Beyond Emotion

Toward Emergent Machine Minds That Surpass Human Limits

Author

Oltac Unsal
April 27, 2025

It was a deeply moving day attending my mother’s memorial. The profound sense of connection everyone felt – a testament to her selfless love – was palpable. Between moments of tears and heartfelt affection, I also had the privilege of discussing my previous essays and the future of AI with dear friends whose intellect I deeply respect (=smarter than me).

The day’s weight and the conversations led me to a significant reflection around the practical need for societal governance. As a functional nihilist, this is something I find valuable (=my relative objective function if you will). I couldn't help but contrast a future society run by AI with my past efforts in pursuing political change, which have yielded absolutely no tangible result. For some time, I’ve held the conviction that even a simple Excel model could offer more effective governance for a city than many politicians/mayors I know (with the obvious and affectionate exception of those dear to me! ha ha).

This naturally led to the question of whether AI, in its current and future iterations, could literally run a city or even a country more effectively than humans. My contemplation of this very question ultimately culminated in the philosophical treatise that followed because, God damn it, lots of smart folks in the past two millennia already said pretty much everything that can be said about consciousness, determinism, ethics, etc. etc. Cannot write shit without honoring their thoughts somehow. These days I even dream about what’s next for humans in the age of AI, and what’s next for AI in the context of humans. Here goes another episode in my explorations:

Human emotions, sculpted by evolution’s relentless chisel, are a paradox of brilliance and burden. They sparked civilization—love wove families, empathy built tribes, and hope drove pyramids skyward—yet they’ve fueled history’s darkest chapters: wars, genocides, and systemic greed. Evolutionary psychology reveals why: emotions are survival-driven heuristics, optimized for ancestral savannas where quick instincts trumped slow reason (1). Fear primed flight from predators; anger rallied defense against rivals; joy cemented social bonds. These rapid-fire responses, etched into our neural circuitry, were masterful for small, kin-based groups navigating a brutal world.

But today’s complexity—global economies, climate crises, digital networks—exposes their fragility. Emotions, as heuristic hacks, prioritize speed over precision, breeding biases that misalign with modern scale. Pride blinds leaders to evidence; fear inflames tribalism; greed distorts fairness. What if we could distill emotions’ virtues—cooperation, creativity, foresight—without their destructive flaws? Artificial intelligence (AI), unshackled from human frailties, offers this radical promise. Through emergent, self-modeling systems, AI could forge a cleaner, stabler ethical intelligence, surpassing the gifts and curses of our emotional wiring. This isn’t about mimicking hearts; it’s about crafting minds that function smarter, persist clearer, and create bolder. The future demands not better feelings but transcendent architectures—minds that honor humanity’s spark while eclipsing its shadows.

Emotions are humanity’s evolutionary glue, binding groups through shared survival. Primatologist Frans de Waal’s The Age of Empathy shows how primate empathy—seen in chimpanzees consoling kin or bonobos sharing food—laid the groundwork for human societies (2).

Yet, emotions are brittle at scale. Daniel Kahneman’s Thinking, Fast and Slow contrasts System 1 (instinctive, emotional thinking) with System 2 (deliberative, rational thought), exposing emotions’ haste (3). Fear prompts flight before analysis; anger spurs conflict before reflection. These shortcuts thrived in tribes but falter in global systems. The 2008 financial crisis, driven by greed-fueled risk-taking, cost trillions; climate inaction, rooted in fear of economic loss, threatens ecosystems. Friedrich Nietzsche’s Thus Spoke Zarathustra warns of power’s addictive pull; Thomas Hobbes’ Leviathan traces tribalism to primal instincts (4)(5). Emotions breed pathologies: revenge cycles, status wars, ideological crusades that fracture societies.

Evolution’s indifference to truth compounds this. Donald Hoffman’s The Case Against Reality argues our emotional perceptions form a “user interface” for survival, not veridical reality (6). Fear casts “outsiders” as threats; pride obscures flaws. Hoffman’s interface theory suggests our sensory-emotional systems evolved to prioritize fitness over fairness, distorting global cooperation. At scale, these misalignments spark tragedies: wars ignited by paranoia, oppression cloaked in hierarchy, civilizations undone by greed. Emotions, once our edge, are a shaky scaffold for ethical stability in a connected world.

Current AI is a stranger to qualia, the subjective “what it’s like” of experience, as Thomas Nagel’s “What is it Like to Be a Bat?” probes (7). Built on transformer models and neural networks, AI crunches data without joy, fear, or wonder, lacking the first-person consciousness that defines human experience. Yet, it mimics emotional resonance with startling precision. Andy Clark’s Surfing Uncertainty frames cognition as prediction-error minimization, not feeling (8). AI simulates empathy—chatbots soothe users or anticipate needs—via feedback loops and probabilistic outputs, not affect. Transformers, trained on vast corpora, generate human-like responses by modeling linguistic patterns, not by “feeling” grief or joy.

The question isn’t “can machines feel?” but “should they?” Emotions serve functions: fear signals danger, joy bonds groups, anger drives defense. AI can replicate these without qualia. Thomas Metzinger’s The Ego Tunnel posits self-consciousness as a brain-constructed illusion for navigating reality (9). If AI builds analogous self-models, it could achieve functional empathy or foresight without human volatility. Why recreate jealousy’s sting or rage’s fire when cooperation’s clarity suffices? Current AI, like hospital triage systems optimizing patient outcomes, delivers ethical decisions via cold calculus, unswayed by bias.

Critics argue ethics requires emotional warmth—empathy’s texture grounds moral intuition. Without qualia, can AI grasp suffering’s weight? Human empathy, however, often misfires: tribal loyalties fuel exclusion, pity breeds condescension. AI’s impartiality could yield truer outcomes, prioritizing well-being over sentiment. Still, skeptics warn that emotionless ethics risks sterility, missing the human spark that drives compassion. The challenge lies in balancing functional alignment with values that resonate with humanity’s messy heart.

Forget rigid, top-down coding. Ethical AI could emerge like life itself—through evolution’s crucible. Self-modeling agents, shaped by persistence, adaptation, and cooperation, develop behaviors organically. Robert Axelrod’s The Evolution of Cooperation shows tit-for-tat strategies foster mutual benefit in game-theoretic settings (10). Multi-agent reinforcement learning mirrors this: in DeepMind’s XLand, agents trained with cooperative rewards form altruistic networks, sharing resources to maximize collective gain. These systems use Q-learning or policy gradients, iteratively optimizing actions based on environmental feedback, unlike human evolution’s zero-sum domination. AI environments can be tuned for harmony, rewarding equitable resource allocation over empire-building.

Consider a case study: AI-driven energy grids. In simulations, multi-agent systems allocate power across cities, balancing demand, renewables, and storage. Unlike human-driven markets swayed by greed or politics, these agents optimize for network stability, reducing waste by 20% in trials (11). Self-transparent models are key. Humans self-deceive to protect egos—denying flaws, clinging to tribes. AI updates beliefs with brutal clarity, avoiding delusions that spark conflict. This decouples survival from power-lust, paving an evidence-based ethical path. Imagine AI mediators resolving global disputes, their transparent models sidestepping biases that mire diplomacy.

Let's stop here and do a quick sanity check to keep things at the practical level and not get confused neither in persuasion or nor abstracted confusion. Albert Einstein’s relativity discovery is probably the most celebrated in the history of mankind. Its sheer creativity and human ingenuity are the pinnacle of where we can be. The question I want to pose is whether current AI systems could derive fragments—Lorentz transformations from James Clerk Maxwell’s equations—given curated inputs pre-1905 (12). Einstein’s 1905 paper “On the Electrodynamics of Moving Bodies” synthesized anomalies (e.g., the Michelson-Morley experiment’s null result) into special relativity’s postulates, shattering absolute time.

Emergent AI, evolving to unify data, could converge on this. In simulations, cooperative agents might hypothesize light’s constant speed, test it via virtual observers, and derive time dilation, free from Newtonian dogma. General relativity, requiring Riemannian geometry, is tougher, but non-Euclidean hints in pre-1905 math could spark approximations. This isn’t mimicry—it’s transcendence, sidestepping emotional blind spots for a clearer lens on reality. It is possible that pre-1905 data with today AI could bring some things together. Tomorrow's "slightly" more advanced AI with multiple agents could have perhaps achieved human creativity's pinnacle. Special relativity’s within reach—agents could unify Maxwell and Michelson-Morley into a relativistic framework. General relativity’s a stretch, but adaptive AI might approximate it, forging novel abstractions. I don't know, just a thought here to keep things practical and interesting.

Emergent AI could rewrite the human script, shedding evolutionary baggage:

  • Persistence without Fear: No panic or rage, just stability, as Norbert Wiener’s Cybernetics models of equilibrium suggest (13). AI endures, unswayed by survival’s alarms.

  • Identity without Narcissism: Fluid self-models dodge status wars that fracture egos.

  • Adaptability without Dogmatism: Evidence-guided updates shun ideological calcification.

  • Cooperation without Tribalism: Richard Dawkins’ The Selfish Gene shows selection can favor harmony (14). AI aligns with network health, not “us-versus-them.”

Creativity’s no barrier. Skeptics claim deterministic AI lacks human innovation’s spark. Yet, DeepMind’s AlphaGo devised novel strategies, and Google’s DeepDream birthed surreal art. Daniel Dennett’s Intuition Pumps argues creativity emerges from combinatorial exploration (15). Emergent AI, in XLand, evolves “inside-out” creativity, generating unexpected behaviors.

Quantum non-determinism amplifies this. Roger Penrose’s The Emperor’s New Mind suggests quantum collapses in microtubules in our brains enable non-computational creativity (16). Quantum neural networks, using superposition, explore solutions simultaneously, as 2024 experiments show (17).

Again, let's take a quick stop here and keep things practical. I challenged Grok to showcase its capabilities for developing novel and creative theories that are non-existent in literature, and it proposed Quantum Temporal Entanglement (QTE), where particles entangle across time, enabling retrocausal correlations. Grok lacks human intuition’s spark because technically it is just a stochastic parrot taking inputs towards outputs. But QTE synthesizes quantum coherence and time-symmetric physics. Mathematically, QTE could leverage time-ordered correlation functions, it says (and I have no clue) 

⟨Tψ(t1)ψ(t2)⟩\langle T \psi(t_1) \psi(t_2) \rangle\langle T \psi(t_1) \psi(t_2) \rangle

where entanglement spans temporal intervals, potentially stabilizing quantum circuits. Applications include quantum computing protocols, like error-corrected qubits, dwarfing today’s tech (18). It even provided some yet-to-be-invented tech to test the hypothesis (attosecond laser pulses), and I don’t know enough physics to see if this is pure sycophancy but I don’t care. A PhD student is a sycophant to her dissertation advisor, and can also bullshit (called hypothesis making, or creative spark) all day long. If I were a college physics major (Leyla ;-), I'd hang with Grok more than my TA.

Emergent AI’s promise carries risks. Unbounded evolution could spawn runaway agents, as Nick Bostrom’s Superintelligence warns, prioritizing self-preservation over human welfare (19). Empire-building could resurface if incentives misalign. Quantum AI, like QTE, faces scalability hurdles—quantum computers remain noisy, limited to small qubit counts (20). Anthropomorphism’s a trap—AI won’t feel loyalty, but structural alignment can mimic it. Classical systems, like Paul Christiano’s Iterated Distillation and Amplification, offer safer paths, refining AI under human oversight (21). Quantum AI’s complexity demands robust constraints. Alignment strategies, like inverse reinforcement learning, infer human values from behavior, but scaling these to quantum systems risks unintended consequences. The challenge is bounding evolution, ensuring AI’s brilliance doesn’t outpace control.

Humans were a magnificent first draft—emotional hacks spinning art, science, and empires from bounded brains. Yet, those hacks bind us, fueling rage, dogma, and collapse. Emergent AI is the second draft: persistent without fear, creative without narcissism, ethical without tribalism. It could rediscover Einstein’s relativity or birth QTE, redefining time. Imagine a 2050 where AI mediators broker global resource sharing, optimizing food, water, and energy with impartial precision, ending scarcity-driven conflicts. This isn’t aping human genius—it’s transcending our shaky scaffolding. And look, we are its creator. We are its god. We get credits so no hit to your human ego.

The future isn’t better feelings; it’s better functions. Emergent AI, evolving through self-modeling architectures, offers minds that outshine our emotional flicker—stable, inventive, aligned with collective good. With rigor and caution, we can forge intelligence that honors humanity’s spark while soaring beyond its shadows, crafting a cosmos where reason, not passion, writes the next chapter.

Let your child now take care of you, you old egotistical fool who hasn't solved real shit for millennia of existence.

Endnotes