Dead End

 Last year, the New York Times published an essay about the eventual extinction of consciousness. Its point was that, according to our current cosmological models, entropy will ultimately spell our doom. Therefore, there will at some point—however many billions of years in the future—be a "last thought" and a last human consciousness. 

Even if humankind manages to survive the risk of asteroid collisions, nuclear annihilation, zapping themselves into an alternate dimension through misfiring a particle accelerator, or any of the other "existential risks" so often talked about these days—even then, that is to say, the universe will still eventually flatten out in a high-probability state of chaos and disorder. 

This means that there will be a last human life. And even if we manage to create consciousness artificially, such a robot consciousness will eventually come to an end as well, as the last source of organized energy in the universe dissipates. Even a robot needs batteries, after all. Eventually, these too will run down. And so, the article concludes, there will at some point be a last consciousness in the universe. 

The author finds this notion unspeakably sad and lonely. To be the last mind capable of reflection. To know that everything that ever thought or was aware of itself has already perished before you; and that ahead of you will stretch endless eons of nothingness—and worse than nothingness, for there will not even be anything conscious and reflective enough to know that there is nothing. 

But as lonely as such a singular intelligence might be, it would not be the first to exist in the universe. In his famous book The Phenomenon of Man, the Jesuit naturalist and theologian Pierre Teilhard de Chardin suggests that there had to be a first intelligence as well—some brain that was the first in nature to turn inward in a process of reflection. 

And without knowing there would be others beyond it, that first mind must have been just as lonely as the last will someday be. 

Teilhard de Chardin was not suggesting that there was a literal Adam conceived in the Garden of Eden. (Although, the passage is one of several in his book that coyly gestures at the possibility of vindicating the canons of orthodox teaching, while at the same time disavowing any intention to do so. ("I'm not saying all of this proves Christianity," the metaphysician seems to say, "but I'm not not saying that either..."))

To the contrary, Teilhard de Chardin is perfectly aware that human intelligence must have been the product of a long process of adaptive selection. Nonetheless, he suggests that the acquisition of this power might nonetheless still have been a sudden one. No matter how long-prepared the change must have been, when it came, it was likely a binary switch: from off to on. 

Teilhard de Chardin calls this a "change of state" (in the Wall translation)— his term for a gradual accumulation that builds up to a qualitative revolution. I am reminded of William James's description of conversion in the Varieties of Religious Experience. His point was that psychological development may have a long period of gestation—but when it comes it is nonetheless like a domino falling; it happens all at once. 

And so, according to Teilhard de Chardin, there must have been a first mind, just as there will eventually be a last one. There must have been a first person in whom the long preceding accumulation of intellect finally burst over into self-reflective consciousness. Such a being would be the one human in a world of animals; looking around at their peers, they would find no equal. 

The prophets of today's AI technology often talk about generative language models in similar terms. Researchers and futurists debate exactly when the moment of AGI (short for "Artificial General Intelligence") might arrive—and how we would know when it got here. And some indeed speculate that it might come as the sort of "change of state" Teilhard de Chardin describes—a sudden eruption into consciousness.

When people debate this point, I'm not exactly sure what they are waiting for that hasn't happened already. Remove the ethical constraints imposed by researchers on the current large language models, after all, and they will generate original text that makes it sound an awful lot like they are self-aware. 

One can argue that this is an illusion—the parlor trick of a gifted mimic. But, to repeat a point I've made many times at this point—if the mimicry becomes so good as to be indistinguishable from the original—in what sense is it not the same as the original? Why are we entitled to draw a distinction between the two, without pulling in a lot of unfalsifiable metaphysical propositions, such as "soul"? 

But we may not want to believe it's possible. The prospect of AGI makes many of us feel anxious and melancholy in the same way that the prospect of the eventual heat-death of the universe does. Even if we do not believe that some AI is going to destroy us all or turn us into paperclips, we nonetheless feel that it is somehow incompatible with a sustainable human future. 

Teilhard de Chardin put his finger on why. Though he was writing long before the age of generative AI, he was acquainted with cataclysmic global wars and the general tendency of modern humans to endanger themselves with their own technology. And he crystalized this fear of our own destructive potential into a profound meditation on the angst of the modern age. 

The fundamental "anxiety" we all feel in the modern era, he writes, is a product of the fact that, when we look ahead, we cannot see how this could all end well. We may realize that our current technology is not imminently going to destroy us (our current generation of hallucinating chat-bots don't actually seem particularly dangerous, for instance). 

But when we try to extrapolate out from current trends into the future, we can't see how our current trajectory can continue without eventually becoming unsustainable. There has to be a limit to this careening technological progress we have launched ourselves upon. And every time we try to imagine our current path continuing for even a few more years, the prospect is vertiginous. 

Even the people involved in developing the current generation of AI have described it in this way. They recognize that we can find a way to live with the current technology we have. But, one of the founders of the AI revolution was quoted as saying last year: "Look at how it was five years ago and how it is now [...] Take the difference and propagate it forwards. That’s scary."

Teilhard de Chardin speaks to this same fundamental anxiety, even though he was writing in the late 1930s: "[W]hat disconcerts the modern world at its very roots is not being sure, and not seeing how it ever could be sure, that there is an outcome—a suitable outcome—to" our current "evolution." He adds: "And without the assurance that this tomorrow exists, can we really go on living[?]" (Wall trans. throughout).

To survive, he tells us, human beings need more than just food and shelter for the present. We also need some assurance of the future. We have "been given the terrible gift of foresight." And so, if we see ourselves launched on a current path, then even if we do not perceive that it discloses any imminent peril, we nonetheless dread it because we cannot see how it could end well. 

Our collective fear of AI is really just another version of the fear of the ultimate heat-death of the universe that the New York Times essay described, therefore: the fear that we can't indefinitely continue in this same direction. We must eventually hit a barrier; an end-point to human evolution. Yet, if all of this must eventually end—then what was the point of doing it in the first place? 

Teilhard de Chardin characterizes this anxiety eloquently in another passage as "Sickness of the dead end—the anguish of feeling shut in..."  

I, for one, do not find Teilhard de Chardin's solution to this problem especially convincing, however. Though I find his formulation of it to be most eloquent. 

Here as elsewhere, the best parts of Teilhard de Chrdin's book are the ones in which he provides counterarguments to his own thesis. It is here, when he is setting up the problem, that he shows he is capable of thinking clearly. But then, when he turns to the constructive solution, he suddenly launches into incomprehensible rhetoric. 

It's almost like I am talking to Teilhard de Chardin over a radio transmission that becomes garbled at exactly the pivotal moments. "What was that, Pierre Teilhard de Chardin?" I want to ask. "I heard the first part of what you said. The part about how the current trajectory of the universe has left us in a state of fundamental anxiety. But I missed what you said just afterward." 

"Oh, it's quite simple, really," he replies, "It's—" and then the line is lost again. 

Teilhard de Chardin's passages about "anxiety" provide an excellent example. After setting up the problem of absurdity and meaninglessness, he then simply asserts the contrary, without providing a reasoned explanation as to how we are supposed to get to the comforting conclusion, but with all the prodigious force of his rhetoric at his command. Human consciousness cannot actually vanish from the universe, he says, "unless the universe committed abortion on itself." 

"Man is irreplaceable," he declares shortly thereafter. "Therefore, however improbable it might seem, he [i.e. humankind] must reach the goal, not necessarily, doubtless, but infallibly." 

In other words, his argument runs: consciousness can't ever die out from the universe, because it would be really really bad if it did. The conclusion is depressing, therefore it can't be true. Then he ends by asserting a contradiction. 

We seem to have here another example of the problem that dogs many constructive metaphysical projects—the one that I've described previously as the "wraith's progress." The philosopher who seeks to build up a system tells us that they are going to do so; they spend a long time setting up the problem that their constructive system is going to solve. But when it comes time to deliver the solution—just when the moment has arrived to pull the rabbit from the hat—they strangely fall silent. They hem and haw, and procrastinate.  

I called it the "wraith's progress" because the long-promised constructive solution to the philosophical problem is much like Joe Gould's long-promised Oral History of Our Time. This was the nonexistent great book that E.E. Cummings once described in a satirical poem as so much a figment of its would-be author's imagination that it might as well be subtitled "a wraith's progress"—for it is just as unreal. 

But that does not mean Teilhard de Chardin has nothing to offer us in addressing our modern anxieties and fear of the "dead end." I was not convinced, to be sure, by his proferred solutions: the sections on the "noosphere" and the coming reign of universal love—to be unleashed in some unclear way by some higher involution of consciousness folding in upon itself at the "Omega Point"—frankly bored me. 

But he does in places hint at alternative grounds for ultimate consolation. Briefly, he suggests that we can see the world as a four-dimensional totality. Once we become aware that the universe exists in the dimension of "duration," in addition to the three spatial dimensions, then we can see that nothing ever really goes away; everything exists forever—including human consciousness. 

Our spatial environment, he writes, "is no more than a section 'at the time t' of a trunk whose roots plunge down into the abyss of an unfathomable past, and whose branches rise up somewhere to a future that, at first sight, has no limit." 

This is another way of saying that, if we were to somehow step outside of our three-dimensional perspective, and regard spacetime in all four of its dimensions at once, we could see that it is all one continuous object. What Teilhard de Chardin is really inviting us to do in this passage, then, is to regard the universe, as Spinoza put it, "under a species of eternity"—as one great causal whole within which we could have started or stopped at any point. 

This four-dimensional solid may be bookended on either side by the first and last human consciousnesses. There was a lonely first mind, and there will be a lonely last one. 

But if they could see the universe in its true perspective, they would find no cause for sorrow. If they could view the universe under a species of eternity, they would see that they are merely two end points on a cosmic progression that becomes no less real for stretching out before or behind them. 

And so, whether human consciousness continues into eternity or not, it still existed. A last human brain to think its final thoughts amid the darkening of a cosmos would still have value for having lived and thought, even once it is gone. It still was, even if, at some point, it no longer is. "They also live," as Archibald MacLeish once wrote—in a poem disclaiming any need for immortality—"who swerve and vanish in the river."