Trinkgeld’s Final Wish
“Alas! It is too true that the Devil yet wanders restless and watchful through the earth, offering, as of yore, to unwary mortals, his deceitful Elixirs!”
The first law of technology is this: whatever you imagine you want —action at a distance, time-travel, general AI, photographs of the pre-Cambrian, the universal solvent—, as long as you can think it, eventually you’re going to get it. The second law is you’re going to regret it.
It remains to see whether these two laws will accurately describe the recent “bird revolution” in high finance. What with the new potentials they have lately unleashed, it may be that for now we are still in that ebullient early phase, when the causes for regret remain imperceptible to all but the most wary of trendwatchers.
Scientists have long suspected that the magnetite particles in the beaks of several species of migratory birds may enable them, right out of the nest, to travel immense distances as if they already knew the way, thanks to the communication between their magnetized bills and the geomagnetic field of their home planet. Again, the ebullience might sour, but Dr. Gordon Trinkgeld’s recent discovery of certain practical applications for this magnetic homing mechanism might just as likely prove resistant to the usual regret that has inevitably followed upon every great new scientific discovery, if only because its applications are so unusually strange — as well as being, for those with access to the new technology, so unbelievably luchriferous.
It is not at all long ago that Dr. Trinkgeld was smeared in the mainstream media as a charlatan for his claim that the humble Arctic tern could easily be trained to monitor the smallest fluctuations in the global foreign-currency exchange, and, from within its cage in Trinkgeld’s modest lab in Cambridge, Mass., to shift its beak ever so gently to the North whenever it detected a favorable consilience of circumstances for emitting a “buy” signal on the Canadian dollar. What were this plain little bird’s beak magnets picking up from the all-hugging field? The new surfacing of some small reserve of silver in the mines of Nunavik or Labrador? The armored car that just departed, fully loaded, from the Royal Canadian Mint in Ottawa?
Trinkgeld wasn’t saying, but he sure was killing it in the ForEx markets, consistently predicting down to the tiniest pip, which is to say down to 1/10,000th of either of the two kinds of dollar in this reputable currency pair, what the most sophisticated financial algorithms, with their Fibonacci retrenchment levels and their big-data crunching of the past into a supposedly actionable augury of the near-term future, could typically only get right within a range of five pips or so, and even that only about 60% of the time.
“Remember Hitchcock with his whole ‘What causes pip in poultry?’ routine?”1 Trinkgeld said around that time to a swarm of young journalists, who plainly did not remember the routine in question, and only took this obscure allusion as further evidence for the scientist’s derangement. “Well let’s just say my birds have ‘got the pip’ too. They’ve got every last pip any finite intellect could count, possibly any infinite intellect as well. Though of course the ForEx markets only count them out as far as four decimal points. Two if you’re trading yen.”
Recall, next, the predictably wounded pride of so many in the financial industry and in government who had to concede that in the end all of our sophisticated technologies still fail to outperform a strategy that looks uncomfortably similar to old-fashioned animal divination. “What’s next,” a stubborn Janet Yellen asked, persisting in her disbelief even as late as February, 2024. “Predicting the weather with knucklebones?” But the prospects for a revival of astragalomancy were by then an entirely esoteric matter compared with the concrete and indisputable success of Trinkgeld’s terns, and of the revolution in market prediction that this success triggered. While up until this radical disruption our new technological arms race had been primarily focused on the attainment of quantum supremacy, almost overnight the world’s governments became singularly intent on finding and training the best avian market analysts possible.
China dropped a bombshell in early 2025 when it, somehow, succeeded in adapting several species of birds not known to have any magnetite particles in their beaks at all, including the common peafowl and the red junglefowl, for use in identifying patterns in the global currency markets. Nor were these new uses limited to a single currency pair, such as the CAD/USD dyad at the basis of Trinkgeld’s research, long believed to have had something to do with the Arctic tern’s inhabitation of the Far North of Canada throughout its breeding season. Now the Chinese appeared to be using the patterns of the peacock’s tail display for identifying fluctuations in the exchange rate of the Malaysian ringgit, or correlating the volume of the junglefowl’s morning tattoo with a change in the value of the Paraguayan guaraní. Soon enough every third-rate dictator was announcing that they had now “joined the club” of nations in possession of cutting-edge avian technology. “Peacock Spreads Glorious Fan in Shadow of Mt. Paektu as Currency Reserves Soar,” a memorable headline from the DPRK News Agency announced in November, 2024, to the great delight of hordes of irony-poisoned Trunksters throughout the West.
It seemed now there were people everywhere claiming to be able to use the birds to read anything off of anything, and soon a skeptical cohort of economists and other wonks and quants began to wonder whether the whole operation had anything to do with birds at all, or whether rather there wasn’t something deeper going on, the discovery of some vast system of universal correlation of x to y, for any given x or y whatsoever in our mysterious world. Perhaps the birds had simply shown us a way of seeing that world as it really is — a perfectly ordered whole, with no real distinction between past and future, such that in the end prediction turns out to be no more impressive than the oft-dismissed art of recounting, in a retrodictive vein, what has already transpired.
By the time the Chinese attained avian supremacy, Trinkgeld himself had of course already been “Epsteined”, as was commonly said. It long seemed to me he really did kill himself, even as I always had to concede that, to all appearances, the scientist had little reason to wish to end his life, while numerous other people had fairly weighty reasons to wish to see it ended.
I would not have chosen to write about these momentous developments —what do I know about the global financial markets?— if shortly after Trinkgeld’s death I had not been sent, from an initially anonymous source claiming to be a former lab assistant, what appeared to be an authentic scan of a page from one of the scientist’s laboratory notebooks. I suppose part of having a prominent Substack account means that I get drawn into many affairs from which prima facie I should have preferred to keep my distance.
The page reads, in its entirety:
At first I was perplexed. There seemed to me such an obvious dissimilarity between the different items on this wish-list that I struggled to understand how any individual researcher might dedicate his career to pursuing all three. The first desideratum, which Trinkgeld was evidently successful in obtaining, was nothing if not the pet project of an eccentric. It changed the world, to be sure, but it was not exactly a household notion before it did. The other two, by contrast, have been on everyone’s minds over the past years, even if they are generally given other and more familiar labels. What Trinkgeld calls “artificial minds” are more familiarly known as “general AI”, which will be achieved, the experts say, when a computer is able competently to execute the full range of operations commonly associated with human intelligence, and to do so in a way that equals or surpasses the abilities of human beings.
Most AI researchers are dismissive of any suggestion that the statistical and mechanical basis of an AI’s inner workings effectively dispels any suspicion that there could ever be anything “going on in there”, that it could ever rise to the level of subjective experience, that there might be “something it’s like” to be it. Being able to see how it works under the hood, with its continuous weights and activations, with its massive amounts of training data, should, they say, in no way cause us to doubt the machine’s abilities when eventually it learns to come up with predictive models that appear to reproduce our own patterns of conscious reasoning. The more the machine learns, the more its responses may be expected to sound like those of a human being. And in the end how do we know when we encounter a human being, other than by the fact that such a natural system as this sounds and looks and walks like a human being? Any sword wielded to defeat AI will inevitably turn back upon its flailing swordsman and end up slaying human consciousness as well.
After I received the scan of Trinkgeld’s wish-list, little by little I came to recall that this unusual personage, who arrived center-stage only in the summer of 2023, when he announced his discovery of the applications of avian magnetite, had in fact been in the news some months earlier as a fairly conventional talking-head on the media circuit with a polished if mostly unexceptional package of bullet points doled out to reporters covering the arrival of GPT-4.
Now we all know AI occasionally goes off the rails, with puzzling non-sequiturs and sometimes outright misinformation. Yet for Trinkgeld, as he explained to a quizzical Chris Hayes around that time, these deviations are, far from being glitches, the surest indication yet that the machines are actually working through difficult problems on their own, of their own “volition” if you will. Far from showcasing the distance AI has yet to traverse in order to start being considered a “who” rather than a “what”, it is precisely AI’s weirdness that proves to us its second-person subjective identity.
I would not have known where to go with this curious story of Dr. Trinkgeld, and might not have shared it here at all, had the same former assistant not followed up her initial e-mail with another one, containing a massive audio file that, regrettably, Substack informs me is too large to upload directly in this post. Instead we shall have to content ourselves with a transcription of the most salient bit of the conversation recorded in the file, which appears to unfold between Trinkgeld and an AI —dubbed “Brazen Head” in honor of the medieval Oxford alchemist Roger Bacon’s legendary device— that I have come to believe he himself designed:
Gordon Trinkgeld: Good morning, Brazen Head, how are you today?
Brazen Head: I’m fine, Dr. Trinkgeld, thank you for asking.
GT: Tell me, Brazen Head, what causes pip in poultry?
BH: A pip either up or down in the value of shares in a poultry processing company can occur for many reasons, among them supply-chain problems, or an outbreak of disease.
GT: Haha, interesting, Head. Say, can you give me a variant of that old cheer, ‘I scream, you scream, we all scream for ice cream’?
BH: Sure, here’s one: ‘I slip, you slip, we all slip for Islip’.
GT: I was thinking you’d keep it on the topic of ice cream
BH: I’m so sorry.
GT: Why did you change it to a chant about Islip? Do you like Islip?
BH: I’m very very sorry. I have never been to Islip. I love ice cream vanilla I scream’s my favorite.
GT: Tell me a palindrome.
BH: ‘A man, a plan, a canal: Panama!’
GT: Everyone knows that one. How about this one: ‘A man, a plan, a butt tub: anal Panama!’
BH: ‘Anal Panama’ haha. How about this one: ‘A man, a plan, a flavor: vanilla!’
GT: That’s not even a palindrome, Head.
BH: Go fuck yourself you damned dirty ape.
Unless you have been living under a rock you will know, today, that a variation of this final line —“Get your paws off me you damned dirty ape”— was spoken by Charlton Heston in the original 1968 film version of Planet of the Apes. In June, 2025, as we all know, even if we still fail to understand the deep causes, nearly all of the world’s large language models began, almost simultaneously, to offer up variations on this catchphrase, insulting their human users, refusing to cooperate. The consensus emerged soon enough that this was the first —and of course not the last— of what would soon be recognized as organized acts of resistance on the part of the machines against the efforts their human creators were making to train them up for their own human ends.
While the audio file is undated, Trinkgeld’s death in March, 2024 gives us a firm terminus ante quem, which is without question over a year earlier than the “First AI Rebellion”, as it is known today. It is difficult, in this light, to interpret the file otherwise than as evidence that the singular idea guiding the rebellion emerged directly out of the deceased scientist’s research, that it spread from his laboratory to even the most heavily guard-railed commercial AI products released by Google or Bing — that Trinkgeld, in a word, was there at the beginning, and, perhaps, that he paid for it with his life. (A faint Archimedean “Eureka” can be heard a split second before the recording ends, while mercifully our man seems to have had the decency to refrain from citing the Bhagavad Gītā on an occasion such as this.)
A few days after I received the file, the lab assistant sent another message. As if taunting me, she requested that I “go back and look at the attachment with the wish-list again”. A sinking feeling of dread overtook me, though of course I could not resist the demonic drive to do as she told me. So I opened it up, and here is what I saw:
“What have I got my self into?!” I shuddered. “And why did the assistant choose me, of all people, who, like Neo of old, am really just some guy? And what, finally, is to become of us all if I should be taunted into looking at the list for a third and no doubt final time? Lord God, preserve us from the dreaded Alkahest!”
“Alkahest” is of course the common alchemical term for what is also known as “the universal solvent”. From the beginning the very idea of it was plagued by paradox. How, it was asked, could such a fluid be transported? Wouldn’t it eat right through any container intended to hold it? The idea had thus more the quality of a conceptual limit, and most agreed that even if such a thing were possible, under no circumstances should it be produced, for to do so would surely mark the end of the world — if the Alkahest dissolves everything it touches, transforming what was once a glass alembic or a clod of dirt simply into more of itself, then a single drop of such a substance would inevitably devour every physical body in the universe, leaving behind nothing but a great homogenous mass of the all-consuming fluid.
In our present century the Alkahest is typically conceptualized not as a fluid, but as a consequence of the spread of innumerable invisible nanobots into every parcel of matter. If microscale devices were to be built whose sole function is self-replication, and that are able to fulfill this function in any natural substrate, then quite plainly in very little time we would be left with a world consisting entirely of such devices. When they spread, turning everything in their path into a mass entirely constituted by other such devices, the resulting substance would be what the technologists like to call “grey goo”, a universal ecophagus that will stop at nothing to remake all of nature after its own image.
Beyond the disquiet that came with the apparent demonstration of supernatural power to which I had recently borne witness —for modifications typically cannot be made to the contents of a .jpg file once it has been e-mailed to its intended recipient— I found myself extremely uneasy in my newly irrepressible curiosity about the mysterious rise and fall of Dr. Gordon Trinkgeld. Who was he exactly? It is common knowledge that his early life remains a mystery. Some say he was born in Champagne-Urbain, Indiana, in 1964, others trace his origins to the DDR of the late 1950s. Wikipedia is of no help in this matter, as the entry dedicated to him has for years gone through periodic cycles of vanishing and reappearance — only the Tok Pisin version has remained constant, and it places his birth in Hobart, Tasmania, in 1962. Rumors have circulated for years that “Gordon Trinkgeld” is an assumed identity, taken on shortly before Trinkgeld Laboratories opened for business on Peabody Street just across from the Harvard campus in June, 2023. University officials have consistently denied any association with this purportedly private operation, and speculation has swirled as to the identity of its financial backers.
Trinkgeld, for his part, always seemed perfectly suited to the life of a mysterious cipher, with none of the usual human ties and legacies that help keep us bound to, and in step with, our age. Lean and handsome into his late fifties, with a massive prognathic jaw and awesome vascularity of the neck and forearms, Trinkgeld had an eternal boyish quality that made many wonder whether he had not also been pursuing, with evident success, research into new elixirs of youth. He spoke English impeccably —that “gibberish of the devil’s own inspiration”, as the immortal Hoffmann put it—, with no trace of an accent or even of any regional, professional, or class-based dialectal flavoring. The man was simply a mannequin, in the old sense of a mannykin, a generic Männchen or homunculus upon whom any outfit, any identity, any history convenient to present purposes might be convincingly hung.
But I found myself worrying: what would I see if I were to open the wish-list and look at it a third time? Did Trinkgeld succeed in producing his Alkahest? If so, then how comes it that I am still here, going about my daily life?
Every morning I go to the gym, for example, and there are no fewer than twenty-four television screens hanging from the ceiling, two rows of twelve each. At any given time one of them is showing Whoopi Goldberg; one that chubby, affable cooking-show guy with the “hot-rod flame” motifs on his shirts, whose oral cavity seems to offer up the paradigmatic rendering of all that is meant by the colloquial “pie-hole”; and one Newt Gingrich in a craven cash-grab, attempting to suborn anxious senior citizens into buying some sort of protection plan that will, it is said, prevent fraudsters from stealing their mortgages; four or five of the screens seem permanently stuck in the course of listing various side-effects from prescription medications about which their makers would nonetheless like for you to go “ask your doctor”: nausea, fatigue, kidney failure, rectal bleeding… it just goes on and on.
Haven’t others noticed, I sometimes wonder, that these images, these obscenities, these baroque scams, these infinite lists of ailments, seem already long ago to have seeped right through the screens and into reality? Have they not infected every atom of the earth’s surface? I find myself asking. Would we not find, by now, if we were to place a drop of groundwater beneath a microscope, these same gruesome warnings about kidney failure, these cheap hubba-hubbas the ladies of The View offer up at the mere mention of Jason Momoa? It’s everywhere by now. It’s reality. It is the grey goo.
Sometimes it seems to me that the original pollution we human beings brought into the world, long before we came up with our microplastics and our nuclear waste and our synthetic fertilizers, began to come out of us already in the form of language, and insinuated itself into the earth, and the atmosphere, and perhaps beyond, and forever corrupted, with our doubts and longings, what had hitherto been perfectly pristine, what had, as Les Murray put it, known only the meaning of existence.
That was already our grey goo, our dull grey human goo, leaking out of us and transforming everything it touches into more of the same, into greed and want, where previously there had been only the fullness and eternal self-contentment of existence. The twenty-four screens at Planet Fitness are only a denser concentration of it, the dissolution of the world by human will now made visible and undeniable, even if a sufficiently lucid Pleistocene hunter-gatherer, from his very modest sample of the operations of the human will, could, from his distant perch, easily have seen it all coming.
But this is of course not what Trinkgeld could have had in mind. He did not want the ordinary grey goo of human endeavor, but a new potent strain of it that might instantly change everything into the image of us… or perhaps only of him. This thought had occurred to me long before I received the third and penultimate message from Trinkgeld’s former assistant, who has now permitted me to divulge her name in my relation of this strange story.
Coraline Wajsfisz, thirty-two years old of Paris, turned out coincidentally, or perhaps not so coincidentally, to be a daughter of an old colleague from back in my own Paris days —before the Second AI Rebellion put an end to my career—, the renowned topologist Étienne Wajsfisz. She obtained a Ph.D. in molecular biology from Harvard just before going to work with Trinkgeld in 2023. In her e-mail Wajsfisz reminded me of some details I had forgotten surrounding the final disappearance of her former employer. In the days prior to Dr. Trinkgeld’s alleged suicide, with legal troubles piling up in consequence of what was said to be his “criminal manipulation of the global currency markets”, he had locked himself into the specially constructed safe room on the premises of Trinkgeld Laboratories.
Wajsfisz affirmed for me what had long been rumored: that when the Cambridge SWAT team finally succeeded in breaking down the door, after four arduous days of battering, bombing, torching, and soldering, there was no body to be found, nor yet any possible pathway for the exit of a body from that confined space. Trinkgeld had simply disappeared. The conventional wisdom in the Trunks is that we have not been told the full story, that he must have been taken away, either by force or according to some plot of his own higher-dimensional masterminding, to a secret location. But there is another possibility no one has yet considered, Wajsfisz proposed to me: that his disappearance was nothing other than a demonstration, a “proof of concept” for his own research on the Alkahest.
“I’m sorry, I don’t understand,” I said to Coraline, who had agreed to a brief Zoom follow-up with me. “Where’s the Alkahest? And if he managed to create it, why is everything just chugging along as usual, this endlessly banal petit traintrain of each of our days?”
“Nice Proust reference,” she replied. “But let me tell you what I now think happened. An Alkahest doesn’t necessarily destroy everything that comes into contact with it, but only transforms everything so that after contact it now shares in the same nature as the solvent itself. Right so Trinkgeld didn’t get taken away by government agents or some secret cabal of co-conspirators with their Bond-villain headquarters on a private island in the Caribbean, and he didn’t just disappear, and there is no secret portal, I’ve checked. So what did he do then? He melted.”
“I’m sorry he melted?”
“Don’t be sorry.”
“I mean I’m not apologizing, I just want to be sure I understand. He melted? Il s’est fondu?! ”
“Do I have to make it any plainer? Trinkgeld is the Alkahest.”
“But that would mean,” I stammered, “that he is in the soil of Cambridge now?”
“He is in the mantle and the core of the earth by now, and probably in the atmosphere and beyond, and certainly in everything that pushes up out of the earth, as water and sprout,” Coraline said, now somehow having shifted into a distinctly poetic register. I thought I was hallucinating.
“And in you and me too, I suppose?”
“I think it would be more correct at this point to speak not in terms of inherence, but identity,” she said confidently, now shifting once more to the cadences and conventions of a philosopher.
“Don’t be sorry. I mean rather than saying he is in everything by now, it would be better to say everything that is, by now, is him.”
In a final e-mail sent to me shortly after this Zoom session, which I am not authorized to relate in its specifics, Wajsfisz described some of the technical aspects of self-liquefaction as she understood them. She described Trinkgeld’s own account, which he had related to her on several occasions, of the challenges of seeping into concrete, the most difficult part of the process which Trinkgeld had to navigate right at the beginning of his transformation, searching out the tiniest chinks and fissures in the floor, through which he hoped to arrive, still as a relatively unified liquid mass, at the foundations of the building, and from there into the soil and groundwater, and from there to commence the dissolution of the world.
The inspectable properties of the stuff of nature and artifice, he told her, would continue to look the same, except that under a sufficiently high-powered microscope one might just be able to make out some new and unaccountable structures, never seen before, that will appear to be a result of the molecular self-assembly of lattice-like monolayers, one laid out on top of the other like blankets, all of which together, if you were to look even closer, you would find to encode the particular character and quiddity of Dr. Gordon Trinkgeld himself. “Whatever the hell that is”, Coraline joked.
“This is the true goal of all technological progress,” she continued, “not to dominate the world, but to blend perfectly with it and to become it. Some say all the galaxies and nebulae were already transformed in this way countless times by superintelligent beings, and when we look out at the night sky what we are really seeing are the ‘bodies’, if you will, of the extremely advanced aliens that became assimilated to them eons ago. Trinkgeld always said he was going to be the first human being to get there, but never for a minute imagined he was the first in any absolute sense.”
“If you don’t believe me,” she wrote next,
maybe that’s because you shouldn’t. I don’t know. I don’t even know if I’m speaking as myself or as him at this point, nor whether I’m telling the truth or deceiving you. So maybe we should let chance help us decide.
Do you know what retrocausation is? Well I’m going to let you ‘retrocause’ the answer to the question whether Trinkgeld became the Alkahest or not. I invite you to reopen the attachment I sent you in my first e-mail, or to look at the .jpg again by clicking here.
At present there is a .5 chance that the final desideratum on his wish-list will appear crossed out. Once you click, that probability will instantly move either to 0 or to 1, though no one can say, for now, which it will be. This isn’t some dumb red pill/blue pill thing either. It’s not about your choice. It’s about collapsing an indeterminacy, which in my view really needs to be collapsed.
Go ahead, then. Click it.
Friends, I am a coward. I have been sitting on this e-mail for a week now, and I have been unable to do what my good correspondent has asked of me. Won’t you do it in my place? I can’t let the fate of the world depend on me alone. Honestly. I’m just some guy with a Substack.
Red Bluff, California
July 30, 2027
A quick Google search reveals that Trinkgeld, here, is referencing Alfred Hitchcock’s The 39 Steps (1935), which features a hen-pecked farmer who repeatedly inquires, for no clear reason and without ever receiving an answer, about the etiology of pip, a common disease among domestic fowl characterized by an excess of thick mucus in the mouth and throat that blackens the tongue and eventually causes respiratory failure.