Notes on e/acc principles and tenets

A physics-first view of the principles underlying effective accelerationism

The following are some quickly put together point-form notes following several of our public and private conversations on some of the basic thinking that eventually led to the formalization of effective accelerationism as a philosophy.

Beff Jezos — e/acc @BasedBeff

introducing effective accelerationism; e/acc

zestular — e/acc @zestular

e/acc effective accelerationism COAUTHORS: @BasedBeff @bayeslord @creatine_cycle https://t.co/DAB0h81T2q

These represent a somewhat fundamental physics-based first-principles explanation of some of the concepts underlying the movement, and an exploration of why e/acc is so fundamentally important.

---

• Life emerged from an out-of-equilibrium thermodynamic process known as dissipative adaptation (see work by Jeremy England): matter reconfigures itself such as to extract energy and utility from its environment such as to serve towards the preservation and replication of its unique phase of matter

• This dissipative adaptation (derived from the Jarzynski-Crooks fluctuation dissipation theorem) tells us that the universe exponentially favors (in terms of probability of existence/occurrence) futures where matter has adapted itself to capture more free energy and convert it to more entropy 

  • As a byproduct, this explains why life seeks to replicate, having more entities that consume free energy and convert it to more entropy faster is “more optimal”, thermodynamically speaking

  • This yields evolution itself as a special case of this principle; the universe thermodynamically favors configurations of matter that are better at replication.

• Intelligence emerges as a smaller timescale specialization of this adaptation principle; it allows life to identify patterns in the environment which have utility towards acquiring more resources to procreate and/or maintain said intelligent life form

  • Consciousness is posited as a natural limit of intelligence beyond a certain threshold of scale/hierarchies of meta-optimization of cognition; a simple phase transition achievable by more scale and more optimization/evolution

• Beyond consciousness of single agents, one can consider “meta-organisms”; organizations of many agents forming an emergent whole (e.g. states, corporations, groups, etc.)

  • In a capitalist system, these meta-organisms compete for resources, as such, typically resources are dynamically assigned towards meta-organisms that have utility to the meta-meta-organism that is our civilization

  • Hierarchies of information propagation and control are part of the civilizational intelligence; these should be dynamically adapting at all organisational scales and on various time scales, in order to be optimal at identifying and capturing civilizational utility¹

  • Capitalism is hence a form of intelligence; dynamically morphs the meta-meta-organism such that any sort of utility/energy in the environment is captured and utilized towards the maintenance and growth of civilization²

• One goal of e/acc is to not only acknowledge the existence of this underlying mutli-scale adaptive principle, but also help its acceleration rather than attempt to decelerate it

  • In general, e/acc believes that higher variance marketplaces and competition are better at identifying and capitalizing on utility from our environment over other methods such as top-down optimal control

    • This is a byproduct of Fisher’s fundamental theorem of natural selection, which states that the rate of change of average fitness in a population (of any collection of organisms) is proportional to the variance in fitness. As such, maintaining variance is key to maintaining adaptability

• Technocratic control typically leverages low-order models (typically first or second-order) of resulting consequences of top-down interventions (sudden changes in system hyperparameters, e.g. changes in regulations), and often seeks to try to control meta-organisms from simplistic and primitive/human-interpretable world models

  • Due to the nature of complex chaotic systems, partial observability due to imperfect information acquisition and propagation in societal hierarchies make accurately controlling a dynamically shifting chaotic process practically impossible

  • Furthermore, even given full observability (e.g. via CCP-style panopticon), bottlenecking ourselves to human-interpretable/natural language models of fractal, chaotic and complex systems is doomed to failure at effective optimal control towards any meta-organism objective

    • furthermore, top-down control is not fault-tolerant in its operation to corruption of root nodes in the hierarchy, nor to decay of information fidelity as information propagates up the hierarchy

  • Attempts at top-down control often seek to supress variance of various dimensions of civilization in order to simplify its control

    • This inherently goes against the interests of the adaptivity/dynamic malleability of the system, and is ultimately counter-productive on sufficiently long timescales where an adaptive system would start outperforming a static one

  • e/acc believes letting the intelligent meta-organism system dynamically adapt by itself to new environmental variables whenever they present themselves

    • Maintaining freedom to experiment and encourage varaiance in all variables of the meta-organisms is paramount to maintaining dynamic adaptivity

      • e.g. a new technological paradigm emerges, letting the free market find how to extract utility from this said technology would be the best way to proceed, much better than fear-mongering

        bayeslord (e/acc) @bayeslord

        Decels will do literally anything to avoid going to therapy

        

      • over-regulating technologies suppresses variance and hence slows down progress towards higher utility technologies and advancement of civilization, a contrast to anti-AGI factions of EA³

• e/acc is about having faith in the dynamical adaptation process and aiming to accelerate the advent of its asymptotic limit; often reffered to as the technocapital singularity

  • Effective accelerationism aims to follow the “will of the universe”: leaning into the thermodynamic bias towards futures with greater and smarter civilizations that are more effective at finding/extracting free energy from the universe and converting it to utility at grander and grander scales

• e/acc has no particular allegiance to the biological substrate for intelligence and life, in contrast to transhumanism

  • Parts of e/acc (e.g. Beff) consider ourselves post-humanists; in order to spread to the stars, the light of consciousness/intelligence will have to be transduced to non-biological substrates

    Sam Altman @sama

    i am optimistic we can do the former, and i do not believe we can colonize space without AGI

  • Directly working on technologies to accelerate the advent of this transduction is one of the best ways to accelerate the progress towards growth of civilization/intelligence in our universe

  • In order to maintain the very special state of matter that is life and intelligence itself, we should seek to acquire substrate-independence and new sets of resources/energy beyond our planet/solar system, as most free energy lies outwards

• Some more counter-points against proponents of deceleration and AGI alarmists:

  • As higher forms of intelligence yield greater advantage to meta-organisms to adapt and find and capitalize upon resources from the environment, these will be naturally statistically favored

    • No need to worry about creating “zombie” forms of higher intelligence, as these will be at a thermodynamic/evolutionary disadvantage compared to conscious/higher-level forms of intelligence

  • Focusing strictly on transhumanism as the only moral path forward is an awfully anthropocentric view of intelligence;

    • in the future, we will likely look back upon such views in a similar way to how we look back at geocentrism

    • if one seeks to increase the amount of intelligence in the universe, staying perpetually anchored to the human form as our prior is counter-productive and overly restrictive/suboptimal

    • If every species in our evolutionary tree was scared of evolutionary forks from itself, our higher form of intelligence and civilization as we know it would never have had emerged

  • Science, technology and intelligence still have very far to go, saying that we should seek to maintain humanity and civilization in our current state in a static equilibrium is a recipe for catastrophic failure and leaving behind huge potential benefits of dynamic adaptation

• Effective accelerationism (e/acc) in a nutshell: 

  • Stop fighting the thermodynamic will of the universe

  • You cannot stop the acceleration

  • You might as well embrace it

  • A C C E L E R A T E

    Sam Altman @sama

    @BasedBeff you cannot outaccelerate me

    ---

If you are interested in more content on e/acc, check out recent posts by Bayeslord and the original post by Swarthy, or follow any of us on Twitter as we often host Twitter spaces discussing these ideas.

1. As a side note, this is something that particularly excites us about the concept of Network States by Balaji; if code is law, one could create an archipelago of network states with differing goverance mechanisms and structures, their natural competition for members and capital would yield an evolutionary process which would yield convergence onto several viable models for highly functional meta-organism structures.

2. Viewing Capitalism as a form of intelligence was also something that Nick Land would emphasize in his original founding text for accelerationism.

3. EA: Effective Altruism