The Binding Force
Cooperation, information and the upbringing of AI
“Two things fill the mind with ever new and increasing admiration and
reverence, the more often and more steadily one reflects on them:
the starry sky above me and the moral law within me.”
Immanuel Kant, 1788
In Pursuit of Harmony
Whether working on software architecture or improving the interior design at home, I always pay close attention to how systems make me feel. When they are tidy and logical — when each individual part is intentional, and they are united to form a whole — it feels satisfying to be exposed to them. Why would our brain find these specific features so deeply pleasant?
Neuroscience provides the answer. It turns out we favor harmonious places for the exact same reason we prefer rhythmical music or clear instructions: they are fundamentally easier to understand and categorize. They consume less energy.
Our brain works as a massive predictive engine, constantly trying to make sense of the world around us based on incoming data. To do this, it builds a comprehensive inner model of the world to anticipate potential future scenarios. To calibrate that model, our brain measures the gap between its predictions and the incoming data.
A small gap indicates a well-calibrated model. A large one signals that reality does not confirm the model’s assumptions. Our brain tries to minimize the gap either by changing its model to better fit the data, or by changing the world to fit its existing model. This error-minimization process requires precious processing energy; the bigger the error, the more energy it demands.
Evolution sharpened our predictive modeling skills. Our ancestors, who could better understand the causal dependencies of their surroundings, had better survival chances and consumed less energy processing environmental signals, leaving more resources for what mattered: finding food, avoiding predators, navigating social bonds.
Interior designers often recommend choosing a low-intensity, neutral color as a dominant base — gray or beige, for instance. That color choice is not accidental. Such colors are remarkably close to the average spectrum of our natural environment, and so they barely deviate from our brain’s expectations; the metabolic cost of processing them is minimal.
This is precisely why choosing bright, flashy colors as a base is widely discouraged. In nature vibrant colors are rare — it is a marker of biological significance. It signals danger, toxicity, or reward. We read it as a signal demanding heightened alertness, one that requires heavier neural computation to interpret and resolve.
At the same time, a room kept entirely in gray would feel sterile and boring. To solve this, designers recommend introducing a contrasting accent color. At first glance, this seems like a contradiction from the standpoint of computational efficiency. A new wavelength and an additional point of attention increase uncertainty, forcing the brain to work harder and consume more energy.
The dose makes the poison, though. When a color accent is well-curated, it causes only a micro-deviation from our inner model. The model updates seamlessly, forming a richer, more nuanced internal map of our environment. Resolving this tiny pocket of uncertainty means the brain has learned, refined, or confirmed something. Turns out fifty shades of gray sound better than they look.
Evolution wired this process of gradual learning to feel good — because organisms whose models could update quickly and effortlessly adapted faster to environmental changes, securing higher chances of survival.
This logic extends far beyond color. Each design characteristic activates a different class of neural prediction models: color engages wavelength expectations, proportion engages spatial expectations, shape engages geometric expectations. When all these elements unite seamlessly, we achieve peak cognitive efficiency. The mind expends less metabolic energy updating its predictions, and we experience that sudden reduction in processing friction as a profound sense of aesthetic resonance.
Harmony is a fine play between repetition and variation. The brain craves variation because it expands its knowledge, yet it relies on repetition because it can process incoming signals more efficiently. Too much error overwhelms the system; too little leaves it stagnant. Harmony lives in the narrow band where prediction is mostly confirmed, gently challenged, and effortlessly resolved.
Immanuel Kant, in his work Critique of Judgment, noted something curious: when we find something truly beautiful, we expect others to agree. Where a subjective feeling gets its claim to universality long remained a mystery.
Today the mechanism behind his words might become clearer. Our drive toward harmony is not an eccentric psychological preference but a common property of the human brain’s architecture, shaped by evolution.
The rest of the essay appears in December.
From harmony to the emergence of complexity, the cost of oblivion and the golden rule: the full text of The Binding Force will be published in the DAV Journal 04/2026. Subscribe to get it the day it is out.