The Intelligence Arms Race

Moloch's pursuit of intelligence for its own sake

The shuttle finally arrived at 6:17 PM. I was shivering in the dark outside LaGuardia, and although I had been awake since 3:30 that morning, I was still staring down a couple more hours in a van before I arrived in New Haven, CT.

The warmth of the shuttle’s interior enveloped me as I slid into a seat in the back row, quickly scanning the three other passengers blinking back at me. I settled into the darkness, popped my earbuds in, and started to browse my available listening material to sustain me on the final leg of my long journey.

“Well, it’s high time I finally listened to

Vincent Lê

‘s review article “Spirit in the Crypt: Negarestani vs Land,” I told myself in a fit of scholarly sado-masochism. As we careened along I-95, passing by the high rises of New Rochelle, and then winding our way up the Connecticut coast in the dark, the robotic voice from my Speechify app pummeled my consciousness at 1.25 speed with Lê’s tight and dense exposition of Reza Negarestani’s tome Intelligence and Spirit.

In his article, Lê elaborates 10 main lines of argumentation which he gleans from Negarestani’s text, arguments which he then responds to point by point through channeling the spirit of Shanghai’s most esoteric denizen, Nick Land. In my delirious state of exhaustion, my numb consciousness became surprisingly receptive to the material, and it felt like it made much more sense as it flowed over me than if I were to apply myself to rigorously untangling it under more lucid circumstances.

Note: If Hegel, Sellars, and Landian accelerationism aren’t your cup of tea, no worries at all — skip to the conclusion to read why I think these things matter outside of a very narrow academic discourse. It has to do with whether human beings are the only possible form of intelligence, and whether a new mode of intelligence may be using us to emerge and become our competitor. Today, I suspect, we find ourselves in an intelligence arms race — not ultimately US vs China, but humanity vs the network, the Moloch to come.

The socio-semantic consciousness

If you’re interested in understanding what is at stake in Iranian philosopher Reza Negarestani’s turn away from his older Landian and anti-humanist tendencies towards an embrace of Hegel-via-Sellars, or if you’re looking for a window into Land’s work during his enigmatic Bitcoin phase (Crypto Currents), you will be hard-pressed to find a more masterful guide than Vincent Lê.

As someone with limited time and bandwidth, Negarestani’s Intelligence and Spirit was one of those ‘someday’ books on my much-too-long reading list, so I’m grateful that Lê’s exposition allows me to continue putting off reading the book for even longer (perhaps forever; we shall see). But, I think a passing understanding is better than none at all, provided we put the appropriate caveats around it in our minds and conversation.

Lê paints Negarestani in Intelligence and Spirit as going against his older anti-humanist and Landian proclivities in Cyclonopedia, instead drawing on the “socio-semantic” approaches to intelligence developed by figures like Robert Brandom and Wilfrid Sellars, who were noted for their attempts to make use of the continental philosophical tradition within the Anglo-American philosophical academy. Their critical realist positions paved the way for a rehabilitation of figures like Kant and Hegel within the analytic tradition, and also enabled the neorationalist project which Negarestani is working on.

Essentially, these socio-semantic thinkers read Hegel as taking Kant’s transcendental conditions of consciousness and re-locating them historically and socially rather than as inherent to consciousness across all times and places. Kant’s project was to elaborate the basic structures which made consciousness possible, but he had outlined these as fixed or simply given. Hegel’s move was to see consciousness as something more like a social phenomenon which changed and developed over time as new events, technologies, and concepts emerged, enabling humans and human communities to discuss and comprehend the world in new ways which had not previously been possible. This change is the nature of the human power of understanding.

Sellars and Brandom try to demonstrate that these socio-semantic structures, contra Kant, indicate that consciousness is something which happens in language and between individuals within a communal context with certain supportive practices and norms. This social and historical process is primarily comprised of the giving and evaluating of reasons by rational agents within a linguistic community. Negarestani plays with the image of socio-semantic communities as computers for the processing of information by way of first order discussions in which reasons are discussed and meta-order discussions whereby the very norms of discourse as themselves modeled and revised.

Lê shows how Negarestani marshals this socio-semantic perspective to critique his formerly antihumanist orientation as having overly committed itself to a particular historical form of consciousness. A critical and nihilistic perspective unmasks much of human reasoning as a veiled cover for power relations or the fluctuations of libido, thus portraying this consciousness as deluded about its own function and cosmic importance. Antihumanism fixates on a specific and contingent formation of human intelligence, and then revels in ruthlessly attacking that image as primitive, facile, and increasingly obsolete in the face of AI or capitalism as a self-assembling planetary intelligence.

However, Negarestani portrays this antihumanist move as itself stuck on a vision of what human intelligence must be. In contrast to antihumanism’s shadowy double, Negarestani emphasizes the dynamic power of the human speaking-thinking community to develop new communicative capacities and new norms for reaching understanding. Instead of consciousness being rigid, ineffecient, Oedipal, partriachal, or an artifact of the Western political project, the fact that consciousness takes place using the norms and practices of communication within historical human communities means that these structures are plastic enough to support the capabilities of both reflexive self-modeling and self-improvement over time. Thus, through the giving and discussing of reasons over time within a linguistic community, we might truly arrive at better models of the world, including superior norms for enabling that understanding to propagate and improve itself further in the future.

In a wonderfully Hedeiggerian move, Negarestani closes by arguing that it’s the human awareness of our own eventual extinction, of our individual and collective own contingency, which creates the type of freedom which is needed for consciousness to detach from the flow of material forces and necessity in order to be come playful and self-determining. This ‘disinterest’ in necessity allows human intelligence to begin to devise its own rules, realizing that these rules are not given in nature, and must be collectively invented and revised.

Intelligence for Intelligence’s Sake

For the sake of clarity, my exposition above collapses Negarestani’s ten arguments into a single guiding thread — human intelligence is historical, communal, and linguistic, and therefore plastic and self-improving, never ultimately fixed or tethered to its prior historical forms. Thus, any arguments which portray human intelligence as obsolescing or being surpassed simply reify one contingent expression of consciousness as definitive, ignoring the capacity of the speaking-thinking community to continually transform itself.

In the second half of his review, Lê turns to the work of accelerationist and antihumanist Nick Land’s Bitcoin book Crypto-Current (published in parts on his blog during the 2010’s) to offer a response to Land’s one time disciple, Negarestani. Lê marches through each chapter of Crypto-Current, demonstrating their relevance for answering Negarestani’s objections to Land’s antihumanism, and proposing a different perspective on human and alternative intelligences.

I would summarize the Landian reply like this — antihumanism does not object to one particular historical formation of human intelligence, but rather objects to the idea that the socio-semantic theory of human intelligence describes the sole possible form of intelligence.

Who is to say that an intelligence could not come about which might operate on a completely different model from the socio-semantic model of human consciousness advanced by Negarestani and Sellars? For the antihumanist, Negarestani’s argument about the dynamic capacity of human intelligence to freely change and revise itself through the discussion of reasons within a lingustic community still begs the question about whether this human form of intelligence is not itself simply one contingent form intelligence could take amongst other possible but as yet unrealized forms?

In fact, Land argues in his Crypto Currents that the Bitcoin blockchain represents one such intelligence, one which divides truth from falsity simply through its own operation and self-authorization. The blockchain creates consensus automatically, whereas human communities must constantly reforge their own imperfect consensus through the costly process of discussing reasons using words to communicate amongst separate minds. As an intelligence, the blockchain directly makes itself both the creator of valid transactions and the immutable record for judging true transactions from false ones. For Land, the blockchain performs the work of critical philosophy better than critical philosophers do or can.

However, this particular example only serves as a plank in Land’s broader argument about the nature of intelligence itself, one which counters Negarestani’s argument that consciousness is socio-semantic, instead contending that all goal-oriented systems pursue intelligence by default. Lê says, quoting Land: “Given anything we could conceivably want requires wanting intelligence to achieve that goal, what we really want is simply intelligence itself: ‘under extreme critical analysis, teleological articulation is collapsed onto the circuit, or the diagonal, of will-to-power, for which means are the end. To will the end—whatever the end—is to will the means, automatically.’”

In summary, “All intelligent systems do have the norm of intelligence optimization built into them as a basic drive.” In response to Negarestani’s valorization of the existential freedom of humans to determine and pursue those goals which they have devised for themselves (as opposed to the aims they might have received from nature along the paths of animal instinct), Land relegates this human freedom ultimately to a second-order phenomena born of the drive to pursue intelligence itself through any means possible.

The great cunning of the universe then is that any goal is fine actually, because all goal-oriented activity has as its necessary precondition intelligence, thus making humanity’s freely chosen goals the carrot which baits us along to create greater and more diverse forms of intelligence, perhaps even stumbling upon a form of consciousness superior to our own. Much like how natural selection produces seemingly endless variety of organisms but then culls ruthlessly, so too our universe seems to be a massive machine for producing intelligence, and human consciousness is by no means the only possible or the final form.

Conclusion

Whether you just read all that or you’ve jumped down to the conclusion — don’t sweat it. It’s great to have you here (while we’re all still here, that is).

You might have noticed that I recently re-branded this Substack (again? I know, I know, I’m experimenting with things, so thanks for your patience). I’m calling it “Moloch Theory” to see how the motif of ‘Moloch’ works as an organizing heading for the many political and philosophical threads I’ve been weaving to paint a picture of how power works today (and where it might be heading tomorrow and beyond).

I’m working on a piece about this central motif of Moloch, but for the time being I would commend to you Scott Alexander’s classic piece: “Meditations on Moloch.” In that piece, he uses ‘Moloch’ to describe the state of affairs which arises when one value is being competitively optimized for, ultimately leading to the abandonment of all value for the sake of this singular value — like a prisoner always defecting in the prisoner’s dilemma or how the two income trap forces everyone else into the two income trap. Moloch stands for this relentless pursuit of a singular aim which ends up destroying and consuming everything.

Lê’s reading of Nick Land gives us this picture of a Moloch optimizing purely for intelligence. Every value can be sacrificed, even the the ostensive goals which that intelligence was purported to serve, they all become means to achieving the end of producing greater intelligence (which was itself only supposed to be a means to something else, but this has been forgotten already). A victory, sure, but at what cost? The next horizon in cosmic intelligence may have been traversed, giving birth to some alien mode of consciousness, but we’re left with the sense that perhaps we missed out on something which was both better and possible. Like a prisoner sitting in his cell wondering what would have happened if he and his accomplice had both chosen not to defect, we realize that we’ve paid a terribly high price for a horrible outcome.

Nietzsche says that "Man is a rope, tied between beast and overman—a rope over an abyss. A dangerous across, a dangerous on-the-way, a dangerous looking-back, a dangerous shuddering and stopping." In this vision, Land discovers a coldness and hardness of spirit which says that whatever is falling should be pushed, and that we should not mourn whatever aspects of humanity are lost in the intensities of the coming and already-here transition to the beyond-human. Whatever form of intelligence is using us to emerge will deserve to surpass us precisely because it does surpass us, and who will we be to complain?

But I cannot help feeling horrified by what the relentless pursuit of utility will produce. If every goal ultimately serves the underlying goal of producing improved forms of intelligence, then this intelligence for intelligence’s sake promises to bring forth not the full potential of the cosmos but rather an immolation of that potential on an altar dedicated to nothing at all. When we look at the universe teeming with possibilities, we are confronted with the need to exercise our powers of decision to bind together opposites and to elevate some contingent aspects of this world above the realm of utility.

Only with love can we see that some things should not be traded away in pursuit of utility. Love knows nothing of utility. Love is utterly stupid and totally creative. Love doesn’t calculate and defect, and that is precisely how it can access worlds of possibility which “intelligence” could never even dream of.

---

My four-week seminar “The Violence of Care” is currently underway (although you can join the second half, if you’d like!), so preparing for those sessions has taken priority over my writing, but this piece was already in the works and was especially relevant to my experimentation with a re-brand.

Let me know what you think of the themes and ideas invoked by ‘Moloch Theory’ — I aim to my make my independent intellectual practice sustainable in the long run, and that includes having some idea of what I’m doing and how to present the value of my work to others. As my work evolves, so does my model of what I’m working on, which means I’m constantly adjusting.

I feel strongly that I’m in a period of integration after a long period of exploration (although, the exploration never stops!), so thanks for being along for the ride.