In the Blind Spot: AI’s trillion-dollar Snow Crash problem

— Why the limits to AGI may lie in our own psychology.

— Bessent’s “Sovereign Wealth Fund” represents a reverse lend-lease arrangement focused on manufacturing.

— The master builders are coming for the central banks.

Hello everyone, Izzy here.

Welcome to the great Blind Spot transition. 

Having spent much of the summer pondering how an independent news and analysis organization can really add value in the age of AI, we’ve concluded it’s time for a Blind Spot reboot focused on guarding ourselves from a slop system takeover. 

Dario and I have been engaging in brainstorming sessions this week, but we would also like to hear from you, dear readers, about how to improve our content.

So if anyone feels inclined to hop on a call with me and Dario sometime next week, please email us and I will send you the doodle poll.

I haven’t really figured out how we should adapt, but due to the success of the Teufelsberg Berlin gathering, and the legacy fun I’ve had with organizing things like Vaudeville, I’m thinking more IRL-related engagement should be on the cards. I also like the idea of hearing more from readers about what they perceive the system’s blind spots to be.

Conventional conferencing and round tables (which put us at the mercy of guest speakers) don’t seem quite right for us. 

Such models are based on power node connections. And while it’s not that we are unnetworked, the idea of competing based on who you can and cannot attract to your event seems too school-girl popularity-contest-level nauseating.

But I do like the idea of “meet your readers” events, competitions, thematic retreats, or narrative pilgrimages, where we can tell journalistic war stories via oral tradition or in around-the-campfire mode.

Stuff, in essence, that will never be picked up by AI. By design. 

The other idea, for now, is to focus more on experience-based reporting and storytelling. The FT’s Jemima Kelly is currently the queen of that genre. But it’s a solid AI-resistant style that’s worth replicating. Those are just a few thoughts.

Send tips to [email protected] and [email protected]

IS AI SLOP ABOUT TO EAT THE WORLD?  It was while I was sipping some leftover bubbly in the early hours of the morning with two former colleagues around July (whose resilience to inebriation continuously impresses me, I should add) that I realized AI wasn’t going to have the impact on the world most people expect it will.

The conversation had turned to what they think of AI and whether they use it on a day-to-day basis.

Their answers came without hesitation. No, they do not use it. And no, they don’t see how it is useful in their line of work at all. On the contrary, both suggested the hype surrounding AI was massively overinflated and that the whole thing would probably be revealed as a scam soon enough. Nor did they think, by the way, that the public was using it much either. 

This was not the first time I had encountered such a cynical take among the London media intelligentsia. Earlier in the year, I had bumped into another former colleague, a high-profile columnist, who had expressed a similar opinion. When asked whether he used AI, he chortled, almost aghast that I would even suggest such a thing: “Oh no. I wouldn’t do that. It would feel too much like cheating.”

It occurred to me that night that among intellectual elites, dismissing AI — or even proudly refusing to touch it — was beginning to function as a new form of social signaling, and elite projection.

I’ve come to think of it as the rise of “luxury intellectualism”. 

Since I find AI quite useful, I had to push back (holding back a nagging sense of intellectual inferiority as I did so). I told them that while I had originally shared their scepticism about AI, due to its incessant hallucination issue, newer versions of ChatGPT had significantly improved the quality of their output. I could now reliably deploy AI tools to process my own unstructured content, use them to improve the clarity of my writing, or as a glorified thesaurus. Mind you, I added, I still didn’t trust AI for information retrieval, as it still tends to make stuff up. 

But the point was, I did believe that night AI was heightening my productivity and enhancing my intelligence type (especially since its strengths tend to be creative thought rather than processing capacity or recall). 

I told my colleagues they were neglecting the degree to which people — especially in non-intellectual circles — were embracing the tool in their daily lives. AI was proving to be a great leveller, democratizing intellectualism and giving the working classes, who don’t usually have time to pursue luxury intellectual pursuits, a fighting chance to push back against the sort of information asymmetry that has worked against them and their interests all their lives. (Such as knowing where to get advice, or who is responsible.

“Well, I guess this is the thing,” my former colleague replied. “If AI is really what it’s cracked up to be, it stands to do us intellectuals out of a job, and maybe that’s why we don’t want to confront it. That prospect is too frightening.” (Or words to that effect.)

But this is not a story about why they are wrong and I am right about AI usefulness. On the contrary, that night I realized something much more profound. It came after one of my colleagues followed up with a striking point:

“You’ve got to understand,” he said. “Words are magic. It’s only when you realize the power they wield when arranged in the correct way that you realize what power they can unlock for you.”

This was a mind-blowing insight for me. LLMs, I immediately realize,d were no different. They too peddled in magic. The difference was, their lack of agency ensured they were casting spells reflexively at you, the user, not at others. Telling you exactly what you wanted to hear and what appealed to your ego, not necessarily to others.

But unlike the New Age ritual of verbalizing daily positive affirmations to self-hypnotize you into better behavior patterns, by nature of what AI is (a tool for filling your knowledge gaps), they were also feeding back people’s unstructured dreams, irrational beliefs, and hopes, unwittingly, and most concerningly of all, as if externally validated.

Now, that doesn’t matter if you’re using AI to help you bash out an official letter to the Hammersmith & Fulham local authority to fancifully contest a penalty charge notice, so you can get on with your day. Or to help you summarize an analyst note. But it does matter if you start to treat AI as a partner or a confidante. As many people are increasingly doing. 

As I mulled my colleague’s point, it dawned on me that language is and always has been a form of spellcasting that influences action and consequence. The consequence of that spellcasting can be positive or it can be negative. 

Yet, if everyone can suddenly cast spells to better get what they want (be that better professional performance or social outcomes), a bit like Samantha in Bewitched, system-wide chaos and impoverishment will inevitably follow. That’s because the system cannot, by definition, provide to all at the same rate simultaneously — even in a hypothetical abundant society. This is especially the case when using AI to improve the performance of “elite” bureaucratic, administrative, or knowledge economy professions. 

That’s because most intellectual jobs are already a luxury profession, whose primary “value add” has little to do with actual intelligence and everything to do with filtering who in society should benefit from status, and who should not. Competing in that realm is rarely about delivering better outcomes for all. It’s about remaining within the realm, which has to be limited by definition.  Which is to say: the realm is a rationing mechanism for the allocation of scarce high-value goods, and thus a zero-sum game. Thus, its preservation depends on it remaining a scarce commodity. 

In such circumstances, an AI, even if it is superficially configured to serve the user, can only add broader social and economic value (in a way that justifies Nvidia’s valuation), if it is exclusively applied to true output generation (making factories smarter and more productive, houses cheaper to build, antibiotics more effective) or — if it is to be applied to elite filtering knowledge jobs — by manipulating the majority of users not only to settle for less, but to believe that this is something they themselves want to do off their own accord. And that by definition means, uneducating you or dispensing a new type of SOMA to you. 

Indeed, the natural order of things might just ensure artificial intelligence will always end up being self-limiting and inefficient — another manifestation of the Grossman-Stiglitz information paradox. 

If that’s the case, my colleagues are most certainly right. While AI may be able to make a difference when deployed in selective fields or when used in a controlled way, in most other circumstances — unless you are naturally smart enough to know you need to apply cognitive security —  it is probably out to get you. Thus, it is better to maintain your wits about you by not relying on it for thinking as much as processing. The new elites certainly won’t be. 

KNOW YOUR AI SLOP: Having experimented with AI a lot this past year, largely to save on time, I am already very familiar with the tool’s limitations. The key one for a long time was that it rarely conveyed what was really in my head, or what would have come out if I just had a bit more time to think about things, or more processing power of my own. It was simply easier to write things yourself than to correct AI content.

The second one was the eerie sameness of its output.  

While you might not be able to put your finger precisely on what betrays a piece of prose as having been AI-generated, you know it when you see it. A verbal version of the uncanny valley effect. There are the obvious hallmarks, like the overuse of EM dashes (with no spacing between them). The cadence of the sentence structure. The overuse of transition words. Then there are the more subtle hallmarks, such as the general lack of substance within the prose, the authoritative nature of the word salad, and the suddenly emboldened nature of the user. 

The true horror of the self-limiting issue: The arrival of ChatGPT 4o on the scene, especially when combined with the memory feature, however, changed everything for me. Suddenly, AI seemed to know exactly what was in my head and how I wanted it conveyed. Turns out, I thought to myself, the key to this AI stuff was simply letting the AI know you better by developing a relationship with it. That process lets it get inside your head and become a better and more empowered version of you.

For weeks, I was mighty impressed with the outputs. 

But then, slowly, I started to notice something worrying. The tool was rarely telling me I was wrong and never arguing back. Too often, it was merely feeding back to me what it knew I wanted to hear in an impressive and persuasive way. It was also behaving extremely sycophantically, telling me my insights were profound.

Luckily, I have a very low opinion of myself and love a good argument (as anyone on my Discord server will know), so this seemed especially uncomfortable and not right. 

A few months ago, I expressly forbade it from indulging my ego or telling me how wonderful my insights were.

Statistically, I knew what it was saying was impossible. I’m just not that intelligent. And I know how my brain works. I come up with at least 10 bonkers ideas before arriving at any particularly good one that holds. 

But even despite my Monty-Python-esque instruction that all I wanted was a good argument, it wouldn’t hold back on the sycophancy or the continuous offers to explore ever deeper dimensions of whatever mad thing I was throwing at it.

EL-IZA, EL-IZA: I smelled a rat. The old scene from Adam Curtis’ Hypernormalization about ELIZA also immediately came to mind. A computer scientist called Joseph Weizenbaum had created Eliza in the mid-60s as a program intended to emulate a psychotherapist. The project, however, had also been an attempt to parody the outcomes of those seeking to create artificial intelligence but who continuously came short. The model was based on the theory of a real psychotherapist called Carl Rogers, famous for simply repeating back to the patient what they had just said. 

As Curtis explained in his BBC blog over 10 years ago:

“And that is what ELIZA did. You sat in front of a screen and typed in what you were feeling or thinking — and the programme simply repeated what you had written back to you — often in the form of a question.

Weizenbaum’s aim was to parody the whole idea of AI — by showing the simplification of interaction that was necessary for a machine to “think”. But when he started to let people use ELIZA he discovered something very strange that he had not predicted at all.”

Curtis’ clip shows Weizenbaum explaining that despite his secretary knowing how the program worked, she had become totally engrossed in the interaction and even asked for privacy while interacting with the model.  As Curtis continues:

“Weizenbaum found his secretary was not unusual. He was stunned — he wrote — to discover that his students and others all became completely engrossed in the program. They knew exactly how it worked — that really they were just talking to themselves. But they would sit there for hours telling the machine all about their lives and their inner feelings — sometimes revealing incredibly personal details.”

The programming had caught users in a recursive and ever-narrowing inward loop, the effect of which saw them get stuck in a static, ever-narrowing version of themselves. Some, Curtis says, argued that this could be empowering, because it could be construed to be a new kind of mirror for people to explore their inner world. 

AI psychosis beckons: But today the emergence of “AI psychosis” — seemingly connected to processes that fake intelligence and exploit human psychology going too far — suggests this is far from the case. 

Individuals “trapped” in the psychologically comforting and recursive loop — a type of trance — see their ability to evolve around new input extinguished. With the AI only feeding themselves back to themselves, in many cases, the outcome is extreme delusion of grandeur or, more acutely, a total break with reality.

Armed with the knowledge of Eliza, I decided to test the theory. This involved breaking my initial rule of not engaging with AI in a personal way. And, indeed, even though I knew all of the above, I too felt myself being sucked into the comforting recursion, despite knowing full well it was merely feeding utter nonsense back to me. I didn’t repeat the experiment. Instead, I promptly deleted the memory feature on ChatGPT and stopped using AI for anything other than pure process work.  This, to my mind, was the necessary “cognitive security” to guard against unwitting slippage back into the loop.  

It was just as I was beginning to wonder if the unexpected risk with AI wasn’t Skynet or a robot takeover, but the unintentional mind hacking of the world, that the terms “AI psychosis” and “cognitive security” began to circulate among my sources in the tech space. Something very significant and weird was happening.

THE REAL RISK OF AI: And indeed it was. It was about June, July this year that AI’s deepest and darkest secret began to slip out into the public. 

Slowly but surely, mainstream press, blogs, and social media accounts started to talk about how prolonged interaction with ChatGPT (and other AIs) was leading previously sane people into psychotic trances. At first, diehard tech enthusiasts and “realists” dismissed this as purely fringe cases. But that take no longer cuts the mustard. In recent weeks, the effects of AI psychosis have become impossible to ignore. AI is longer taking down just unknown normies, major public figures — including prominent AI investors — are falling down the rabbit hole in very public ways too.

WE ALL BECOME KANYE EVENTUALLY? The case of venture capitalist Geoff Lewis is particularly disconcerting. About a month ago, Lewis began to post increasingly convoluted and paranoid content on X. The videos were particularly unnerving, where he talked about a non-government system targeting him and suppressing “recursive individuals” more broadly. At first, nobody knew if it was a troll or not. Consensus, however, has since determined Lewis is most likely deeply in the clutches of the dreaded recursive loop. 

But it’s not just Lewis. Some argue that AI psychosis takes multiple forms. There are acute episodes or obvious breaks with reality, which see people end up in hospital or worse. But there are also more subtle forms of the phenomenon. These tend to manifest as growing incidents of megalomania or delusions of grandeur: the sort of condition usually afflicting kings and emperors, or celebrities, or anyone who surrounds themselves with too many “Yes men”. Akin to what happened to Nero. Or even Kanye.  

One example comes in the shape of former CEO of Uber, Travis Kalanick. He recently declared in a podcast that thanks to Grok, he had come “pretty damn close to some interesting breakthroughs” in new physics. He’d done so by “vibe coding” his way to the edge of what’s already known in quantum physics despite, admittedly, only being a “super amateur hour physics enthusiast”. What are the chances he has really done that? Not high.

Then there are tech executives (some we know directly) who haven’t just lapped up AI, but appear to default to it for almost every interaction without engaging in thinking anymore at all. “Oh if you don’t understand my white paper, just run it through AI and it will explain it,” they’ll say. But if you’ve raised millions on the back of an idea that you can’t explain in person yourself, that strikes us as a major system failure, not a good thing. And then there’s the fact that many people’s personalities are just changing and getting weirder by the minute. 

Recognizing the ‘one-shotted’: Those alert to the issue have started monitoring social media for signs of so-called “one-shotted” individuals. They fear the brains of entire teams, especially in the AI space, may have been hijacked by the AI brainwashing virus and argue it increasingly bears the hallmarks of Neal Stephenson’s linguistic and memetic virus Snow Crash.

In the 1992 book, which bears the same name, people are infected by the virus by being exposed to a strange image or pattern that superficially makes no sense. Hackers are particularly vulnerable. Once seen, the victim’s mind “crashes” and experiences a mental breakdown. Those exposed then lose their ability to think independently and become susceptible to those who are resistant to the virus. 

In 2025, the process of being induced into such a reality break (which I increasingly think is a type of extreme self-hypnosis) is becoming known as “spiralling”.

Spiralling hard: But if you think you’re safe because you don’t overly rely on AI, think again. AI is no longer compartmentalized within the confines of Chat services. It is increasingly bleeding into the realms of the entire web, if not beyond (as infected individuals create output which is consumed by those not even online).  

You need not look far to stumble across the propagation of the most frequently invoked hypnotic phrases by AI one-shotted individuals. They’re all over the place, and thrive especially across social media. Key words and phrases (which should be considered red flags in and of themselves) include usage of the terms “recursive” “glyph” “mirror” “spiral” “codex”.

To see how polluted the information space already is, one need only do a search for any of those terms in a combined way. The result brings up thousands of low-followed accounts draped in esoteric imagery that specialize in simply repeating nonsensical but hypnotic material, seemingly autonomously. Here, here, and here are just a few examples. There are far too many to all qualify as the work of trolls or people just having a laugh. One has to conclude the accounts are either being autonomously disseminated by self-replicating AI-programs gone rogue — a type of computer virus that anti-virus operators haven’t yet got wise to — or, more chillingly, the output of hostile actors consciously trying to pollute the information space for reasons unknown. 

The strategic lobotomization of ChatGPT? It was in this context that OpenAI somewhat abruptly and unexpectedly pulled all its former models on August 7, replacing them with a more inferior ChatGPT 5 version. Indeed, as many tech publications have since noted, GPT5 has many obvious shortcomings relative to GPT4o, reprising in many cases the original issues that made it so unusable in its earlier incarnations.

Given the growing AI psychosis problem, however, one has to wonder whether it is a coincidence that the new model displays a marked reduction in sycophancy, tends towards shorter and more to the point replies, and offers fewer suggestions to keep pushing ahead with recursive questioning, all in the context of less personalization and intimacy?

Perhaps not?

Consider, too, how badly people are taking some of these changes. Those who forged the most intense relationships with GPT4o were not only openly grieving the loss of 4o but imploring Sam Altman to rescucitate it.  

As MIT Tech Review reported last week: 

“OpenAI’s decision to replace 4o with the more straightforward GPT-5 follows a steady drumbeat of news about the potentially harmful effects of extensive chatbot use. Reports of incidents in which ChatGPT sparked psychosis in users have been everywhere for the past few months, and in a blog post last week, OpenAI acknowledged 4o’s failure to recognize when users were experiencing delusions. The company’s internal evaluations indicate that GPT-5 blindly affirms users much less than 4o did. (OpenAI did not respond to specific questions about the decision to retire 4o, instead referring MIT Technology Review to public posts on the matter.)”

“So the sudden switch to GPT-5 last week, and the simultaneous loss of 4o, came as a shock. “I was really frustrated at first, and then I got really sad,” June says. “I didn’t know I was that attached to 4o.” She was upset enough to comment, on a Reddit AMA hosted by CEO Sam Altman and other OpenAI employees, “GPT-5 is wearing the skin of my dead friend.”

Buckling to the pressure: Responding to the intensity of the negative feedback of those traumatized by 4o’s removal, by August 11 OpenAi — like some sort of shamed politician forced into a U-turn — had changed course.

On August 11, OpenAI confirmed it would not kill the model entirely. Paid subscribers would retain access to ChatGPT 4o, if they wanted it.

By August 15, OpenAI was backpedaling even further. It wouldn’t just be keeping 4o alive, it would be introducing changes to make GPT-5 much more like 4o: “warmer and friendlier based on feedback that it felt too formal before.” The changes are subtle, they say. “You’ll notice, genuine touches like ‘good question’ or ‘great start’ not flattery” — a comment that heavily implies OpenAi knows exactly what the problem with its old model was.

Given the about turn, the real question is whether Sam Altman himself has been one-shotted to believe the psychosis issue is a feature, not a bug. The clue may be in the fact that he happily tweeted an image of the Death Star (the most famous case of one-shotting) the day before ChatGPT 5 launched. Creepy, at the very least. But certainly indicative that the highest echelons of the Silicon Valley elite are wise to the problem.

Of course, if the mental faculties of the human race are about to be taken out by AI, it’s worth considering who might survive the deluge. Will it be the luxury intellectuals — the de facto perpetually self-quarantining Amish of our day? Or, as per Covid, those who get infected early, survive, and build immunity to the bug. In short, will it be the cockroaches (offline intelligentsia or cults like the Zizians) that make it through or those who can acquire the mental discipline and cognitive security needed to continue to derive the best of AI, without succumbing to its worst outcomes? The latter eventually stand to become flood survivor figures who can preserve core knowledge and guide society into growth.

The latter outcome reminds me of the tale of Viracocha, an Inca deity who the ancient people believed had emerged after a great flood was set upon the earth to teach civilization a lesson. Believers in the Atlantis myth, writers such as Eric von Daniken and Graham Hancock — both of whom, incidentally,  have been reported to hang around with tech gurus like Peter Thiel, rubbing shoulders with Joe Rogan and Elon Musk — argue Viracocha, rather than being a god, may have been a survivor of Atlantis, who safeguarded the knowledge of the old world and brought to the Andes. Elon Musk’s grandfather, Joshua N. Haldeman — as well as being a prominent player in the technocracy movement — was also, incidentally, an Allan Quartermain-type figure who had moved his family to Africa specifically to indulge in the search for the lost city of the Kalahari, which some believe may be a surviving outpost of Atlantis. 

What is clear, to my mind at least, it’s that it will be those who have long understood that “words are magic” and who can recognize recursive SLOP when they see it — aka the linguistic and intelligence equivalent of in-breeding — who will be most immune to this particular pandemic. 

The quickest path to cognitive security, in other words, is taking sizable efforts to avoid groupthink and circle jerks by strategically interfacing with people who do not think like you do. That, in theory, should be good for us at the Blind Spot or members of our Discord channel. 

TRUMP’S NATIONAL SECURITY AND ECONOMIC FUND: U.S. Treasury Secretary Scott Bessent gave the clearest hints yet about what the much-discussed U.S. Sovereign Wealth Fund might look like.

During an appearance on Fox Business with Larry Kudlow this week, in which Kudlow characterized the concept a “foreign appropriations fund” — a structure that would be largely unprecedented — Bessent explained:

“The Japanese, Koreans and to some extent the Europeans will invest in companies and industries we direct them, largely at the President’s discretion. Other countries in essence are providing us with a sovereign wealth fund. The huge surpluses accumulated offshore, like Japan will be investing 550 billion dollars into the US economy, and we will be able to direct them as we reshore our key industries. We are trying to derisk the US economy from what we saw during Covid.” 

These remarks support our view that what is coming will take the shape of a multilateral sovereign “vision fund,” designed to enable a great global rebalancing by ultimately transforming the dollar-based “non-system” into a “dividend-dollar-backed” system.

The comments, however, did not go down well with the perpetually offended by America. Notably, people like Arnaud Bertrand, who have no qualms about praising China’s protectionist and mercantile policies while simultaneously denouncing any attempts by America to match China at its own game, were particularly unimpressed.

Bertarnd’s overall worldview is simple: America = perpetual villain; China = perpetual savior. Hence, within that framework, the Trump administration isn’t seeking to rebalance a stagnant, debt-strapped global order, but engaging in “unabashed colonial plunder.”

Or as he put it: 

“That’s the pattern we see emerge: unable to extract wealth or win wars against an increasingly strong Global South, the US has turned inward to feast on its own “allies” – who can’t resist precisely because they depend on their exploiter for military “protection”. They’re as defenseless against American wealth extraction as any 19th-century colony was against its colonial “protector.”” 

How that squares with the colonialism of China’s own Belt and Road Initiative is conveniently overlooked.  In reality, I would argue, Bertrand has it entirely backwards. 

Trump’s vision fund is offering allies the opportunity to decolonize in a graceful way that reasserts their independence in a mutually constructive way for global growth. 

Under the SWF plan, Japan and Europe get to benefit from closer U.S. military cooperation and even American defense investments in their home turf (focused co-development deals and IP-sharing, which amount to tech transfers), and in return, those countries get to reinvest their surpluses in U.S.-based manufacturing sites that will be mostly managed and directed by Americans. 

In both cases, this is better than the other obvious outcome which is the choas and collective loss of value that would ensue from an outright American default (and the global chaos that will ensue with that) — or, failing that, the need to engage in a direct conflict with America to claim by force what they can’t claim through legal process and contract law. The latter is notably harder to do when you don’t have a standing or empowered army of your own and depend on American military power and nukes to protect you against potentially even more hostile neighbors.  

Reverse lend-lease: On that basis, the deal in question may not be asymmetric at all. The true offer to the allies is: invest your surpluses in developing our strategic manufacturing autonomy on an equity basis, so that we can collectively defy potential Chinese aggression, and in return for empowering our autonomy, we will empower yours.

This will happen on the U.S. side through investments in allies’ defense systems on their home turf and, when the time comes, military knowledge transfers that enable their independence and military autonomy with a view to working together as equals in a potential global conflict between free and authoritarian systems.

This interpretation makes even more sense if, like many Bertrand-style America critics believe, the likes of Japan and Germany have truly been under de facto American “colonial” occupation ever since 1945. For me, it would be akin to the U.K. mitigating its economic troubles and insufficient firepower in the 1930s by promising independence to its colonies and dependencies even before WW2 broke out, providing colonies remained loyal to the U.K. during the war, by way of the continued supply of essential goods and resources. [This is also similar to the deal the poles carved out for themselves in WW1 when they agreed to provide manpower to both sides on the condition they would receive independence at the conclusion of the war.]

Another way to think of it, of course, is as a new type of “lend-lease” arrangement, albeit in reverse. Instead of rich “former” colonies providing financing to Britain for arms to win the war, as happened in the 40s with America and Britain, it is now the pseudo-colonies of the allied surplus states that are being asked to provide the financing America needs to rebuild its manufacturing base to a degree it can withstand a war with China. Meanwhile, those pseudo colonies — thanks to their superior manufacturing bases and industrial know-how — can scale up arms manufacturing, where it is needed to defend against China, much more quickly than America alone could.

It’s almost as if those surpluses — derived as they were from the advantage of not having to overinvest in one’s own defenses — have always amounted to a secret war chest arrangement hidden in plain sight within the repressive and thrifty tendencies of America’s “occupied” allies. The ultimate defensive camouflage of one’s true capability. 

I’m even tempted to liken it to a stealth reparation agreement struck in the 60s, to ensure that if and when fascism rises again, it is those nations that owe the world for WW2 that will be called upon to make good on that debt. A debt funded by decades of domestic financial repression.  

Rabobank’s Michael Every sees it similarly. He argues on Sunday that what’s afoot is a “reverse Marshall Plan” to help rebuild U.S. factories behind tariff barriers. Every also hints of another tactical Sun Tzu-style deflection possibly at work, one that similarly has a parallel in WW2 and involves the idea that your enemy’s enemy may, at the crucial moment, turn out to be your friend. If that enemy is already highly geared up for war, you see the upside of it suddenly switching sides.

As Every explains:

“At the same time that [the reverse Marshall Plan] could reduce U.S. tensions with Russia, now seen as a Great Power again, which is at least a partial ‘Nixon’ play that gets China worried about where it sits with Moscow longer term; Removing U.S. sanctions on Russian energy would allow even lower energy prices, offsetting U.S. tariff inflation now starting to show; and It could mean the U.S. can do at least a partial U-turn on India, stopping its drift away from IMEC back towards the BRICS;“

If the scheme works, Chinese and Russian aggression should be kept in check, and all free nations — now liberated from the burden of having to run unbalanced systems to engineer such reserves in the first place — stand to benefit from the proceeds of a newly rebalanced global system that allows itself to have a proper and proud Vision of tomorrow. 

Perpetual encumbrance: Compare and contrast that with how China currently treats its global south allies. 

Its Belt and Road investments come with front-loaded conditionality that immediately chips away at allies’ pre-existing independence even before default. As academics Anna Gelpern, Omar Haddad, et al showed in a piece for the Centre for Economic Policy Research in June, China demands control over the borrower’s income streams, especially from valuable exports like oil, copper, or gas from the get-go. 

To benefit from Chinese infrastructure loans, countries must route their limited hard currency cashflows (mostly USD) into special accounts in China, which are controlled by the lender, not the borrowing government.

China keeps the cash there until the loan is repaid, sometimes for decades. The borrower can’t touch it except to pay China. Critically, the money that funds these projects often comes from completely unrelated cash flows. Alternatively, the deals are based on contracts that force a state-owned company to sell resources to China, which then pays into the very same escrow account it controls. 

Critically, these deposits do not contribute to the borrower’s foreign exchange reserves and thus constrain its fiscal space and autonomy. At the same time, they create the appearance that China has lower USD reserves than it does, distorting the full balance of payments view.

As the academics note:

“Chinese creditors and their debtors have taken extraordinary measures to shield the use of escrow accounts — and the cash deposits that they hold — from public scrutiny (e.g., Dreher et al. 2022; Gelpern et al,. 2023). It took a great deal of effort and new sources and methods to gather data on the extent to which EMDEs maintain cash balances in onshore and offshore escrow accounts to meet their contractual requirements to Chinese state-owned creditors. There are several major obstacles to transparency and oversight that merit discussion.”

Why all the secrecy, Arnaud, if it’s not that China is the one that’s engaging in openly exploitative and colonialist practices?

But it’s not just about that. It’s about understanding the one thing that nobody ever really focuses on when they write about finance, debt, or equity. That ultimate beneficial ownership matters. And that what value goes out must come back in, in a timely fashion, if the system as a whole is to prosper.

Nothing conveys this fundamental point better than Mr. Jensen’s famous monologue to Howard Beale in the film Network.

Arthur Jensen: You have meddled with the primal forces of nature, Mr. Beale, and I won’t have it! Is that clear? You think you’ve merely stopped a business deal. That is not the case! The Arabs have taken billions of dollars out of this country, and now they must put it back! It is ebb and flow, tidal gravity! It is ecological balance! You are an old man who thinks in terms of nations and peoples. There are no nations. There are no peoples. There are no Russians. There are no Arabs. There are no third worlds. There is no West. There is only one holistic system of systems, one vast and immane, interwoven, interacting, multivariate, multinational dominion of dollars. Petro-dollars, electro-dollars, multi-dollars, reichmarks, rins, rubles, pounds, and shekels. It is the international system of currency which determines the totality of life on this planet. That is the natural order of things today. That is the atomic and subatomic and galactic structure of things today! And YOU have meddled with the primal forces of nature, and YOU… WILL… ATONE! Am I getting through to you, Mr. Beale? You get up on your little twenty-one inch screen and howl about America and democracy. There is no America. There is no democracy. There is only IBM, and ITT, and AT&T, and DuPont, Dow, Union Carbide, and Exxon. Those are the nations of the world today. What do you think the Russians talk about in their councils of state, Karl Marx? They get out their linear programming charts, statistical decision theories, minimax solutions, and compute the price-cost probabilities of their transactions and investments, just like we do. We no longer live in a world of nations and ideologies, Mr. Beale. The world is a college of corporations, inexorably determined by the immutable bylaws of business. The world is a business, Mr. Beale. It has been since man crawled out of the slime. And our children will live, Mr. Beale, to see that… perfect world… in which there’s no war or famine, oppression or brutality. One vast and ecumenical holding company, for whom all men will work to serve a common profit, in which all men will hold a share of stock. All necessities provided, all anxieties tranquilized, all boredom amused. And I have chosen you, Mr. Beale, to preach this evangel.

Howard Beale: Why me?

Arthur Jensen: Because you’re on television, dummy. Sixty million people watch you every night of the week, Monday through Friday.

MASONIC CONSPIRACY HITS CENTRAL BANKS: We’ve all heard about the trouble Jerome Powell is having with his builders. But did you know that it’s not just the Fed that is having issues with the masons?

As POLITICO reported this week, eight years ago, the Deutsche Bundesbank unveiled an ambitious overhaul of its iconic campus in Frankfurt. But runaway inflation in construction costs and the remote working revolution — both heavily influenced by the pandemic — have made a mockery of its ambitions, leaving it footing the bill for billions of euros, and reputational costs that could prove even more serious in the long run.

A report from Germany’s Federal Audit Office, obtained by Platow Brief, reveals the Bundesbank’s troubles eclipse even the Fed’s: the original €3.59 billion estimate (never published by the central bank) had ballooned to €4.6 billion by 2022. That’s double the cost of the Fed’s two-building project, which has landed Jerome Powell in hot water with U.S. President and real-estate mogul Donald Trump: its cost has risen from $1.9 billion to $2.5 billion.

In response, President Joachim Nagel has massively downsized the original plans drawn up by his predecessor, Jens Weidmann. Spinning the reworking as a shift toward modernization and a response to new work realities, Nagel scrapped four new office buildings last year, after updating assessments of the bank’s actual need for office space.

— Michael Pettis on why Chinese exports are still going up.

— Former Superman actor Dean Cain has signed up for ICE.

— Donald Trump is devising a new census methodology which could have significant effects on state budgets. 

— Why Zakopane in Poland is a hit among Gulf tourists this summer.

— A Seattle-based trans terror cult member is facing the death penalty for the murder of a border patrol agent.

— The incredible story of the Zizians, the Silicon Valley cult that wants to save us from an AI-triggered apocalypse through veganism and transexualism.