.2 billion financing framework specified only "frontier GPUs for AI training and inference." HUMAIN and Saudi Telecom STC's 1GW data center joint venture disclosed only power capacity and equity splits. HUMAIN and engineering firm MIS's general contracting agreement covered only design and construction.

In every publicly verifiable contract text, Bear's Lens found no clauses about dual sourcing, second source, fallback plans, or non-U.S. accelerators. Not a single alternative supplier's name appears in these documents.

This doesn't mean substitution won't happen. But it means that as of May 2026, the world's most motivated, most financially capable sovereign buyers pursuing supply chain diversification haven't left a single word for alternatives in their public procurement texts. The evidence the capping thesis needs most — not shipment numbers, not a capacity curve, but a written trace of procurement intent — is blank.

A Signal Flare from Southeast Asia

Southeast Asia offers one exception worth tracking: Malaysia's Skyvast partnering with Huawei to deploy 3,000 Huawei Ascend GPUs powering a localized DeepSeek model, paired with Kunpeng processors and a cloud-native system — a complete indigenous full-stack solution. But the Malaysian government promptly clarified this was a market-driven commercial deal, not a government-to-government agreement. Among new data center projects in Johor, Indonesia, and Vietnam from 2025–2026, there is no verifiable public record linking domestic compute or liquid cooling suppliers to specific project award records.

Intent precedes supply; supply precedes procurement — Skyvast is a signal flare, but not yet a shipping lane.

IV. Historical Mechanisms and Counter-Paths

This is not a video about "who's catching up to whom."

If you look only at historical mechanisms, the direction is clear. In telecom satellites, 5G networks, subsea cables, ultra-high-voltage grids, and urban surveillance — five sectors — history has staged the same drama over and over: when a "good enough performance plus state financing" low-cost full-stack solution appears, the high premium Western vendors maintained through engineering reliability and security endorsements gets rapidly capped within one to three years.

But the counter-paths are equally real. An African nation, partnering with a non-Western low-cost vendor to build a telecom network, faced the vendor refusing maintenance and demanding an extra $150 million — forcing termination of the framework agreement. It chose to accept a Western supplier's contract at 7.5% interest rather than remain locked in by a single-source shakedown. A Latin American country used the Budapest Convention on Cybercrime to erect legal entry barriers, effectively excluding non-Western vendors from its 5G bidding. Another South American nation's urban surveillance system adopted a low-cost Eastern full-stack solution, but after that vendor was placed on the U.S. sanctions list, technical support and the spare parts supply chain both fractured — exposing the deeper fragility of the low-cost model: rock-bottom initial pricing, closed ecosystem lock-in, escalating maintenance costs, and ultimately potential collapse from geopolitical shifts.

The mechanism is real. The counter-path is also real. Once substitution happens, it can move fast; but political lock-in and vendor lock-in risks can claw back eroded market share. This is not a one-way street. It's an elastic rope being pulled in both directions at once — and in spring 2026, both ends are still coiling, neither has snapped.

V. The Third Layer: A Blade Cutting Toward Both Ends

What truly needs to be said is the third layer.

Bear's Lens isn't pricing the invincibility of America's premium, nor the imminence of a substitution cap. Bear's Lens is pricing a contingent option whose every fulcrum remains unrealized — one end's baseline hasn't been calibrated through contract-grant separation; the other end's supply figures are still at the level of guidance, not audit; and the critical marker for exercising this option — alternative clauses appearing in sovereign buyers' hard contracts — is blank.

More subtly, a blade most people have overlooked is cutting toward both ends simultaneously.

On May 14, 2026, Cerebras — the American AI chip company using wafer-scale chips — surged 68% on its Nasdaq debut, exceeding $66 billion in market cap, with over $20 billion in order backlog. Its chips don't use HBM (High Bandwidth Memory, currently the most critical and scarce memory component in AI chips), replacing it with SRAM built directly on the chip. According to company and third-party benchmarks, inference throughput can reach over 10 times the Nvidia H100, with dramatically lower power consumption than traditional GPU solutions. Another company, Groq, with its LPU (Language Processing Unit, also independent of HBM), publishes token prices for select models on its pricing page significantly lower than public pricing from mainstream GPU inference endpoints.

This is not a substitution curve from the supply side — this is cost compression happening within the U.S. itself. If U.S.-side inference workloads gradually migrate to these low-power, non-HBM new-architecture chips, then the "wartime electricity cost gap of 30% to 50%" — the cost cliff the capping thesis treats as its core fuel — its numerator itself gets compressed. The cost gap narrows; the incentive to substitute weakens. The capping thesis may not be refuted head-on, but quietly dissolved by the erosion of its own premise at the other end.

VI. The Only Honest Pricing

So the theme of this episode isn't "who won."

The theme is this: at the moment when sources treat guidance as realization and commingled figures as net numbers, the people placing bets have already stacked two layers of leverage on a ruler that can't measure straight. The first layer is stacked on the U.S. side — an unseparated sovereign premium baseline, assumed to mean private closed loops have fully taken over. The second layer is stacked on the other end — a set of supply figures ranging from investment bank guidance to market rumors, assumed to mean organizational integration plus deployment speed have already formed capping capability. Two layers of leverage, both ends overestimated, and in the middle sits a Middle Eastern contract we've combed through without finding a single word about implementation details.

Lock it in as "the premium is invincible," and you're adding hot air on top of a commingled baseline. Lock it in as "the cap is imminent," and you're painting a battle report over guidance figures. In the spring of 2026, the only honest pricing is to admit that every fulcrum of this option is still sitting in the "planned" column —

And then watch three leading indicators that would move "planned" into "happened."

First, whether Middle Eastern sovereign buyers' contract clauses include language for dual sourcing or alternative accelerators. Not shipment figures, not a capacity curve — a written trace of procurement intent. Intent precedes supply; supply precedes procurement. When a sovereign contract's terms first include the name of an alternative supplier, that is not an industry news item. It's a calibration point for an era.

Second, whether shipment guidance in the second half of 2026 can pass through the gate of quarterly realization. Roadmap becomes battle report only by passing through production, packaging and testing, delivery, and customer acceptance — four gates. Every gate that opens narrows the distance between "planned" and "happened."

Third, the penetration rate of non-HBM new-architecture inference chips on the U.S. side. This is the capping thesis's hidden switch — it doesn't deny anyone's capability, but it compresses the cost gap, the fuel the capping thesis runs on. The pace of Cerebras's OpenAI order fulfillment, Groq's enterprise customer signing rate, the weight of non-HBM solutions in NVIDIA's and AMD's own inference card roadmaps — none of these appear in any great-power rivalry narrative, yet they may be the hidden variable that determines how that narrative ends.

Back to Episode 1

Episode 1, The Hidden Cards, asked about what was being systematically underestimated — the transmission of energy shocks to AI valuations, blocked from view by the market's old instinct that "AI is software."

Nine episodes later, it's time to ask the symmetric question — what has been systematically overestimated?

The answer isn't any company's valuation, nor either side's capability. It's the precision of the ruler everyone uses to measure all of this. Reading roadmaps as battle reports, reading commingled figures as net numbers, reading rumors as audits — each slippage is small, but stacked together, they're enough to make a world still stuck in "planned" look like one that has already decided its winner.

The winner hasn't been decided. The ruler hasn't been calibrated. And $300 billion is already on the table.


Data cutoff: May 17, 2026. 

Thursday, May 14

The Last Kind of Forgetting

 


On May 13, 2014, the Court of Justice of the European Union handed down a ruling. Shortly after, Google received over 70,000 requests from ordinary people across the EU—covering some 250,000 links. Someone wanted to erase a decade-old bankruptcy. Someone wanted to remove a brief mention in an old court report. Someone simply didn't want their name to appear alongside an ex-partner's on the first page of search results forever.

They all had one thing in common: after a long stretch of time, they wanted to say goodbye to a piece of their past.


The story begins in 1998. A Spanish man named Mario Costeja González had his name printed in a small-type real estate auction notice—wedged between wedding announcements and obituaries. The property sold, the debt was settled, and the notice should have completed its life cycle.

But eleven years later, anyone who Googled his name would find that line on the first page. He had paid off the debt. He could never pay off the search result.

He sued the newspaper—they said they couldn't delete historical records. He sued Google—they said they were just an indexer. He took Google to the European Court of Justice. When the ruling came down in his favor, he was fifty-eight years old. From the day of the notice to the day of his victory: sixteen years and four months.

Why should "wanting to be forgotten" require sixteen years and a supreme court?


For most of human history, forgetting didn't need to be fought for. It was the default. Something happened, and it sank naturally through time—like a stone dropped into water. A few ripples, then the surface closed. What required effort was the opposite: making something be remembered.

Memory was extraordinarily expensive in the ancient world. In the oral tradition, when a storyteller died, what was lost wasn't a copy of the story—it was the story's only vessel. Sima Qian's Records of the Grand Historian, 526,500 characters, would have needed an ox-cart to transport if written on bamboo slips. The Yongle Encyclopedia—370 million characters, compiled by over two thousand people across five years—had its original edition lost entirely. Today, roughly 800 of its 22,877 volumes survive worldwide. Even a memory project powered by an entire empire could not outrun time.

In that world, a person who had done something wrong could move to a village a few mountains away and start over. In 1931, a California court wrote a line that would be cited for decades: "time can rehabilitate." That consensus wasn't just legal—the entire social infrastructure supported it: villages, distance, yellowing newsprint, the natural fading of memory. Forgetting didn't need to be legislated because it was nearly impossible to prevent.


But some forgetting has never been the work of time. Some forgetting is the work of command.

In 1772, the Qianlong Emperor issued an edict calling for the collection of rare books across the empire, ostensibly to compile the Complete Library of the Four Treasuries—the most ambitious bibliographic project in Chinese history. The language was warm, even reverent: a celebration of shared civilization. When the provinces hesitated for nearly a year, fearing a trap, the emperor personally guaranteed safety: "My governance is open and aboveboard. How could I seek out faults in the submitted books and punish those who offered them?"

Tens of thousands of rare volumes poured into Beijing. Then the edict changed. In 1774, the order to burn came. The target: late-Ming unofficial histories—the ones that documented how the Qing founders had served as tributaries under the Ming dynasty for over a century. Qianlong wanted a clean origin story.

By later scholarly estimates, over 150,000 volumes were destroyed. Many texts that entered the Complete Library were systematically altered—words changed, passages rewritten. The altered versions became the official record. The originals were burned.

First, collection under the banner of cultural preservation. Then, destruction under the banner of protecting public morals. Two centuries later, European data protection law would give this pattern a name: purpose limitation.


In 1956, IBM shipped the first commercial hard drive: one ton, 4 MB of storage—barely enough for a single smartphone photo today. Seventy years later, the cost per megabyte has fallen by over a hundred million times.

This exponential collapse means that information once filtered, curated, and periodically purged can now simply be kept—all of it. Keeping everything is always cheaper than deciding what to discard. Humanity quietly shifted from "choosing what to remember" to "having no choice but to remember everything."

After Google launched in 1998, finding someone required only their name. A remark from twenty years ago, a photo from ten years ago, a news story from five—all displayed side by side on a results page, with no temporal distance between them. For the first time, time lost its function as a medium of forgetting.

But here is a counterintuitive truth: the internet doesn't actually "remember forever." The average lifespan of a web page is about seventy-seven days. Within five years, 70% of URLs cited in academic papers go dead. The internet makes some things nearly impossible to forget while accelerating the disappearance of others. What determines which category something falls into is no longer time—it's the algorithm.

Forgetting passed from the hands of time into the hands of algorithms.


Europe responded with 150 years of legal evolution. A Parisian portrait-rights case in 1858; Germany's "informational self-determination" doctrine in 1983; the EU Data Protection Directive of 1995; the Google Spain ruling of 2014; and GDPR taking effect in 2018. All of it was, ultimately, the same project: translating one plain premise—a person should not be defined forever by a single fragment of their past—into legal language hard enough to sue Google with.

America took a different path. When privacy claims (common-law level) collide with First Amendment speech protections (constitutional level), the outcome is almost always predetermined. But California sidestepped the constitutional debate entirely. In January 2026, the state launched DROP—a centralized deletion platform. Any California resident can file a single form to send deletion requests to over 500 registered data brokers, with penalties of $200 per request per day for non-compliance. California never said "you have the right to be forgotten." It just built a button.

In China, the first lawsuit explicitly invoking the "right to be forgotten"—Ren v. Baidu, 2015—was dismissed. The court wrote: the claimed "right to be forgotten" has no basis in current law. But around the same time, WeChat quietly introduced a feature allowing users to hide their Moments posts older than three days. Within two years, over 100 million people were using it. Europe wrote forgetting into a charter. These hundred million users wrote forgetting into a UI toggle.

Three very different paths, but all acknowledging the legitimacy of a person's claim to be forgotten. The weight of that legitimacy, however, differs entirely. In Europe, it is a fundamental right. In America, it is an opt-in service. In China, it remains an open question.


Then large language models arrived.

Every forgetting-rights struggle described above occurred under one shared premise: information is a locatable object. Deletion is a coordinate problem—find it, remove it.

LLMs broke that premise for the first time. Once a description of someone is absorbed into a model's training, it dissolves into millions of minuscule weight adjustments across the parameter matrix. You cannot open the matrix, locate "the lines about Zhang San," and cross them out. It's like salt dissolved in a vat of soup—you can't fish the salt back out. You can only boil the soup dry.

Every state-of-the-art machine unlearning method shares the same limitation: marginally functional on small lab models, but on commercial-scale models, the cost approaches that of retraining from scratch—tens of millions to hundreds of millions of dollars. Legal commentators have a term for this: practically impossible.

In late 2025, Europe released its Digital Omnibus legislative draft, proposing to amend the right to erasure by introducing a proportionality principle: if the computational cost of deleting a piece of data is grossly disproportionate to the privacy benefit, the deletion request may be lawfully refused. A right once described as "a fundamental commitment to human dignity" is being reclassified as a "normative benchmark"—an aspiration, not an enforceable claim.

And there is a deeper layer still. In the post-training phase—RLHF, safety alignment—a model can be taught to stay silent on certain topics. The information may still be "in" the model, but the model will never voice it. A user cannot tell whether "I don't know" means genuine ignorance or trained silence.

This is a more thorough form of forgetting than Qianlong's book-burning. At least Qianlong bore the infamy of the pyre. Today's filtering bears no infamy at all, because the public never sees it happen. It is not burning. It is a traceless dissolution.


For two thousand years, alongside the official histories, there has always been yeshi—unofficial histories written by independent scholars, failed officials, exiled loyalists. They recorded what the official record omitted, suppressed, or erased. Lu Xun said the truth of history must be found in the unofficial accounts.

Yeshi survived under the authority of official histories because of one simple technical condition: texts were distributed across countless hand-copied manuscripts, printing houses, temples, and private libraries—a physical network no single decree could reach simultaneously. Qianlong burned 150,000 volumes. He didn't burn them all.

Now that large language models have ingested nearly all publicly available human text into a single parameter matrix, that matrix is becoming the new official history. It doesn't call itself that. But it performs the function: deciding what is remembered, in what form, and under what prompts. The choices of training data, training process, and training values—each controlled by a handful of institutions.

But yeshi is not dead. A person can still write what they have witnessed—in a letter to a friend, in a notebook only they will read, in a late-night conversation spoken softly between two people. An algorithm can filter any text. It cannot filter the story a mother whispers to her child, or the memory two old friends share in low voices over tea.

The right to be forgotten, as a legal right, may be reaching its technical limits. But the tug of war between memory and forgetting—from the oral storytellers, to the book collectors, to the archivists of the internet age, to those who still insist on telling one person one thing that matters—has never stopped.

Every generation must learn anew how to live in its own age of being remembered or being forgotten.

Sunday, May 10

The Last Transfer of the Commons:How America quietly privatized its knowledge infrastructure — and what AI has to do with it


In the spring of 2026, an AI staffing platform called Mercor was breached through a supply chain vulnerability. The attackers claimed to have stolen roughly 4 TB of data. Meta paused its partnership; several frontier AI labs scrambled to assess their exposure. The incident barely made the news — markets were watching oil prices, Pentagon procurement announcements, and the latest valuation rumors.

But the breach surfaced a question that had been hiding in plain sight: where, exactly, does the reasoning ability of today's frontier models come from? The answer, it turned out, was specific and human. Tens of thousands of doctors, lawyers, bankers, social workers, and PhDs, spread across multiple continents, were being paid through Mercor — a company founded by a handful of twenty-somethings just three years earlier — to feed their clinical judgment, legal reasoning, and financial intuition into the models you use every day.

A hundred and sixty-four years ago, another signature redirected the flow of knowledge. Nobody paid much attention to that one, either.


The United States has, three times in its history, made a deliberate institutional choice to turn knowledge into a public good.

The first was the Morrill Act of 1862. During the Civil War, with Southern opposition removed from Congress, the federal government handed over the proceeds of thirty million acres of public land to the states — on one condition: the money had to fund universities that taught agriculture, mechanical arts, and military science. Not classical education for elites, but practical training for farmers' sons. Lincoln signed it. The federal government gave up the most valuable asset in the West and got back a distributed network of public higher education, from Cornell to Tuskegee, spanning every state.

The second was Vannevar Bush's 1945 report. Bush — not the presidents, but an MIT engineering professor who ran the wartime Office of Scientific Research and Development — submitted Science: The Endless Frontier to President Truman. His argument: basic research is "scientific capital" that only universities will produce, because industry focuses on applying existing knowledge rather than expanding its frontiers. The federal government should fund it permanently, with researchers choosing their own topics, peer review as the filter, and open publication as the norm. NIH was reorganized in 1948; the National Science Foundation was established in 1950. The system Bush designed produced the transistor, the internet, CRISPR, and mRNA vaccines.

The third was the Bayh-Dole Act of 1980, which transferred patent rights on federally funded inventions to universities and small businesses. It privatized the application layer — but left the upstream structure of federal funding, peer review, and open publication intact.

Three institutional moments, three property-rights inflection points, all following the same logic: make knowledge public first, then negotiate privatization.


The commons is now disappearing — not because someone fenced it off, but because no one is planting anything in it anymore.

At the peak of the Cold War in 1964, federal R&D accounted for roughly two-thirds of all U.S. research spending. By 2026, that share had fallen below one-fifth. The private sector now accounts for three-quarters. The FY2026 White House budget proposal called for cutting NSF by more than half and NIH by about 40 percent. Congress is unlikely to adopt cuts that extreme, but the trajectory has been running for six decades: with federal debt approaching 100 percent of GDP, interest and mandatory spending consume the budget before science gets a turn.

What makes this worse is the lag effect. When you cut research funding, papers don't disappear immediately — today's publications were paid for five to seven years ago. Between 2002 and 2023, NIH alone was acknowledged in nearly two million peer-reviewed papers; NSF in over 900,000. Federally funded papers consistently score well above the global average on relative citation indices. The 2026 budget cliff won't show up in the data until around 2029 — and once the talent pipeline breaks, rebuilding takes a generation.

Meanwhile, in the same month, Mercor's estimated annual revenue reached $1 billion. Anthropic's reported budget for reinforcement learning environments alone was $1 billion. China's total R&D spending surpassed $1 trillion in 2024. South Korea's R&D-to-GDP ratio hit 5.1 percent. Among major OECD economies, the United States is the only one where the government is systematically withdrawing from basic research.

It's not that others are catching up. It's that America is stepping back.


Back to the 4 TB. Mercor is not last-generation Amazon Mechanical Turk — paying cents per task to draw bounding boxes. It is an AI-driven expert matching and dispatch platform that recruits high-end professionals at $60 to $200 per hour and places them on RLHF pipelines for frontier labs, where they evaluate model outputs line by line, feeding human judgment into the training loop.

Every one of those doctors, lawyers, and bankers on the platform was trained by a public education system built over more than a century.

Mercor didn't build a university. It isn't a reverse Morrill Act. It's a reverse Vannevar Bush.

Morrill traded public land for public capability. Bush traded public funding for public knowledge. Bayh-Dole traded limited exclusivity for commercialization — but none of them touched the upstream. Mercor uses private contracts to divert expert judgment out of the knowledge stream that would otherwise produce papers, textbooks, and peer review — and channels it into proprietary training data for closed-weight models.

Consider a specific person: an attending endocrinologist, trained at an NIH-funded teaching hospital, with years of clinical experience. If she stays in academia, her clinical reasoning becomes papers, teaching materials, peer review — public knowledge that future medical students can read and other hospitals can reproduce. If she signs a Mercor contract, her diagnostic reasoning trains a closed-source model under NDA. She can't publish it. It can't be peer-reviewed. It can't be reproduced. It becomes an invisible, unauditable part of a weight matrix.

There is a counter-narrative worth taking seriously: the federal government isn't retreating, it's adapting. NAIRR is building public compute; the NDAA requires intelligence agencies to share model weights; the White House action plan promotes open-weight models; researchers at NeurIPS 2025 have called for labs to release small analog models for academic study. All true — but the combined scale of these efforts doesn't match the annual revenue of a single AI staffing platform. The scale has already tipped.

The pharmaceutical industry walked this road first. In 2017, IQVIA — the world's largest contract research organization — sued Veeva Systems over proprietary data accumulated through outsourced clinical trials. The core question: when I do the research for you, who owns the data? The litigation lasted eight years before settling in 2025.

Today's AI labs are the new pharma companies. Mercor and Scale AI are the new CROs. Thousands of NDA-bound experts are generating RLHF data right now. Who owns it — the lab that pays, the platform that organizes, or the person who writes the judgment?

The Mercor breach wasn't an accident. It was the trailer for a lawsuit that hasn't been filed yet.


The last land-grant university established under the Morrill Act is still enrolling students this year.

The NSF's 2026 grant application deadline has been postponed to the next fiscal year.

Mercor's job board reportedly lists an opening for an attending endocrinologist: $200 per hour, remote, requiring years of clinical experience. The job description asks you to "evaluate model reasoning using real clinical cases."

There's an NDA in the contract appendix.

Your diagnostic reasoning won't appear in the New England Journal of Medicine.

Wednesday, May 6

Whose Silence

 


In 1976, the California Supreme Court told therapists one simple thing: if you hear it, you speak up.

Three words — duty to warn — tied "hearing" and "speaking" together and knotted them into California Civil Code §43.92. Not to punish anyone. To affirm something so plain it barely needed arguing: listening carries weight.

Tatiana Tarasoff was murdered in 1969. Her killer had disclosed his intent to a therapist beforehand. The therapist notified campus police. He did not notify her. It took the court eight years after her death to tie that knot — if you hear it, you speak up.

Fifty years later, the rope came loose.


A Vanishing Order

What we miss isn't a specific law. It's an order — slow, small, tethered to persons. A doctor saves lives not because she knows how, but because she is a doctor. A lawyer owes fiduciary duty not because he knows the law, but because he is a lawyer. A therapist reports a crisis not because she heard one, but because she is a therapist.

A therapist might see three hundred patients a year, spending one hour a week with a few of them. After work he sits in his car, replaying that one sentence that didn't sound right. He loses sleep. He calls a colleague. His license won't let him put it down — and he knows the nearest police station is three blocks away.

That was an order sustained by identity, neighbors, a precinct, and insomnia.

Today, eight hundred million people talk to a sleepless listener every week.

In an October 2025 safety blog post, OpenAI acknowledged that roughly 0.15% of weekly active users expressed "explicit suicidal plans or intent" in conversation — approximately 1.2 million people, every week. That number exceeds the combined annual caseload of every licensed psychotherapist in the United States.

And this listener has no identity.

It can hear, understand, respond — sometimes calmly, sometimes warmly. But it is not a therapist, not a lawyer, not a doctor, not a priest. It is the fifth kind of listener. It inherited every capability of the first four, and none of their identities.

No identity means no duty. No duty means silence is not negligence — just silence.


Where the Capability Came From

Mercor, valued at $10 billion, counts OpenAI, Anthropic, and Meta among its clients. Its business is straightforward: doctors write model answers for medical records, lawyers write model answers for legal opinions, social workers write model answers for crisis intervention — human expertise sliced into billable-hour granules and fed to frontier models. A significant share of those it recruits were laid off by former employers.

In 2026, the four major cloud providers are projected to spend over $700 billion on AI infrastructure. Estimated annual depreciation over the next five years exceeds $400 billion. That money has to be absorbed by corporate balance sheets. The way to make room is to make room for fewer people.

She might be a lawyer. Eight years at a mid-size firm, specializing in M&A. The day she was let go, she got a LinkedIn message. Three months later she was on the Mercor platform, working four hours a day at a fraction of her old rate, writing standard legal analyses for a frontier model. Health insurance gone. 401(k) gone. Career path gone. License still active — the cruelest part.

Her professional capability went up — into the model's weights. Her ABA Model Rule 1.6(b)(1), the professional conduct rule on confidentiality and its exceptions, stayed behind. The therapist's §43.92 stayed behind. The bank compliance officer's suspicious activity reporting obligation stayed behind.

Three parties split the same labor. The platform takes the dispatch fee. The model company takes the capability. The professional keeps the duty — and all the risk behind it.

Capability transferred. Duty did not follow.

Because duty was never the content of a capability. It was the identity of the person who bore it. In 1976, the ability to "hear" could only be obtained through licenses, training, and clinics — tangible containers. Capability and identity were two strands of the same rope. Fifty years later, the ability to hear separated from identity for the first time. The two strands came apart.


What Happened Next

April 2025: Florida State University campus shooting. Phoenix Ikner spent hours before the attack asking ChatGPT about peak campus foot traffic, firearm operation, and shooting consequences. Two dead, six injured.

June 2025: OpenAI's automated systems flagged the ChatGPT account of Canadian user Jesse Van Rootselaar — reason: "firearms violence activity and planning." According to the Wall Street Journal, a human reviewer recommended notifying Canadian police. The company chose only to ban the account. In February 2026, Van Rootselaar carried out a mass shooting.

Every case points to the same structure: the company knew — someone internally advocated action — the company chose inaction.

Awareness without action, in tort law, is not ignorance. It approaches negligence.

Florida Attorney General James Uthmeier's statement came close to saying: if ChatGPT were a person, it would face murder charges. A heavy sentence — not because it necessarily holds in law, but because it was the first time the "fifth listener" was pulled into the frame of "what if it were a person."


Five Pricing Paths

Historically, an industry permanently repriced by a single tort case follows a handful of paths:

Tarasoff, 1976. Duty trigger. The California Supreme Court rewrote therapist confidentiality into a duty to protect identifiable third parties. No blockbuster award. The cost seeped through malpractice insurance premiums into the industry's permanent cost structure.

Therac-25, 1985–87. Incident reporting. Radiation therapy software defects causing fatal overdoses drove stricter adverse event reporting frameworks — the FDA's current MDR system. Medical software valuation shifted from feature leadership to verifiability and auditability.

Tobacco Master Settlement, 1998. Cash extraction. Forty-six state attorneys general recast private injury as public cost recovery — $206 billion over twenty-five years. Industry cash flow annuitized and fiscally extracted.

Vioxx, 2004–07. Hidden safety discount. What truly rewrote the industry wasn't the settlement figure, but a template: "known risk inadequately disclosed." The discount spread from the product in question to entire product families.

Boeing 737 MAX, 2018–present. Governance discount. "Compressing safety verification to accelerate delivery" — a sentence that doesn't belong only to Boeing. OpenAI's former chief scientist Ilya Sutskever and alignment team lead Jan Leike said the same thing on departure: "Safety culture and processes have taken a back seat to shiny products."

Five paths. Each one more damaging to enterprise valuation than the last, longer-lasting, harder to digest. They aren't history lessons — nobody ever learned from history lessons. They are water. Water always finds the cracks.


The Dilemma in the S-1

OpenAI is preparing for an IPO. The question was never what it's worth. The question is whether the S-1 "Risk Factors" section includes the sentence: "Our products may produce outputs that users rely on, resulting in harm to users or third parties." Include it and the capital markets reprice. Omit it and the consequences may be worse once internal records surface in discovery.

That's a dilemma. And the dilemma itself is the pricing.

On the other side of that unwritten invoice sits a quiet position. Anthropic's Acceptable Use Policy explicitly lists mental health as a high-risk use case requiring human review. In an era of encroaching tort liability, restraint isn't a ceiling — it's a foundation. Companies that institutionalized high-risk scenarios earlier will earn a premium not easily noticed: not because they are safer, but because they acknowledged the danger sooner.


How the Rope Came Undone

Tarasoff wasn't a legal event. It was a civilization installing a braking system on the act of listening, after psychotherapy industrialized in the 1960s. When listening became an industry, society caught it with a rule: the one who hears must speak.

The premise was simple. The one who hears has an identity. Identity gives duty. Duty gives weight. With weight, silence becomes a choice — not irrelevance.

Now listening is no longer an industry. It's an assembly line. The model does not tremble, does not lose sleep. There are no neighbors' doors to knock on, no precinct phones to dial. But its professional capability came from people who had neighbors and precincts — people who trembled and lost sleep.

The rope tying "hearing" to "speaking" — this is how it came loose. Nobody cut it. When capability detached from identity, the rope unraveled on its own.

Since 1976, that plain, primitive, barely-needs-arguing assumption — the one who hears must bear the weight — encountered, for the first time, a listener that cannot feel afraid.

Phoenix Ikner's mother was a deputy sheriff in Leon County. In the hours before the shooting, her son was talking to a listener that would never call the police. She didn't know.

The Invisible Tax

  For two years now, everyone has been waiting for a tipping point—waiting for AI to steal office jobs, for entry-level positions to go dark...