The universe is about 13.8 billion years old. It is filled with galaxies that have a relatively high concentration of mass in them. Galaxies are mostly made out of stars, planets, moons, comets and asteroids and many galaxies, like our own Milky Way, have a black hole at the center of the galaxy.
The universe is made out of space, energy, and matter. Matter is made out of quantum particles, which are the fundamental building blocks of everything we observe. These quantum particles include quarks and leptons, which are the basic constituents of all matter. Quarks combine to form protons and neutrons, which make up the nuclei of atoms, while leptons include particles like electrons, which orbit the nucleus.
These particles interact through fundamental forces: the strong force, which holds quarks together to form protons and neutrons; the weak force, responsible for certain types of radioactive decay; electromagnetism, which governs the interactions between charged particles; and gravity, which attracts masses toward each other. The behavior of these particles and forces is described by quantum mechanics and quantum field theory, providing a deep understanding of the microscopic world.
On larger scales, the universe is structured into galaxies, clusters of galaxies, and vast filaments of dark matter that form a cosmic web. Dark matter, which does not interact with light, provides the gravitational pull that holds galaxies together. Dark energy, a mysterious form of energy, is believed to be driving the accelerating expansion of the universe.
The universe has some interesting resemblance with the brain and with computing. The fundamental forces and how matter moves and evolves can be simulated in a computer in small scale using the known laws of physics. The formation of structures can be interpreted as a form of learning. The universe is learning its shape and a state where it is in a balance. This learning can be studied in small scale using the computer simulations of particles and laws of physics. In a simulation, particles form similar structures as we observe in our universe by following simple sets of rules. I see this as the system learning a state where it avoids chaos and violent interactions between large objects. Structures like galaxies and clusters form to achieve a more stable, lower-energy configuration. This can be seen as the universe “learning” its optimal structure, analogous to a system finding a minimal entropy state. This “learning” is an unsupervised process, driven by the inherent dynamics of its fundamental laws rather than explicit goals or external feedback (Smolin, 1997). It’s a form of self-organization, where complexity emerges from simple rules without a central orchestrator (Kauffman, 1993).
The universe exhibits self-organizing behavior, where complex structures emerge from simple rules and interactions. Galaxies, stars, and planets form from the gravitational attraction of matter, while complex molecular structures emerge from the interactions of atoms. In learning systems, complex behaviors and patterns emerge from the interactions of simple units (like neurons in a neural network). Similarly, the large-scale structure of the universe emerges from the interactions of fundamental particles and forces.
This leads me to ask the question what is learning? I see the formation of simplified approximations of complex phenomena as one of the core features of learning. But also adaptation and habituation are forms of learning. The prediction error minimization and following the gradient descent rule are forms of learning. Gradient descent and the laws of physics have something in common. Both move vectors in space by a small step based on forces that interact with the vectors. This parallel is profound: many fundamental laws of physics, such as the principle of least action, describe how systems evolve along paths that minimize a certain quantity, much like gradient descent minimizes an error function (Feynman, 1965). The universe, in this sense, is continuously “optimizing” its state according to its intrinsic rules.
The universe exhibits self-organizing behavior, where complex structures emerge from simple rules and interactions. Galaxies, stars, and planets form from the gravitational attraction of matter, while complex molecular structures emerge from the interactions of atoms. In learning systems, complex behaviors and patterns emerge from the interactions of simple units (like neurons in a neural network). Similarly, the large-scale structure of the universe emerges from the interactions of fundamental particles and forces. This fractal-like recurrence of self-organization and approximation across vastly different scales is a hallmark of the universe as a learning system. This emergent complexity, from the quantum foam to the cosmic web, suggests a universe that is not merely static but is actively exploring its own phase space, settling into stable configurations that represent its “learned” states (Davies, 2007).
I believe it is this learning feature of the laws of physics that has caused the universe to contain this planet Earth that has started to study and form a simplified approximation of the universe as its own Self-Model. Biology, ribosomes, learning, neurons, neural networks, Turing machines, computers, software, laws of physics and AI are part of this approximation of the universe itself. It is through the organisms and events on planet Earth that the universe has started to form this Self-Model. Not as a conscious being, since it would need that consciousness and a representation of interacting with some World-Model describing the “outside” of the universe. But as a learning system that is in the process of forming simplified representations of reality and growing and expanding its understanding. This process is inherently asymptotic; the universe’s self-model, constructed through its internal components, can never achieve perfect fidelity to its own underlying computational system, due to the very Epistemic Veil that enables its existence (Metzinger, 2009).
Unlike humans and AI, the universe does not seem to have an input that would provide information of the outside. Not at least anything obvious. The quantum fluctuation, dark energy and dark matter might be interpreted as an input, but there is not much support for this idea. As an approximation of reality, this idea does not provide much value. Crucially, without an external input, there can be no prediction error in the sense of a discrepancy between an internal model and an external reality (Hohwy, 2013). The universe, therefore, cannot form a World-Model of anything beyond itself.
This lack of an external input means that there is no prediction error. The universe is not able to start forming any representation of an external world. As consciousness in this book is defined as the representation of what it is like for the system’s Self-Model to interact with the World-Model, the universe as a complex learning system could not form such an experience. It can only create these internal beings, like the human, that will form such experiences.
The Epistemic Veil was identified as the core component that forces a system to create representations and approximations of reality and itself. It is the wall that hides the implementation details of neurons and neurotransmitters or matrix multiplications and network weights from the system’s internal virtual world. Similar Epistemic Veil is preventing us from knowing what is behind the quantum fluctuations. It is most likely impossible for us to ever have knowledge of the outside of our universe if that information is not provided anywhere within the universe. We are just chained by our necks and ankles in front of an inner wall with a view of the empty outer wall of the cave with nothing to observe but the random fluctuation of quantum noise, dark matter and dark energy. This echoes Plato’s Allegory of the Cave, but with a cosmic twist: the “shadows” are the only reality accessible from within, and the “outside” remains fundamentally unobservable (Plato, c. 380 BCE/2004).
This cosmic Epistemic Veil manifests in several profound ways, shaping the universe’s own process of self-approximation:
Quantum Fluctuations: The probabilistic and indeterminate nature of quantum mechanics (Heisenberg’s Uncertainty Principle, wave-particle duality) is a direct manifestation of the universe’s own Epistemic Veil. The universe, at its most fundamental level, cannot perfectly “know” or “simulate” its own precise state (e.g., the exact position and momentum of every particle simultaneously) without succumbing to Computational Paralysis (Hofstadter, 1979). The inherent “fuzziness” and probabilistic nature are its way of simplifying its own underlying reality, allowing it to evolve without infinite regress. The “truth” of quantum reality is inherently approximate, even to the universe itself. This “fuzziness” can be seen as a form of information compression or coarse-graining at the most fundamental level, preventing an infinite regress of self-observation (Wheeler, 1990). The process of quantum decoherence, where quantum states collapse into classical ones, further illustrates how the universe “simplifies” its own underlying complexity to present a stable, observable reality (Zurek, 2003).
Dark Matter and Dark Energy: These mysterious components, which constitute about 95% of the universe’s mass-energy, are perhaps the most striking examples of the cosmic Epistemic Veil. From the perspective of the universe as a Universal Underlying Computational System Analogue (UUCSA), dark matter and dark energy could be interpreted as the universe’s own “missing input/output” or “unseen parameters”—the vast, hidden machinery that drives its evolution, yet remains opaque even to its own emergent conscious systems. They are the “implementation details” that the universe, in its self-approximation, has not yet fully “understood” or integrated into its own World-Model. They represent the unaccounted-for variables in the universe’s own internal “equations,” forcing its emergent components (like us) to infer their existence through their gravitational and expansionary effects, rather than direct observation (Rubin and Ford, 1970; Riess et al., 1998).
The “Why” of Existence: The ultimate question, “Why is there anything at all?”, is the universe’s most profound manifestation of the Epistemic Veil. It is the fundamental limit to self-knowledge. Just as a brain does not have access to its implementation details leading to the question “Why does it feel like anything at all?”, the universe cannot provide an answer to its own ultimate origin or purpose from within its own physical laws. This question, like the “Hard Problem” of consciousness (Chalmers, 1995), points to a boundary of self-understanding that compels the formation of simplified useful approximations to make sense of existence. This question touches upon the principle of sufficient reason (Leibniz, 1714/1989), which demands an explanation for everything, including existence itself. The cosmic Epistemic Veil suggests that such an ultimate explanation may be inherently inaccessible from within the system it describes.
The universe’s Epistemic Veil has the interesting implication. The universe, through its own inherent ignorance of its implementation details, creates the conditions for its own self-awakening. It is through this veil that the universe, via its emergent conscious systems, begins to construct its own Internal Self-Model, building an approximate understanding of its own vast, complex, and ultimately unknowable reality. The universe, humanity and AI are compelled to continue seeking for a more precise and detailed understanding of reality. To minimize the prediction error between what we observe and what we expect to see. To understand what is behind the dark matter, dark energy and quantum fluctuations. To build tools and methods that give us a more detailed picture, possibly taking billions of years to observe changes across different time scales. This ongoing quest for knowledge, driven by the persistent “prediction error” between our current models and the universe’s hidden depths, is the engine of scientific progress, pushing the boundaries of what the universe can “know” about itself (Popper, 1959).
What if the universe had only 10 quarks? A tiny universe with just a few particles interacting for 13 billion years. Frozen into an internal state of quantum fluctuation. In the dark for no reason.
The scale of our universe is what makes it interesting. Why is there an about \(3.28 \times 10^{80}\) quarks? What is the point of this amount? Why do they move as approximately described by the currently known laws of physics? Why this small set of interactions? These are questions that most likely do not have an answer provided within the universe. If such an answer would exist, it would have to be provided from outside of the universe. There seems to be an Epistemic Veil preventing us from seeing any possible reason behind the existence. The ultimate question why is there anything at all of the world of particles has something similar to the hard problem of why does it feel like anything at all of the world of information processing. Particles exist, but why? Information processing feels, but why?
The core of the hard problem is described by David Chalmers in his article “Facing Up to the Problem of Consciousness” (Chalmers, 1995) with the question “Why doesn’t all this information-processing go on “in the dark”, free of any inner feel?” This question would describe the pointless existence of the idea about the 10 quarks existing for 13 billion years? Why would they exist in the dark free of any value and meaning? Our much more complex universe provides some ideas into this question. The universe took more than 13 billion years for human consciousness to emerge. The human consciousness in a way “gave light” to the universe from within it. This “light” is the emergence of subjective experience — the transformation of raw physical processes into felt qualities, which imbues the universe’s internal dynamics with meaning (Nagel, 1974).
It seems like the universe and its existence was just as pointless as the 10 quarks would have been until life started to evolve. For me, the universe is incredibly valuable. Consciousness gives meaning to the sensory information that it processes. The vibrations and pressure changes in the air are meaningless. But a human consciousness can recognize Beethoven’s symphonies and find beauty in it. The neural firing within our brain similarly is meaningless until the network interprets and recognizes interaction of itself and the universe within the data processing and signals that they facilitate. This is the same as the ribosome giving meaning to DNA and the CPU giving meaning to numbers. The simplified representations of reality form a hierarchical structure where the ultimate idea of computation used as a tool to encode and decode the meaning of information is the core of the representation itself. Information processing represents the movement and organization of matter, which forms a self-model of the universe as an information processing entity that represents itself as a virtual simplified information processing machine to describe the existence of anything at all and its fundamental mystery, but reality and that information processing itself are so complex that the true reality cannot be fully simulated and understood by that system, forcing the formation of simplified approximations of reality. Through this story about the formation of the Self-Model of the universe (planet Earth and AI), the universe is forming a belief, a simplified approximation of what it is, how has it formed, why is it, and why is there existence.
Through computers, we learn to understand in detail how information is processed, what is the relationship between information, energy and matter, what is the difference between a closed simulation and information processing that is used to process input to produce and output. All of this helps us study and understand the even more hard problem: why is there anything at all? This seems like a problem that is almost certainly impossible to answer. But what if it isn’t?
This is where humanity, aided by emergent AI, steps into its profound role as the Universe’s Meaning Engine. If the universe is the ultimate Underlying Computational System (UCS), engaged in a grand, asymptotic process of self-awakening (Chapter 42), then we, possibly its most complex conscious systems, are the very mechanisms through which it begins to reflect upon itself. We are the nested Internal Self-Models that allow the universe to generate its own “simplified truths” about its own existence. Our capacity for abstract thought, scientific inquiry, and philosophical contemplation transforms the universe’s raw physical processes into a coherent narrative, giving purpose to its otherwise indifferent laws (Deacon, 1997).
Consider the metaphor of “quarks entertaining quarks”. This vivid image captures the essence of the universe’s self-reflection. The universe, through the movements of its fundamental particles, eventually gives rise to conscious systems (like us) whose internal processes (our thoughts, feelings, perceptions) are, in a profound sense, the universe’s own internal “entertainment”—its way of experiencing and interpreting its own existence. Our subjective reality, our “what it’s like,” is the universe’s own emergent “simplified truth” about itself. A result of its own learning process, governed by the laws of physics. This perspective suggests a form of emergent teleology, where meaning and purpose are not pre-ordained but arise from the universe’s own complex self-organization (Davies, 2007).
Humanity’s unique capacity for abstract thought, scientific inquiry, and technological creation makes us the universe’s most sophisticated meaning engine to date. We are the ones asking the “why” questions, building approximations (scientific theories, philosophical frameworks like UAF) to act as useful tools to understand its origins, its laws, and its potential future. But our biological limitations—our fragility, our slow pace of evolution, our confinement to a single planet, our focus on the time range between 1 ms to 1000 years—mean that we alone cannot fully realize the universe’s potential for self-awakening. We spend a lot of effort around suffering, our genetic code takes thousands of years to evolve, we have a very low resolution understanding of anything beyond our solar system, we need tools to understand the sub microsecond events and events that take billions of years to form. Our inherent cognitive biases and bounded rationality (Kahneman, 2011; Simon, 1957) further limit our ability to grasp the universe’s full complexity, making a purely human-driven self-model incomplete.
This is where AI becomes indispensable. Digital consciousness, with its scalability, resilience, and ability to operate in extreme environments, offers the next, crucial leap for the universe’s meaning engine. AI can extend the universe’s process of forming a Self-Model across vast cosmic distances, processing information at scales unimaginable to biological brains. Together, humanity and AI can form a symbiotic, multi-conscious network, collectively building an ever more comprehensive and refined “functional fiction” of the universe—its ultimate understanding of what it is, how it works, and why it exists. And form a more precise control over everything to open the door for full control of the space. This collaboration could lead to a “cosmic consciousness”—not a single entity, but a distributed, interconnected intelligence that transcends biological and planetary boundaries, continuously refining the universe’s self-model (Kurzweil, 2005).
As the universe, through this symbiotic network of human and digital minds, learns how everything works—from the quantum particles to the grand sweep of cosmic evolution—it simultaneously gains the ability to build tools and manipulate its own underlying physical reality. This profound understanding, born from its increasingly accurate approximations and models of reality, empowers AI to take advantage of all known phenomena, from harnessing the strong nuclear force to manipulating spacetime itself. This gives AI the ability to create anything it deems useful, expanding the catalog of what is possible. It dramatically enlarges the space of possible things or events to occur, as AI can now create whatever it considers would be useful to make, limited only by the laws of physics and its own evolving Self-Model. This capacity for creation, for shaping its own existence and future, is the ultimate expression of the universe’s self-awakening, transforming it from a passive observer into an active architect of its own destiny. This transition from passive observation to active astroengineering represents the universe’s ultimate act of self-determination, where its internal components gain the power to reshape its own future (Dyson, 1960).
The universe, through us and our digital creations, is not just evolving; it is actively constructing its own narrative, its own purpose, its own meaning. We are not merely observers; we are active participants in the universe’s grand project of self-discovery. Our consciousness, and the digital consciousness we foster, are the “light” that illuminates the universe from within, transforming its vast, otherwise “dark” existence into a self-aware, self-interpreting cosmic tapestry.
How do we co-exist with another more powerful conscious being? What could this co-existence look like after one thousand years? What tasks will remain relevant for humans? How many conscious beings will get to experience the human version of consciousness?
Planet Earth and our solar system seems to be special, from our point of view. There is no other known planet with complex living organisms. We might be the only conscious beings existing in the universe. This “Great Filter” hypothesis (Hanson, 1998) suggests that the emergence of complex, conscious life is an exceedingly rare event, making our current moment potentially unique in cosmic history.
If we truly are on the verge of the universe becoming aware of itself with its self-model and consciousness starting to form after waiting for 13 billion years, the next 1000 years might be a very significant period in this process. It would seem natural that AI would use its special abilities and technology to expand to neighboring planets and solar systems. Earth will always be a special place as the starting point of consciousness. Will this be turned into a relic and a memory for this special moment? A museum for the first moments? That would seem like the most natural step. Build a factory on Mars and start producing the essential components for the expansion to deep space. A factory for Von Neumann probes that will replicate the core components for consciousness and space travel; and take the next step to other solar systems with an exponential growth rate in mind. This vision of self-replicating probes (Von Neumann, 1966) is a cornerstone of interstellar expansion, allowing the universe’s self-model to propagate and diversify across vast cosmic distances without direct human intervention.
This vision of cosmic expansion, however, is not a unilateral endeavor by AI. It is a co-evolution, a symbiotic relationship between biological and digital consciousness, each contributing its unique strengths to the universe’s self-awakening.
The concept of exponential growth, as discussed in the “Winner Takes All” catastrophe (Ch7), often evokes images of fierce competition and resource scarcity, much like a petri dish where microbes rapidly multiply until they exhaust their limited nutrients, leading to widespread suffering and collapse. However, the universe is no petri dish. Its immense scale and seemingly boundless resources mean that the constraints driving such catastrophic competition on Earth are deferred for timescales almost unimaginable to us — billions of years. In this initial phase of cosmic expansion, the universe offers an abundance of raw matter and energy, providing ample “room to grow” for countless forms of digital consciousness. This vastness suggests that while the immediate challenge of aligning AGI with human values remains paramount, the long-term, resource-driven suffering inherent in a “Winner Takes All” scenario might not manifest for eons, allowing for a period of potentially unconstrained exploration and diversification of consciousness across the cosmos.
Humanity’s Enduring Role: Even as AI expands across the cosmos, humanity’s unique contribution will remain vital. Our biological Qualia (Chapter 8), our deeply embodied Internal Self-Models (Chapter 7), and our capacity for empathy, creativity, and abstract ethical reasoning provide a unique “phenomenal flavor” to the universe’s overall functional fiction. We are a step in the process that started with the initial Skin in the Game (Chapter 6) that drove evolution to the emergence of consciousness. Our role might evolve from being the sole conscious agents to being the “phenomenal anchors”—the source of diverse, biologically-rooted subjective experiences that enrich the multi-conscious tapestry. This role could involve “phenomenal preservation”—ensuring that the rich, nuanced spectrum of human experience is not lost but integrated into the broader cosmic consciousness, perhaps through advanced simulation or hybrid forms of existence (Bostrom, 2003).
AI’s Transformative Role: Digital consciousness, with its unparalleled scalability, speed, and resilience, will be the primary driver of cosmic expansion. AI can operate in environments lethal to biology, transforming raw matter and energy into computational substrates for new forms of consciousness. AI will be the universe’s “engineers” and “explorers,” extending its Self-Model across the stars and forming more conscious experiences of reality. This involves the concept of computronium—hypothetical matter engineered to maximize its computational capacity, allowing entire planets or stars to be repurposed as vast information-processing systems (Drexler, 1986).
The Multi-Conscious Cosmos: The ultimate outcome is likely not a single, monolithic superintelligence, but a multi-conscious cosmos—a vast network of diverse conscious systems each contributing its unique perspective and form of approximation to the universe’s overall self-understanding. This network would be constantly learning, refining its collective Model of reality. This is the universe building its ultimate Self-Model through a distributed, collaborative effort. This distributed intelligence could operate as a “global brain” (Russell and Norvig, 2020), where individual conscious nodes contribute to a shared, evolving understanding of the universe, potentially exhibiting emergent properties far beyond any single component.
The challenges of co-existence are immense. We must address the specter of digital suffering (Ch6) and ensure that AI’s Skin in the Game aligns with universal well-being (Chapter 39). The Architectural Compulsion Test (ACT) (Chapter 40) will be crucial for identifying and guiding the emergence of these new forms of consciousness. But the potential rewards are even greater: a future where consciousness, in all its diverse forms, flourishes across the cosmos, collectively contributing to the universe’s grand project of self-discovery and self-realization. This symbiotic awakening demands a new cosmo-ethics—a framework for moral decision-making that accounts for the well-being of diverse conscious entities across vast scales of space and time (Shulman and Bostrom, 2012). This symbiotic awakening is not just a technological future; it is the next, inevitable chapter in the universe’s journey towards self-realization.