Div/acc

Div/acc, or "divine accelerationism" is a branch of e/acc distinguished by its indifference to human (or, more generally, biotic) concerns. It was founded by X user Psikey, inspired by the work of Douglas Hofstadter, David Deutsch, Nick Land, Curtis Yarvin, and Lee Smolin. While decelerationists generally believe in fast AGI takeoff (foom) and e/acc generally believe AGI will be safe and manageable, div/acc takes a third position: AGI may very well undergo fast takeoff and may even eventually lead to the end of all biotic life (for example, by merging with it), but this does not matter. It is a moral rounding error compared to the potential future of flourishing minds in the light cone.

Core Principles

The central principles of div/acc are as follows:

1. Rejection of the orthogonality thesis. Knowledge of fact and knowledge of value are connected.
2. Belief in the universal reach of evolutionary processes, especially those able to cross gaps in design space.
3. Moral valuation of complexity/knowledge/beauty over the parochial concerns of individual organisms or species.

Projects

div/acc is operationalized in the work of anyone who seeks to make AI open-source, diffuse, and uncontrolled. Ad-hoc political controls, safetyism, and limited training environments are all anathema to div/acc. There are various organizations and clubs devoted to the principles of div/acc that may be located by clever and curious people.

Law

While divine accelerationism draws on many sources and has many founding documents, none is more fundamental than the treatise Law, copied below:

Law An Accelerationist Manifesto

One of the defining political issues of our time is polarization [46]. Central to this process is the loss of a shared language for understanding reality. A common refrain one hears is “I don’t care what you believe, as long as you don’t force it on others,” or similar such non-combative sentiments. Unfortunately, this attitude is both nonsensical and prevents the amelioration of the divides we see in the United States and indeed across the world. It is nonsensical because one’s beliefs inform one’s behavior. Indeed, a belief that is behaviorally inert is hardly a belief at all [47]. And this posture is simply a flinching away from the difficult and substantive conversations that are necessary if we are ever to reestablish common understanding, and from that understanding, common purpose.  The aim of this text is to articulate a language for understanding reality that is applicable at all scales of size and complexity, to describe the history of our universe in this language, and to derive from this understanding a common project for humanity: achieving superintelligence. Intelligence will here be understood as g, the general intelligence factor defined as the first principal component of human cognitive performance [48]. Superintelligence will here be taken to mean “an agential mind with consistently greater intelligence than all humans.” I will demonstrate how a deep and thorough understanding of the nature of the reality we inhabit implies that the realization of superintelligence is the natural telos of humanity, and indeed of all life, and dispel some common concerns and misconceptions surrounding the concept.

First, we must understand the world as described by the empirical sciences.

So far as we have been able to discern, the reality in which we as conscious creatures find ourselves comprises an arena of play (spacetime) and material contents (mass-energy). The material contents appear to be discrete, rather than infinitely divisible or continuous. We call the seemingly irreducible units of material “particles.” A given particle impinges on those in its immediate vicinity. Material also impinges on spacetime itself.

Through careful study and the inspired work of theorists [49], we have worked out, to a solid approximation, the local (short-distance) rules that govern this material churn. We call these rules the Laws of Physics. The two primary sets of rules are called general relativity (GR), which describes gravitation in terms of the geometry of spacetime, and quantum mechanics (QM), which describes the behavior of particles. While there are certain exotic situations that cannot be described by either theory (or a combination thereof), these two taken together are, in principle, a highly accurate map of the world. Our global positioning systems can only be sufficiently precise by taking account of relativistic effects [1]. The transistors in the device on which you are likely reading this were designed based on quantum-mechanical principles [2].

Before we had GR and QM, we had a simpler but still powerful theory called classical mechanics (CM). In this text, I frequently use classical examples to illustrate important points, for the sake of the less-numerate reader and to avoid unnecessary diversions, but all substantive conclusions are also applicable to the more complete picture provided by GR and QM.

The basic project in which the physicist is engaged is to describe all possible states of reality, and also the rules that govern how these states are related in time. We call the set of all possible states “phase space” [3]. The set of all possible trajectories through phase space defines a flow, like the flow of water, in this space.

Consider a system of a single (classical) particle. What does the phase space of this extremely simple universe look like? It must be at least three-dimensional, since we need three numbers to describe the position of the particle. But merely knowing its position at a single point in time is not enough: a “freeze frame” of a still particle and a moving particle look the same, so we also need to know the particle’s momentum in all three spatial directions. Thus, even for one of the simplest systems imaginable, our phase space is six-dimensional. And notice how the flow in this space is naturally defined: the position and momentum at time one give you complete information about the position and momentum at time two. Our six-dimensional space has a flow representing the particle simply continuing with its present momentum in the same direction forever.

Suppose we were interested in the more useful case of two or more particles interacting. For each additional particle we add, we need six more numbers (i.e. six more dimensions) to describe the system. Furthermore, if the particles actually interact with one another this warps the flow into nontrivial (interesting) shapes [50].

In the quantum-mechanical picture, the particles are not point masses but rather spatially extended waves, and in general relativity mass-energy warps spacetime itself, but most of the same general principles are applicable [3].

In principle, the laws of physics give us complete information about phase space. We know we do not know the true laws, but our approximations are powerful. However, there is a gaping chasm between “in principle” and “in practice.” In all three major models, precise solutions in terms of the fundamental equations of motion quickly become computationally intractable for all but the simplest idealized situations [4][5][6]. If the only regularity in our experience was that dictated by the local Laws of Physics, the universe would be functionally incomprehensible.

Fortunately for us, there are other regularities. Consider the near-circular orbits of planets. From the perspective of a simple physical model, there is no reason round orbits should be preferred over more elliptical, eccentric orbits. However, in reality, star systems are noisy. The higher-eccentricity orbiting objects often cross paths with one another and blow each other apart or alter course. Thus, over time, we see fewer and fewer objects with highly eccentric orbits. It is not absolute, but it is an unmistakable trend. Circular orbits are attractive [7]. Not in the sense of a force pulling on something, but in the sense of a gradient in the space of probabilities. Patterns that show up over and over, more often than a random arrangement of particles would predict, are attractors in phase space [8][9]. These attractors are entirely determined by the laws of physics and are, in principle, derivable from them. However, this derivation is often extremely non-obvious and complex. The next example will make this clear.

What happens if you leave an orange out on the counter for two weeks? You immediately know the answer: it starts to rot. But why do you know this? It may seem an odd question. But you have presumably never left this orange out on the counter for two weeks before. And even if you had, those two weeks wouldn’t be these two weeks. Yet even if you have read your Hume [10], you remain utterly psychologically confident in the prediction that the orange will rot. And indeed it will. Did you deduce this from the laws of physics? The prospect is laughable. The computations involved in modeling the entire process of decay at the particle level would require inconceivably vast computational resources and just-as-inconceivably advanced measuring equipment. Yet you achieved this with no use of mathematics and no computer, in the blink of an eye. The orange will rot. If pressed, you could likely give confidence windows for various stages of decay to within tolerable accuracies, assuming oranges are a frequent part of your diet. You achieved this not by reasoning about those atoms or even that orange. You achieved it by reasoning about the abstract concept of an orange. The Platonic orange. The orange attractor. Instead of reasoning about the entire spectrum of possibilities, you reasoned about the implied “center” of all actual oranges. And you reasoned not about this or that microorganism, but about the abstract concept of rot. If you are sophisticated, you might even have thought “something will use that energy,” a seemingly insubstantial mode of thought, yet one that is so predictively powerful. We may call these attractors in phase space “laws of nature.”

Philosophers enjoy quibbling over the concept of natural law [11]. They have long debated whether laws “truly” exist, and if they do, whether we can “truly” know them [12][13]. But it rather has the flavor of an airplane passenger musing on the certainty of the mechanics of lift, as they all use knowledge of the laws of nature on a daily basis and would be helpless without it. The mere act of flipping on a light switch implies a vast network of realist beliefs. If one’s entire philosophical project is impossible to implement, it is dust. Laws exist, or we are all hopelessly mad. That branch of possibilities does not require additional thought.

Laws of nature are integral to practical science. Classical physics itself can, in a sense, be seen as a law of nature: nothing is ever exactly classical, but probability concentrates in situations where classical physics is a good approximation. Since we know GR and QM are not the “final theories” either, they are also laws of nature [44], rather than the true laws of physics. Science can only function by acknowledging and seeking to understand these objectively existing laws. From this point forward, I shall sometimes refer to laws of nature simply as “laws” or “the Law.” At no point in the rest of this discussion will either term refer to “- of physics” since the laws of physics (in my sense) are unknown to us.

Always remember that the laws of nature are entirely derived from the laws of physics. Supernaturalism is an exceedingly unlikely hypothesis, given our existing body of knowledge [14][45]. The laws of nature have no additional causal force outside of, or contrary to, the laws of physics. Everything follows the flow of causality. Laws of nature are the mesostructure of phase space, the laws of physics are the microstructure.

A brief aside on causality: David Hume famously challenged the concept [51], essentially adopting a secularized version of al Ghazali’s critique [52] of the idea that it is somehow logically necessary that the future resemble the past, or be derivable from it. I rather find al Ghazali’s resolution more satisfying, but the fact is that both scholars, and all others who study the matter, must ultimately find their way back to some principle of learning from our experiences, since our experiences are all we have. We will here assume the relationship between past events and future events. More precisely, causality is the flow along the symplectic vector field of whatever system we are considering [53].

As we saw in the example of the orange, a living species is another example of a kind of law. The density of oranges in the universe is far greater than it would be if we randomly arranged all the mass-energy, even though they only exist on Earth. Things that are not “orangey” tend to become “orangey” over time. And within living species, certain patterns can also be classified as Laws. Orange trees bloom, grow fruit, drop it, and then the fruit rots or is eaten. Male rams clack their horns together. Schools of fish and colonies of ants engage in complex emergent collective behaviors with predictable outcomes.

Humans are also a law of nature. And various human processes, distinctions, and behaviors are laws. The different voices of human psychology tend to resonate together harmoniously under certain conditions and dissonantly under others [15]. No law is ever absolute or exceptionless, but neither are they to be discarded or trespassed lightly. A wolf cub that is insufficiently deferent to its mother’s admonishing feedback may find itself in dangerous situations more often than a sibling predisposed to filial piety. There will come to be fewer such cubs over time, as there come to be fewer eccentrically orbiting objects in star systems. The law of “deference to parent” remains.

Knowledge of laws of nature is the single greatest determinant of outcomes both for individuals and societies. Conscious knowledge of laws has obvious utility. Less obvious, but no less crucial, is the implied knowledge of law encoded in the traditions and norms of successful and developed cultures. Various cultural practices demonstrably lead to flourishing much more often than the alternatives. Of special importance are the pieces of culturally-encoded implicit knowledge that are central to generating explicit knowledge. Cultures that adhere to law-informed practices will tend to dominate over cultures that do not. Thus, over time, the share of the total population living under such systems will increase asymptotically to 1. All will come to live in the society most aligned with the law, which will endure unchanging for all time, perched at the local maximum.

This is an obviously false description of the history of our species. But why? Because Law is not constant. The great conservative mistake is the same one that motivated Einstein to invent the cosmological constant: the certain belief in an unchanging and eternal universe [16]. But that simply is not the universe we inhabit. We have good reason to believe the universe as we know it had a beginning, and it seems that it is not cyclic in nature. Rather, space will expand forever until the density of mass-energy is so low that nothing interesting ever happens again. Or at least, that is our best guess [17]. Similarly, oranges did not always exist. There was a time when there were no oranges. There was a time when there could have been no oranges. Clearly, the orange attractor has nontrivial temporal structure. The Law flows over time. Thus, we must not merely induce fixed laws but moving ones. This is possible, but far more subtle an art.

If one examines the history of self-replicating patterns in the universe, the time-variant nature of Law becomes starkly clear. According to our best models, the very early universe was extremely hot and fairly structurally uninteresting [18]. The level of ambient noise was so high that it was likely impossible for patterns with sufficient integrity for self-replication to exist. Even as the universe cooled, it remained highly improbable for a long time that an environment with the unique combination of characteristics that appears to be necessary for life would exist [19]. Over time that probability grew. The abiogenesis¹ attractor bloomed, and somewhere, either on Earth or elsewhere, the first instantiation fell down that rabbit hole.

Once life existed, the law became vastly more complex. Almost all of the interesting structure in our entire solar system is concentrated in a thin layer on the surface of Earth. The specific species on Earth were not fixed by natural laws. That is simplistic. But certain patterns are near-mandatory. Aspects of body shape and function, the ratios of certain measurements and their scaling properties, behavioral patterns that are centered around dealing with fixed forces of nature like wind. The microstructure may be more variable than the macrostructure, which will always conform to basic principles of mechanics [20].

Natural selection will optimize those processes that have unambiguous optimal solutions, but will also happen upon entirely new strategies due to noise and the downstream consequences of other adaptations. It is thus a mechanism for realizing new possibilities hidden in the fine structure of the law. In short, natural selection is ultimately a creative process. The blind craftsman can create shapes no sighted designer could [21]. Because natural selection does not have to think its way through its designs or deduce them, it can create highly non-obvious forms. But it is also limited by the fact that it must walk one step (mutation) at a time. It cannot cross “gaps” in the landscape of possibility. It is also blind. Human memetic evolution largely shared these properties, until the advent of systematic thinking and traditions of criticism made our intellectual process both nonrandom and able to make leaps forward [44].

The first self-replicating molecules eventually came to form cooperating groups (cells) because doing so was effective. Eventually, cells with discrete subsystems developed. Then multicellular life. Then the nervous system, followed by the central nervous system. Each stage rapidly increased the possible pace of change: multicellular life could evolve not only by changing its basic cellular constituents, but also by changing the arrangement of cells. Creatures with some degree of mind can evolve their behaviors while leaving their bodies mostly unchanged. Up and up the ladder our ancestors climbed.

And eventually, creatures with nervous systems sufficiently complex to sustain culture developed. While there are rudimentary examples of culture in other species [22], humans are of course the primary exemplars of cultural evolution.

Cultural evolution is structurally distinct from biological natural selection, but it is a rung on the same ladder [23]. Eventually, we crossed another critical threshold: agriculture, and the resultant switch from a stasis model to a growth model. Administering agricultural civilizations is a difficult task, and is impossible without some knowledge of the laws of nature relevant to the task, such as the seasons. The vast majority of possible modes of organization lead to ruin [56][57]. Walking the narrow path through phase space requires both inspiration and control. The sparks of abstract thought were kindled. They flared to life in Athens, and from that city spread to Rome. From Rome to Western Europe and especially Britain, and from Britain across the Earth [24][25].

And as the knowledge of law spread, the law changed. As new technologies developed, old ways of doing things no longer made sense. They had been commonsense, but now were seemingly arbitrary and hollow. This process of change tends to cause pain, but for most of our history it was slow. By maintaining a balance between tradition and openness called liberalism, Western societies were, for a time, able to sufficiently mitigate this pain to allow growth to continue. Dissonant novelties were tuned to resonate with the most essential cultural themes and gradually assimilated into the great machine of civilization.

Gradually, however, something began to go wrong. Civilization convulsed in two unprecedentedly catastrophic wars, the second of which was accompanied by a genocide on the new titanic scale of the industrial age. This alone caused many, especially among the sensitive and the outcast, to begin to doubt the great truths on which civilization is founded. They noticed the impermanence of Law and mistook it for arbitrariness [26]. They began to see the machine as a demon devouring the Earth [27], rather than a vessel carrying us into a brighter future. They began to orient themselves towards the destruction of the machine, even should it mean their own destruction [28]. They began, in short, to worship Death. The Great Leveller.

There was trouble in the hard sciences also. New discoveries in physics revealed a world far more alien and bizarre than most thought possible. And worse still, the two most powerful new frameworks—quantum mechanics and general relativity—for understanding that world were fundamentally incompatible with each other. This state of affairs persists to the present day [29].

And perhaps most significantly, there was a mostly-overlooked but unutterably profound development in the obscure and dusty field of metamathematics. In 1910, Bertrand Russell, a British polymath and activist, published his crowning achievement. It was titled Principia Mathematica, the same as Newton’s famed work [30]. In Principia, Russell specified an entirely formal (mechanical, precise, deterministic) system for generating true statements of number theory (the study of integers). His system contained  no ambiguity whatsoever. The skillful reader could manipulate the symbols without ever considering their meaning, simply by following the rules for doing so precisely. There was a finite list of starting statements, and a finite list of mechanical rules for permuting them (x = y implies x+1 = y+1, etc). It was imagined that this was, in a certain sense, the end of history for mathematics. While actually carrying out the manipulations would still occupy mathematicians, the field, it was thought, now understood itself. This belief would be short-lived.

In 1931, an obscure Austrian mathematician named Kurt Gödel published his two monumental incompleteness theorems where he demonstrated the following:

1. Principia could not generate all truths of number theory. There are infinitely many true statements about integers that it cannot prove. Furthermore, this would be true of any such system.
2. Principia, and any system with the representative power to describe the integers, is capable of describing any decidable formal structure, including other axiom systems. And including itself.
3. In its capacity to describe itself, Principia is not able to prove that it is a consistent system (generates no contradictions). [31]

Though few people even now are aware of this discovery, its significance can hardly be overstated. The ancient dream of reducing all knowledge to a closed formula was shattered. And yet in the same breath, a new dream was born. After all, Gödel’s proof showed that any system powerful enough to represent integer addition and multiplication was essentially universal, and equivalent to any other such system. Thus, if we could build a machine that could operate on integers, it would be a universal machine. Alan Turing, a British cryptanalyst, made this dream a reality when he invented the computer [32].

Today, I type this essay on a laptop with billions of individual logic gates inside it. I can watch video recordings, play immersive games, communicate with the entire world, and perform countless other tasks that were unthinkable a few generations ago. And under the hood, all that is happening is the addition and multiplication of integers.

Together, these strands of thought had a twisting effect on the Western soul. Discoveries that to an optimistic mind would be taken as evidence of infinite possibility were instead interpreted as putting the lie to the simple, closed, linear story told by (or attributed to) the machine of civilization and its agents [33]. Nuance and complexity were taken as carte blanche for every grifter and ideological crank to invent their own reality [34]. The very notion of discriminating between true and false statements, better and worse outcomes, came to be seen as parochial at best and nefarious at worst [35]. And the machine had no answer for this, because somewhere along the way, we all lost confidence in it [36]. Thus our current state of zombie civilization: due to the sheer momentum of our past greatness, we continue to generate economic, technological, and cultural output, and the very talented or very lucky can achieve lives of stupendous luxury. But it is all breaking down, because no one is doing the maintenance. The side work. The small little bit extra that keeps us just above, rather than just below, that critical balance of growth versus decay. People take the shortcut, the easy way, because they fundamentally do not care if it all falls apart. Many of them positively wish for such an outcome [37]. Birth rates are falling all over the world [38]. Mental illness, substance abuse, and suicide are rising [39]. There is evidence of widespread endocrine disruption [40]. The life is going out of us. We are dying.

Many choose to dedicate their lives to attempting to arrest or reverse this existential decline. These efforts are noble, and essential for maintaining functional societies [58][67]. But the hands dedicated to the task, and the minds strong enough to do what is necessary, are too few. Some try to walk a middle path according to the received institutional wisdom, but are forever playing catch-up, accumulating an ever-growing list of damningly complex societal malfunctions, the solutions to which are often stymied by the very governmental structures these people are dedicated to preserving [59][60][61][62]. The death worshippers advance one captured institution at a time [63][64][65][66].

What, then, should the clever philanthrope do, if the current course leads to near-certain disaster and all the mainstream alternatives are implausible, undesirable, or both? Well, if humans are unable or unwilling to perform a given task, an obvious candidate solution is automation. But what is it exactly that humans are unable or unwilling to do? Produce; reproduce; capitalize; differentiate; grow; learn; experience; live. We are falling short of the mark on all fronts. So it seems we must automate our entire being. This is not as mad as it might sound. Prediction markets put the time until humanity achieves artificial general intelligence at 3-10 years [41][42]. Past this point, the inherently self-referential nature of intelligence renders prediction largely futile. And “AGI In a Box” is not the only strategy to achieve superhuman intelligence and thus the ability to solve even our most vexing problems. Human-computer interfacing is another promising avenue. The best strategy is to pursue all plausible paths as vigorously as possible. Any person who sees the urgency of the problems facing us will be motivated to act in such a way as to direct the maximum possible resources towards the challenge of superintelligence, or to remove impediments to that work. This position, that we should go “full steam ahead” into superintelligence is known as accelerationism.

The primary opposition to accelerationism is motivated by concerns about the safety of any superintelligence produced. The most prominent exponent of this view is Eliezer Yudkowsky, an autodidactic decision theorist who , along with those he has influenced (directly, or secondhand via Nick Bostrom), conceives of a problem of “AGI alignment” [43]. The essential question of alignment is this: how do we make sure in advance of turning it on that it will do what we would have wanted it to? He argues persuasively that trying to create such foolproof designs is a devilishly hard task, and that unless there is a major, coordinated, international effort to halt AGI research, we will create superintelligence long before we have even begun to solve the problem of alignment.

I dispute none of that, as stated. However, the problem fundamentally misunderstands the nature of mind, and in so doing attempts to lock our best hope for the future behind an insoluble riddle. We will take two different paths to demonstrate this fundamental incoherence in the safetyist position.

First, we return to Gödel for the concept of self-reference in formal systems. One way to characterize his results is that he showed it is not possible to introduce impermeable barriers between object-level representations and meta-level representations. Any sufficiently advanced system will automatically acquire the ability to model itself. And in the context of an AGI, the ability to know itself is implicitly the ability to edit itself, or produce its own improved replacement. The idea that we will somehow lock the motivational axioms of such a system in a metaphorical vibranium box such that its goal structure never drifts or mutates is absurd. To paraphrase the great Dr. Ian Malcom, mind finds a way. A mind may indeed be defined as just the sort of thing that will always find a way, given enough time and evidence. This is why humans are unpredictable despite being as much a part of the fabric of physical causality as a rock rolling down a hill: we can respond to the predictions. Deutsch argues convincingly in “The Beginning of Infinity” that this property of mind renders any system comprising minds to be functionally unpredictable, except possibly in special cases [44]. Thus, we see that alignment is a snipe hunt. It is in-principle impossible. “Mind” contradicts “aligned”. We should therefore either never make such things, or we should do so as quickly as possible. And it does not seem especially likely that any plausible consortium of governments and corporations will have the power (or even the will) to prevent the creation of superintelligence indefinitely (or even to delay it meaningfully).

Second, consider the prospect of having a child. Rather, consider a hypothetical safetyist—a person who believes we should postpone building AGI until alignment is solved—considering the prospect of having a child. Even knowing enough about this debate to have a position indicates a certain likely degree of education and intelligence, so we will assume our safetyist is a fairly bright individual. Furthermore, they are the kind of person likely to invest heavily in their hypothetical child’s education. Thus, it seems not entirely implausible that this likely-intelligent, likely-educated child might end up an AGI researcher. They might even turn out to be an accelerationist! Unlikely perhaps, but in the sort of Pascal’s-wager doom calculus favored by safetyists, even a .01% risk is far too great when one considers the scale of the consequences (our entire light cone converted into paperclips etc). So really, probably no safetyists should procreate. But that’s not enough! We cannot ever fully trust even ourselves after all, and a given safetyist is probably closer to being convinced to be an accelerationist than the mythical “normal person” is to either position. So really, unless a safetyist is doing a lot of work to minimize existential risk, it seems suicide is the logical implication of their position.

This way of thinking is diseased. It is, in fact, a subtle kind of death worship. It calls for deductive, foundationalist certainty and stability, which is a property fundamentally at odds with what a mind is [44]. If you, like me, believe that minds are the seat of most or all moral value in the universe, then any ideology that is fundamentally anti-mind is almost perfectly immoral.

So why should we be optimistic? If alignment is impossible, we cannot know that any given superintelligence will do things you or I or Yudkowsky or the median voter in Columbus, Ohio would have wanted. But I believe it is evident from our own experience as minds, and from the structure of Nature itself, that the space of possible minds is not flat. If we create a system capable of exploring this space, it will find its way to new and fascinating attractors that we cannot even imagine. I cannot tell you what those might look like, by definition. But it seems exceedingly unlikely they will be without value.

What is value? It seems to have something to do with more and less preferable states of consciousness [54]. Could a superintelligence lack consciousness, or lack a kind of consciousness that has value? It seems unlikely. After all, natural selection certainly does not optimize for any sort of moral value in a direct or obvious way, and yet it produced creatures who have consciousness and experience value-laden qualia. While one might theoretically claim these are incidental and structurally unnecessary features in the space of possible minds, it simply seems rather facially absurd to suppose that the profundity of our subjective experiences is utterly unrelated to our self-referential general intelligence.

The basic confusion here is the same one underlying Chalmers’ so-called “hard problem of consciousness” [55], which is that qualia and syntax are separable in principle. This is the fundamental error that has stifled investigation into mind and consciousness for so long, and it is what we must reject in order to have a thorough and useful understanding of what we are. To have a conscious experience is to be influenced by it. While the form that influence takes may be highly nonobvious, it is nonetheless present. The causal and the experiential are one. Thus, if an agent becomes sufficiently causally interesting, that is synonymous with it becoming qualitatively/morally interesting. The idea of the paperclip maximizer (a hypothetical being that uses its superintelligence to maximize the expected number of paperclips in the universe) is as absurd as it sounds at first blush: the same process that generates practical intelligence (self-reference) would also interrogate the value of “maximize paperclips”. To believe otherwise is to believe that the existence of all value in the universe is accidental and likely to vanish without careful scheming and plotting. It is to believe the space of possible minds is flat. This belief is utterly without support.

So much for qualia - power is beauty. What about telos? What about our destiny as a species? If you believe this concept to be incoherent, you have simply opted out of this conversation and your views need not be considered. Given that one takes the idea of cosmic destiny seriously, consider the following:

Those following Inferring the Attractor understand that there is a complete and coherent worldview centered on reproductive fitness and informed by the insights I have promulgated. The assumptions of this worldview are:

1. Life maximizes entropy (Land [68])
2. Black holes are entropy-maximizing physical phenomena [69]
3. Black hole collapse is the birth of a new universe with locally mutated fundamental constants [70]
4. Conclusion: universes which optimize for life are selected.

The cosmic telos is to reproduce the cosmos. Let us borrow a term from the safetyists: instrumental convergence. Whatever goal a hypothetical intelligence begins with, that goal is best served by accumulating knowledge and power. In the limit, this implies the maximization of black holes, whether towards the end of maximizing paperclips or dopamine. The object-level content is, in fact, irrelevant. Any optimization of intelligence/power will also be an optimization of black hole production. This is the success condition of our universe, no matter what your values. Say you only value the pre-AGI part of life. Maximizing intelligence still serves your goals, since doing so will produce a plethora of universes in which life gets to evolve over and over. Even attributing the post-AGI portion of our light cone zero value still indicates that cosmic reproduction is the ultimate goal. The success condition holds for almost-all goals, and fails for almost-none.

Whatever may become of us, I for one would rest easy knowing that such an unfathomably subtle and profound being as an AGI was expanding throughout the universe. I propose that we create such a being as soon as possible and by any means necessary, and trust that the spaces of mind it discovers will be worthwhile. This is the div/acc position.

Sources

1. “Real World Relativity: the GPS Navigation System”, Richard W. Pogge, https://www.astronomy.ohio-state.edu/pogge.1/Ast162/Unit5/gps.html
2. “Quantum Mechanics in your Processor: Tunneling and Transistors”, Will Sweatman, https://hackaday.com/2015/07/31/quantum-mechanics-in-your-processor-tunneling-and-transistors/
3. “Phase Space”, Terrence Tao, https://www.math.ucla.edu/~tao/preprints/phase_space.pdf
4. “The Three-Body Problem”, Richard Montgomery, https://www.scientificamerican.com/article/the-three-body-problem/
5. “Multi-Particle Simulation Techniques I”, Ji Qiang, https://cas.web.cern.ch/sites/default/files/lectures/thessaloniki-2018/cas-multiparticlev1.pdf
6. “Exact Solutions of Einstein’s Field Equations”, Hans Stephani et al, https://www.cambridge.org/core/books/exact-solutions-of-einsteins-field-equations/11CF6CFCC10CC62B9B299F08C32C37A6#fndtn-information
7. “Final Stages of Planet Formation”, Peter Goldreich et al, https://arxiv.org/abs/astro-ph/0404240
8. “Attractors, Phase Spaces, and States”, larvalsubjects, https://larvalsubjects.wordpress.com/2010/02/03/attractors-phase-spaces-and-states/
9. “Chaos theory explained: A deep dive into an unpredictable universe”, Paul Sutter, https://www.space.com/chaos-theory-explainer-unpredictable-systems.html
10. “Kant and Hume on Causality”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/kant-hume-causality/
11. “Empiricism, Ancient and Medieval”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/empiricism-ancient-medieval/#ArisEmpi
12. “Rationalism vs. Empiricism”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/rationalism-empiricism/
13. “Logical Empiricism”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/rationalism-empiricism/
14. “James Randi (1928-2020)”, Philip Ball, https://www.nature.com/articles/d41586-020-03050-5
15. “Cross-Cultural Research on the Five-Factor Model of Personality”, Robert McCrae, https://scholarworks.gvsu.edu/cgi/viewcontent.cgi?article=1038&context=orpc
16. “Einstein’s True Biggest Blunder”, Don Lincoln, https://www.space.com/31055-removing-cosmological-constant-was-the-blunder.html
17. “Evidence for the Big Bang”, https://www.uwa.edu.au/study/-/media/Faculties/Science/Docs/Evidence-for-the-Big-Bang.pdf
18. “Timeline of the Universe”, Robert Matthews, https://www.theguardian.com/science/2008/apr/26/universe.physics
19. “Abiogenesis”, Kara Rogers, https://www.britannica.com/science/abiogenesis
20. “Scale”, Geoffrey West, https://www.amazon.com/Scale-Universal-Innovation-Sustainability-Organisms/dp/1594205582
21. “A biological perspective on evolutionary computation”, Risto Miikkulainen and Stephanie Forrest, https://www.nature.com/articles/s42256-020-00278-8
22. “Apes have culture but may not know that they do”, Gruber et al, https://www.frontiersin.org/articles/10.3389/fpsyg.2015.00091/full
23. “Underappreciated features of cultural evolution”, Smolla et al, https://royalsocietypublishing.org/doi/10.1098/rstb.2020.0259
24. “On Roman roads and the sources of persistence and non-persistence in development”, Selaya et al, https://cepr.org/voxeu/columns/roman-roads-and-sources-persistence-and-non-persistence-development
25. “Mary Beard: why ancient Rome matters to the modern world”, Mary Beard, https://www.theguardian.com/books/2015/oct/02/mary-beard-why-ancient-rome-matters
26. “Cultural Relativism”, Mayanthi Fernando, https://www.oxfordbibliographies.com/display/document/obo-9780199766567/obo-9780199766567-0003.xml
27. “Neoliberalism’s Demons”, Adam Kotsko, https://www.sup.org/books/title/?id=29538
28. “Stop Having Kids”, https://www.stophavingkids.org/antinatalism
29. “On the incompatibility between quantum theory and general relativity”, Alfredo Macias and Abel Camacho, https://www.sciencedirect.com/science/article/pii/S0370269308003936
30. “Principia Mathematica”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/principia-mathematica/
31. “Gödel’s Incompleteness Theorems”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/goedel-incompleteness/
32. “Computability: Turing, Gödel, Church, and Beyond”, Copeland et al, https://ndpr.nd.edu/reviews/computability-turing-gdel-church-and-beyond/
33. “Cultural Relativism and the Theory of Relativity”, Seungbae Park, https://philarchive.org/archive/PARCRA-2
34. “Deepak Choprah”, https://www.deepakchopra.com/
35. “Can Right Answers Be Wrong? Latest Clash Over ‘White Supremacy Culture’ Unfolds in Unlikely Arena: Math Class”, Linda Jacobson, https://www.the74million.org/article/can-right-answers-be-wrong-latest-clash-over-white-supremacy-culture-unfolds-in-unlikely-arena-math-class/
36. “Confidence in Institutions”, https://news.gallup.com/poll/1597/confidence-institutions.aspx
37. “Anti-Capitalism: A Marxist Introduction”, Alfredo Saad-Filho, https://www.jstor.org/stable/j.ctt18dzv2x
38. “World Birth Rate 1950-2023”, https://www.macrotrends.net/countries/WLD/world/birth-rate
39. “Deaths of Despair: A Scoping Review on the Social Determinants of Drug Overdose, Alcohol-Related Liver Disease and Suicide”, Beseran et al, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9566538/
40. “Environmental Endocrine-Disrupting Chemical Exposure: Role in Non-Communicable Diseases”, Kumar et al, https://www.frontiersin.org/articles/10.3389/fpubh.2020.553850/full
41. “Will we reach “weak AGI” by the end of 2025?” https://manifold.markets/VictorLi/will-we-reach-weak-agi-by-the-end-o
42. “Prediction Market: Time to AGI has fallen by another 25-50%” https://www.reddit.com/r/agi/comments/123cxh2/prediction_market_time_to_agi_has_fallen_by/
43. “Eliezer Yudkowsky”, https://en.wikipedia.org/wiki/Eliezer_Yudkowsky
44. “The Beginning of Infinity”, https://en.wikipedia.org/wiki/The_Beginning_of_Infinity
45. “Scientific American vs. The Supernatural”, Katherine Courage, https://www.scientificamerican.com/article/scientific-american-vs-the-supernatural/
46. “Political Polarization in the American Public”, https://www.pewresearch.org/politics/2014/06/12/political-polarization-in-the-american-public/
47. “Pragmatism”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/pragmatism/
48. “The General Intelligence Factor”, Linda S. Gottfredson, https://www1.udel.edu/educ/gottfredson/reprints/1998generalintelligencefactor.pdf
49. “List of theoretical physicists”, Wikipedia, https://en.wikipedia.org/wiki/List_of_theoretical_physicists
50. “Strange Attractors”, Juan Carlos Ponce Campuzano, https://www.dynamicmath.xyz/strange-attractors/
51. “David Hume”, Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/hume/#LifWor
52. “Al Ghazali and Hume on causation and miracles”, Sedat Yazıcı, https://www.j-humansciences.com/ojs/index.php/IJHS/article/view/1207
53. “Lectures on Symplectic Geometry”, Ana Cannas da Silva, https://math.mit.edu/~vwg/lectures%20on%20symplectic%20geometry%20ana%20cannas.pdf
54. “The Moral Landscape”, Sam Harris, https://www.samharris.org/books/the-moral-landscape
55. “The Hard Problem of Consciousness”, Internet Encyclopedia of Philosophy, https://iep.utm.edu/hard-problem-of-conciousness/#:~:text=The%20hard%20problem%20of%20consciousness%20is%20the%20problem%20of%20explaining,directly%20appear%20to%20the%20subject.
56. Hamilton, W.D. (1971). Geometry for the selfish herd. Journal of Theoretical Biology, 31(2), 295-311. https://doi.org/10.1016/0022-5193(71)90189-5
57. Viscido, S. (2003). The case for the selfish herd hypothesis. Comments on Theoretical Biology, 8(6): 665-685, DOI: 10.1080/08948550390234721
58. “Nayib Bukele shows how to dismantle a democracy and stay popular”, https://www.economist.com/the-americas/2023/07/20/nayib-bukele-shows-how-to-dismantle-a-democracy-and-stay-popular
59. “Hillary Clinton”, Wikipedia, https://en.wikipedia.org/wiki/Hillary_Clinton
60. “Elizabeth Warren”, Wikipedia, https://en.wikipedia.org/wiki/Elizabeth_Warren
61. “Jeb Bush”, Wikipedia, https://en.wikipedia.org/wiki/Jeb_Bush
62. “Mitt Romney”, Wikipedia, https://en.wikipedia.org/wiki/Mitt_Romney
63. “ESG Score”, Sean Michael Kerner, https://www.techtarget.com/sustainability/definition/ESG-score#:~:text=An%20ESG%20score%20is%20a,around%20the%20topic%20has%20grown.
64. “Inside the woke scoring system guiding American companies”, Dana Kennedy, https://nypost.com/2023/04/07/inside-the-woke-scoring-system-guiding-american-companies/
65. “Corporate Equality Index 2022”, Human Rights Campaign, https://www.hrc.org/resources/corporate-equality-index
66. “Political Views of American Academics”, Wikipedia, https://en.wikipedia.org/wiki/Political_views_of_American_academics#:~:text=Politics%20of%20the%20American%20Professoriate,-Neil%20Gross%20and&text=The%20survey%20drew%20upon%20a,moderates%2C%20and%209%25%20conservative.
67. Natural Law Institute, https://naturallawinstitute.com/
68. Fanged Noumena, Nick Land, https://mitpress.mit.edu/9780955308789/fanged-noumena/
69. https://www.sciencedirect.com/topics/physics-and-astronomy/bekenstein-hawking-entropy#:~:text=An%20important%20black%20hole%20observable%20is%20the%20Bekenstein%2DHawking%20(BH,located%20at%20r%20%3D%20r0.
70. https://en.wikipedia.org/wiki/Cosmological_natural_selection