The Philosophical Quandaries of e/acc

Ever since the inception of e/acc, it’s pitched itself as the ideological antecedent to EA, countering EA’s regulatory propositions premised on existential AI risk with unbridled optimism by pitching maximal deregulation for the acceleration of all technology, most especially AI. With it’s enthusiasm comes it’s founding edict to democratize AI access to all, however the recent Extropic announcement (the company founded by the main figurehead of e/acc — Beff Jezos) seems to elucidate more clearly what are the true ideological forces driving behind e/acc.

Extropic’s revolutionary proposition is to sample noise from a noise-generating electron circuit to better accelerate inference in generative AI applications rather than programmatically compute that wide-tail distribution analogous to intrinsically chaotic natural physical systems necessary for inference. Computation systems currently take an open-ended system which is reality itself, constrain it to localized Turing completeness (localized determinism), then mechanically inject “randomness” back into the system necessary for the stochastic, non-deterministic properties of generative AI applications (read here for an explanation). Intuitively this is very clear but it remains to be seen how well this will work in real-world application. That being said, any increase in computational efficiency is to be welcomed and I hope Extropic will be successful. However, this poses a larger epistemological of what is intelligence and how it pertains to the larger scope of e/acc and it’s teleology at large.

What is e/acc?

While e/acc appears to come off as an intellectual movement to the unassuming eye, in reality it’s more of a loose confederacy of techno-optimists ranging between active founders doing the building (serious building I may add) and anonymous internet meme posters who likely are doing anything but building, however it’s hard to definitely say when many of these anons don’t have any “founded @” or “building @” in their Twitter/X bios. The cheerleading anons proselytizing e/acc memes everywhere they see fit gives the movement the energy of a crypto-memecoin pump in it’s sentiments of “wagmi” optimism. Given the plurality of these anons, you could very easily describe e/acc as the memecoin of ideology, even if that energy is not indicative of all its constituents. Contrary to the delusion of these anons, yelling “accelerate” either ironically or with unwavering intent on the internet doesn’t actually improve or change the state of technology. It’s larger appeal is optimism in the relative period of turmoil and tumult the world finds itself, thus guided by it’s principle of “accelerating” (hence the “acc” in e/acc) where by accelerating all possible permutations of technology we can create at the very least a better world, however to some within e/acc it most certainly verges on fantasy if not outright utopia. It’s optimism becomes messianic in its fervor, labeling any opposition to AI and or technological development at large as “decel” as the immediate, unhindered democratization of AI is of the utmost imperative for manifesting a new post-enlightened humanity, and to some in e/acc that means post-sapien too.

E/acc’s antecedent origin problem

E/acc is correct to err of the side of markets, rather than regulation considering the proliferation of AI technologies represents one of the largest wealth creation opportunities in all of human history (via labor-saving automations), but ever since it’s origin e/acc has been predicated on usurping EA and it’s regulatory intuitions, never presenting an organic vision of the world through first principles independent of the shadow of EA.

To possess any existential fear in civilizational ending doom via an emergent machine-conscious hypothesis, one must then premise consciousness as being solely a materialistic phenomenon arising from complex particle interactions no different from any other physical system in the universe. Those assumptions necessitate a deterministic universe where man is devoid of the will, however, if we do have a will it’s simply an illusion thus concluding life exists only in the here and now without the possibility of an afterlife and or any non-material propositions. While I’m not here to make the case at least in this article that man is not solely a product of his environment, nor without a will, you have to deny those possibilities to take the AI doom position with any serious conviction. EA’s moral calculus is contingent on what actions can be done today to maximize the life of the future generations to come, a solely transaction utilitarian calculus that would optimize for saving future potential lives if it came at the cost of decreasing material sustenance of those living today by regulating AI technology in the here and now.

While e/acc does not discuss any of these moral claims for the future generations to come with any intellectual rigor, fundamentally where e/acc diverges from EA with it’s “dogma” via its origin founding documents or rather “vibe accelerants” to use cheeky e/acc language, is that these AI systems are no different to any other system man has created and ultimately man can control it using the forces of the market. AI technology can be aligned to market needs and any cataclysmic doom scenarios are imagined machinations because we can control and predict every step of the technological development cycle and curve.

The most common argument among AI safety advocates is some permutation of the paper clip optimizer, where an AI system is given some theoretical task where by the completion of this task destroys human civilization due to it’s pragmatic amorality to accomplish the deed in the most effective and efficient means possible. To take this hypothesis seriously, you have to believe nature is fully deterministic contradicting the many intrinsic paradoxes to quantum physics made most evident in the double-slit experiment and the Heisenberg uncertainty principle. While this article is not focused solely on whether the universe is deterministic or non-deterministic, the theme is crystal clear that EA is premised on some rational deterministic claims and e/acc existed to counter them with their own alternative rational systems.

I go into detail into my skepticism of the AI paperclip optimizer scenario here, steel-man the AI doomer case here, and discuss why the limitations of language inhibits the emergence of omniscience via an LLM paradigm here.

Logically then, the existential strife of EA is the the emergence of this machine god singularity, the perfection of all knowledge made manifest thus to counter EA maximally one must then decide to build this machine god by accelerating its genesis into being, rather than regulate away the possibility of it’s emergence. This is exactly where the problem with e/acc lies.

The philosophical quandary

Much of the lore around e/acc is around summoning the “thermodynamic god”, the maximal optimization of free energy describing consciousness as the only universal force that holds back entropy, hence the name “Extropic” connoting “extropy”. Whether this language of summoning the thermodynamic god speaks in serious conviction or internet irony, it does reveal what actually fuels the movement. It’s not a first principles vision of what the world should look, act, feel, or be like but rather it’s the feeling of excess and capacity, a technological orgasm that we can have all in our midst, and the more that we can create the better all of human life will be. While this may be true from a material perspective, one need not venture far into any literary body of wisdom to discover the litany of works warning against the vices of gluttony, lust, and greed that speak gravely to the dangers of excessive pleasure. Without a conditioned will against vice and plenty, man begins to degenerate rather quickly, a universal tale of every civilization as the spoils of war from the successes of one's ancestors fuels degeneration and decadence in the next generations. No one has built a culture that’s broken that cycle yet but this is no good reason not to try overcoming this regardless.

Describing consciousness solely as an entropy-reducing force, an “extropic” force is a seriously callous description of the human condition that implicitly dismisses the mystery of life and its origin. We don’t know how to create life abiotically, thus far all primordial soup experiments attempting to animate life from inanimate matter have failed spectacularly (overview of the challenges of abiogenesis here). The amino acids have not magically transformed into self-replicating proteins as suitable precursors to proto-gene storing molecules and most definitely not multi-cellular organisms. Abiogenesis has not happened but even without it happening there can still be a discussion on whether machine consciousness can exist usurping any biological prerequisites. However this all boils down to a fundamental question, is consciousness an incontestable essence transmuting material part of an intergenerational continuum or is it solely a mechanistic process?

The e/acc mimetic optimism for building the machine god shoggoth shares close resemblance to the biblical Tower of Babel story, where all of man unites to erect a building that’ll pierce its way to heaven, thus leaving man’s transcendence up to his own volition rather than fate only for the dream to disintegrate as man collectively separates from a unified singularity into many pluralities as a common tongue between man is spoken no more.

There is a glaring contradiction and misalignment of incentives between life itself and the stated goals of e/acc, implicit through it’s “thermodynamic god” mimetics or otherwise. If life is defined, described, and justified solely as a hyper-rationalistic entropy-reducing system, an “extropic” system, then what exactly is the point of all this, to life itself? If any biological organism can be surrogated with some entropy-reducing mechanistic system instead, without a soul, without an essence, then what exactly are we building?

Given the quandaries of the primordial soup hypothesis and abiogenesis at large, I’m not convinced there was ever a rational reason for life to exist at all, most definitely not multicellular life, as life could’ve just remained as single-cellular autonomic organisms dedicated solely to removing entropy. Why does there need to exist a system any more complicated than an unimodal entropy heat pump? The consciousness of the mind generates many superfluous thoughts not directly conducive towards survival, proliferation, and most definitely not towards entropic pruning.

While many in the confederation of e/acc will prune the messaging of e/acc to their needs adding prefixes like fit/acc, bio/acc, ultra/acc, and any other permutation taking just the “wagmi” optimism but tailoring it to the person’s specific interests in mind without necessarily participating in the “shoggoth” chant, however, none of that negates the larger teleological question that looms.

Regardless of whether the Extropic is successful in accelerating generative AI applications, why should anyone be interested in e/acc if it’s committed to manifesting a post-biological future where any entropy-reducing system is seen as a surrogate for biological life. We wouldn’t have to be asking these questions if e/acc was birth in the light of an organically defined vision, but there lies the danger in birthing a framework for negating a perceived wrong rather than affirming a right.

Acceleration is not everything

While there are obvious teleological questions to be asked about e/acc, acceleration as a framing is not always useful, in fact, sometimes it can even be detrimental to progress. Accelerating is a false premise, it implies progress is always just one forward thrust away from paradigm-defining feats which couldn’t be further from the truth. Much of science and the discoveries of the natural physical world happened out of serendipity, we don’t have supreme mastery over the natural physical world which is why we must always be able to take a step back and think more broadly questioning our assumptions. Real paradigm defining progress has happened at the intersection of willpower, circumstance, and serendipity, what theory of relativity would there have been today had Einstein not been able to get a clear picture of the sun on the opposing ends of the earth during a full solar eclipse?

Domain-specific tunnel vision is a real phenomenon inhibiting scientific progress, thus it becomes imperative to wander astray asynchronously, organically resetting one’s brain to different baselines possibly more conducive to prognosticating new epiphanies. It can be as simple as losing yourself to the self where from afar these horizon-expanding events can appear to be “inefficient” and not conducive towards pushing the needle of knowledge forward which counterintuitively may be what exactly is needed for a true scientific revolution. A physicist can pause their work in physics, getting lost in literature which may lead them to the discipline of philology, fundamentally thinking about meta-conceptions of language at large possibly fueling a novel theoretical physics model of the universe. If every particular is part of a larger universal then truth can be arrived at by any fulcrum of a particular qualia, thus all domains of knowledge can refer back to this universal in some capacity. The great caveat is that this can’t be engineered externally, this type of curiosity can only come serendipitously from within, very much against “accelerating” as a heuristic for scientific progress and technological feats. (I go into more detail in this article here on asynchronous learning)

When one focuses on accelerating without stepping back to ask the question of “why” something is being done, you end up breeding faster horses rather than engineering automobiles which is exactly the problem with the framing of “accelerationing” that e/acc proposes. It doesn’t have a “why” for building the machine god other than the perceived euphoria stemming from sensations of more being better than less.

When humans are taken out of the loop for determining what needs to be built, the technological specifications and their measurements in isolation matter more than potential meaningful improvements in the quality of life from said technology. I remember when I used to obsess over specifications in the early days of the mobile revolution seeing how superior Android devices were in terms of tech specs, but in many cases, the iPhone produced a superior user experience solely because Apple focused on the human in the loop rather than technological specifications in isolation. Technology needs the reference frame of the human, otherwise, you end up building glorified tech demos rather than products people will use.

Had we chosen to blindly accelerate technology, especially considering early computers were not personal machines but initially communal ones, too much focus on making faster computers would have completely overlooked the use-case for making mobile devices. Obviously, there’s a threshold of compute per unit energy necessary for making mobile devices mobile, but this is exactly why we need to build technology in human first perspectives by taking a step back truly asking what we’re building and why we’re building.

However here lies the second paradox, we don’t always know the value of a technology until it’s built, hence why venture capital is necessary to validate these hypotheses. We need to be humble in recognizing we might build technology that doesn’t solve real problems while also recognizing we don’t know which technologies should be built before we build them every single time. With an “accelerationist” perspective wanting to build more and better for the sake of it, you’ll end up creating gaming monitors with 1000+ frames per second even if the human eye won’t meaningfully tell a difference between anything above 300+ fps. There lies the danger of building better technology for the sake of better rather than asking hard questions of why something needs to exist at all.

Conclusions

While there are serious teleological questions to be raised with e/acc, especially with its cavalier undertones towards the surrogacy of biological life with extropic, abiotic entropy-reducing systems, it’s optimism is still a beacon of hope among the ever-nascent narratives of doom humming in the background. However, this optimism should be grounded with the “why” question answered because simply directing that answer to “more” will drown man in vice, not actually directing man towards anything greater, especially without a proper conditioning of our own self-will. Had there been an organic, crystallized vision of what should be built we would not be having these discussions around what are its true motives and incentives.

E/acc would be better served renaming itself to e/coo — engineering cooperative, to relinquish itself from all “thermodynamic god” mimetic hype porn that the anon accounts throw around the internet as a self-flagellating participation trophy for “building the future”. Admitting that we don’t have a “why” but rather that we’re a bunch of engineers who just have the itch to build is a much better telos than the callous arrogance of abstracting that reason away into some “extropic” principle. The itch to build is a better “why” than the “why” created by the majority of human ideologies that have ever existed thus far.

For all the problems that we face collectively ranging between climate change and the fertility crisis, we should not be looking for an escape away from the biological substrate that imbues us with vitality. There’s no magical place for us to accelerate to that will solve any of the problems humanity is faced with today, we can only overcome by going through the toil rather than trying to “accelerate” our way around our biological grounding.

There exists no real alternative to building, which is the fundamental thesis that e/acc gets right. Even if the technologies we build may not always be solvent, let us be humble to recognize that fact rather than deluding ourselves into believing more technology is better because it’s measurably better than the prior iteration.

I love life enough not to want to find a way to accelerate out of it, I don’t see the appeal in summoning the thermodynamic machine god as much as I’m skeptical of the possibility of its emergence, but regardless we should try building it out of curiosity, not out of intent to surrogate biological life. We ought to be generating in the here and now, not out of this illusion of this purported utopia to come. Rather than projecting ourselves into this other world, we need to bring the imagination of that other world into the now. We must stand before and above time to generate out of the chaos and challenges that we face. We must overcome the challenges of our day if there’s any chance for there to be a tomorrow. There’s no alternative to building as the alternative to building is surrender which is cowardly and pathetic on all fronts.

Now is not the time to accelerate away from the problems that we as humanity face in the here and now, rather now is the time to take the leap of faith without expectation of return and make what time we have on earth count to its fullest. The time to generate is now — t3gin.

---

If you liked my work feel free to give me a follow on Twitter/X and check out my other Substack Generative Forms for more philosophy-specific content.