What Is Intelligence?

In What Is Intelligence?, Blaise Agüera y Arcas defines intelligence as ‘the ability to model, predict, and influence one’s future’ (10). His argument quickly leads to the assertion that intelligence is fundamentally computational because it is predictive (12).

I’m not going to try to argue for or against this definition. We are in the speculative region of AI research, which means that any answer to ‘What is intelligence?’ will be driven by the practical needs of engineering goals.

Instead, I want to focus on predicting and influencing the future. I’ll save modeling for another post because it deserves attention all on its own.

Any definition of intelligence that centers on prediction—which has been commonplace at least since the emergence of cybernetics—will find ‘error reduction’ as fundamental to the elaboration of intelligence. What is important about Agüera y Arcas’s addition of ‘influencing one’s future’ is that intelligence has a purposive orientation to the future, not just a predictive one.

Combining computation, prediction, and influencing the future leads to computational power becoming the motor of history. When we externalize that intelligence in tools and technologies that are designed to compute faster than humans can unaided by their technologies, the more powerful that motor becomes. The more we combine computation, prediction, and influence into devices designed to move faster than nature, the greater their speed and spread of human influence.

AI marks the historical transition in which our computational power seeks a level of influence that potentially moves faster than we can understand the feedback loops that are the excess of its outputs. Perhaps this is the source of our speculative pessimism or optimism? 

Will to Power and Self-Overcoming

In a crucial chapter of Zarathustra translated as ‘On Self-Overcoming’, Nietzsche traces exactly this movement that I am tracing—from defining ‘intelligence’ through an abstract ‘what is’ question to exposing its fundamental dynamic: the will to power.

Once we have defined intelligence as influencing the future, we are in the realm of Nietzsche’s thought.

Much ink has been spilled on what Nietzsche meant by the will to power, and it is as confusing as the doctrine of eternal recurrence. I have no aspiration to settle a century-long debate over the meaning of these key concepts for Nietzsche. ‘On Self-Overcoming’ read by itself yields some strikingly modern understandings of the relation between life, intelligence, and the assertion of purpose.

“Will to truth” you call that which drives you and makes you lustful, you wisest ones?

Will to thinkability of all being, that’s what I call your will! (II 12, Adrian Del Caro trans., Cambridge University Press, 88)

These opening sentences frame what is to come as an exposure of a barely hidden motivation. The will to truth is first a will that renders the world ‘thinkable’—i.e., the world is intelligible to our thought. Second, and this is crucial for Nietzsche and much of philosophical thought that follows from him, Western thinkability is trapped in the game of ‘being’ not ‘becoming’. Zarathustra’s phrase ‘Will to thinkability of all being’ encompasses that Western bias. Our will to make the world intelligible is the will to find being everywhere—our will to truth sees things rather than motions and relations.

You first want to make all things thinkable, because you doubt, with proper suspicion, whether it is even thinkable. (88)

Nietzsche emphasizes ‘make all things thinkable’ because that is what Western rationality does—it imposes thinkability by imposing a vision of the world as things that are accessible to thought. But that will to truth is actually a making, not just a transparent observation.

Thus begins the transition from will to truth to will to power. Zarathustra will pull us through the transition via the words make and create. The next sentence begins the payoff:

But for you it [the intelligible world] shall behave and bend! (88)

This is the particularly Western manifestation of the will to power—a will that cloaks its ability to assert purpose as a will to truth. To claim that the world is thinkable is to hide the fact that this will is actually trying to make the world ‘behave and bend’ to this other will. We don’t simply try to understand the world, just as AI researchers are not simply trying to abstractly answer ‘what is intelligence’.

We seek these understandings because we are trying to increase our influence.

So, when Agüera y Arcas defines intelligence as ‘the ability to model, predict, and influence one’s future’, he is exercising as much a will to power as a will to truth. To make this claim is not a merely abstract and innocent philosophical (or biological or physical) definition. It is part of the broader expansion of our will to power, our growing capacity to assert purpose and to make our thinkability capable of making things bend and behave as we wish.

Growth will, in fact, become a major concern for ‘On Self-Overcoming’; it will define life itself. But before Zarathustra gets there, he must speak of how this thinkability becomes creative.

Imitation and Creation

Zarathustra continues:

Thus your will wants it. It [the intelligible world] shall become smooth and subservient to the spirit, as its mirror and reflection. (88)

In order for the world to be intelligible and for our intelligibility to make it bend and behave to our purposes, there must be a symmetry between the world and our intelligence. This is the ‘mirror and reflection’ that Zarathustra asserts.

The direction of causation is mixed for Zarathustra. Which comes first, our intelligence or the world’s intelligibility? Is our intelligence the ‘mirror and reflection’ of the world, or vice versa? It is undecided in this passage, and it must be so. If the world were wholly Other than our thinkability of it, then the will to power would not work. Our intelligence would meet a hard boundary that would resist our influence. The world would very likely not bend and behave—it would not be ‘subservient to the spirit.’

The imitative relationship between intelligence and the world begins to emerge here. In the closing sentences of ‘On Self-Overcoming’, Zarathustra will define life itself as motivated by ‘esteem’ rather than mere self-preservation. Esteem, as René Girard will later point out, is a manifestation of mimetic desire—the emulation and imitation of that which we admire becomes crucial to understanding human motivation. For Girard, like Nietzsche, imitation must be assessed in its transformations. Imitation can be emulation or it can be envy that tips over into rivalry. Good and evil are not states of being; they are emergent phenomena that rise and fall out of the movement of imitation.

But let’s not get ahead of ourselves.

At this moment, Zarathustra introduces the key phrase, and the transition from will to truth to will to power is complete:

That is your entire will, you wisest ones, as a will to power; and even when you speak of good and evil and of valuations. (88)

The link between will to power and Nietzsche’s other long-standing concerns of good and evil and revaluation of values is now made explicit. We are tumbling through the history of Western consciousness at breakneck speed in ‘On Self-Overcoming’, and we should appreciate that speed.

You still want to create the world before which you could kneel: this is your ultimate hope and intoxication. (88)

The will to power, cloaked as a will to truth, is fundamentally creative. But creativity for Nietzsche is never neutral. It passes through values, which means it passes through assessments of good and evil—it creates good and evil in its motion. This passage, however, is not the knowledge of true values—good and evil. The passage itself is the result of creative action, which makes any passage an effect of the will to power.

But we should recognize that in this first movement of creation, the will to power creates a world to which we are subservient—‘before which you could kneel’. Our intelligence imagines the world as potentially hostile—one that we can only hope to make bend and behave by worshipping it as gods.

Here the reversal of subservience from the previous sentence should not be lost. For the world to become subservient to our spirit, we must first imagine ourselves subservient to it. We must grant it power over ourselves so that we can overcome that power.

Babylonian Interlude:

Let’s digress to Babylonian astronomy to make Zarathustra’s words less abstract. When the king’s astronomers looked to the sky, they saw omens from the gods. The world becomes intelligible first through positing omens, but omens only serve as signs of preventable events that the gods either foresee or ordain. Where would we draw the line between the will to truth and the will to power? It would be very blurry indeed. The moment of seeing the movement of the heavens as intelligible signs (omens) from gods who foresee the future (or actually create that future) is the moment in which our subservience opens the possibility of overcoming.

Over time, cataloging and interpreting omens will become the recognition of repetition. This repetition will be seen as calculable patterns, and the power of the gods will be greatly diminished when we learn to predict eclipses—which signal the possible death of kings—18 years in advance.

The movement here should be clear: our subservience to the gods is simply a precursor to learning how to overcome them. The will to truth that saw the movement of the heavens as omens eventually becomes a recognition of regularity and predictability. Running through this pursuit of truth is the will to power—by making the world intelligible, we gain better influence over causality.

What is life?

The will to power is at the heart of life. But the will to power is not a state of being—it is becoming. Life is not a static thing that can be reduced to a single definition. It is the experience of reversal: good becoming evil and back again, obeying becoming commanding and back again, the will to power becoming the will to truth and back again. Nothing remains still in Zarathustra’s worldview, and this impermanence is the root of life and living.

And this secret life itself spoke to me: “Behold,” it said, “I am that which must always overcome itself.”

Life is the constant action of overcoming. We are never done—time eternally recurs and with it the motions of life.

Yet we seek permanence because our intelligence seeks being as a resting place. We see things and beings where we should see the constant action of relations. We treat time as our exile from permanence and we wish ultimately to overcome it.

Zarathustra’s solution is not the decadent and weak embrace of impermanence, for that would be the parable of the tightrope walker from the Prologue whose willingness to hover between towers has weakened his initial embrace of danger.

No, Zarathustra’s solution is the eternal recurrence of the will to power, which is life itself. We embrace overcoming as itself a constant striving that seeks resistances, that seeks purposes, but must realize that any act of overcoming is itself the renewal of the need for new challenges.

That I must be struggle and becoming and purpose and the contradiction of purposes—alas, whoever guesses my will also guesses on what crooked paths I must walk. (89)

The pursuit of purpose as a permanent state of being will treat time as its burden. The will to power will treat time as that which must be permanently overcome. This way revenge and ressentiment lie. Its last and most futile expression came at the end of the last century when Fukuyama declared the End of History. We had finally reached ‘liberal democracy’ and the universal power of Hegelian ‘recognition’ as our End.

But this was folly, and ressentiment has followed. The ‘unwillingness toward time’ (112) has returned because it never left.

Emulation and Existence

Zarathustra next takes exception with definitions of life that reduce it to a mere ‘will to existence’. If it were only that, ressentiment and decadence will follow. Creativity would be drained of energy.

Indeed, the one who shot at truth with the words ‘will to existence’ did not hit it: this will—does not exist! (90)

Here he turns to ‘emulation’ as the driving force of the will to power and its creativity.

Much is esteemed more highly by life than life itself; yet out of esteeming itself speaks—the will to power!

At the heart of the will to power is esteeming and emulation and imitation. We seek to be that which we admire, but in this esteeming good and evil are emergent:

Truly, I say to you: good and evil that would be everlasting—there is no such thing! They must overcome themselves out of themselves again and again. (90)

How can good and evil not be everlasting if we must constantly overcome them? This seems contradictory until we realize that at the core of Nietzsche’s thought is how we experience time. By ‘everlasting’ Zarathustra means permanent and separate. Western morality has treated good and evil as Manichean—as separate states of being. Nietzsche’s Zarathustra wishes us to see everything in motion and thus impermanent. But impermanence does not mean weakness, it means that good can become evil just as commanding can become obeying and calls for equality can turn violent.

Thus the highest evil belongs to the highest goodness, but this is the creative one. — (90)

Life is the will to power, which is the eternally recurring call to assert purpose while realizing that the assertion of purpose too easily becomes vengeance when it seeks a non-crooked path and when it treats purpose as a striving for a permanent state—and End of History—outside of time itself.

Life and the will to power are creating a new passage with AI that we do well to attend to. Defining ‘what is life’ and ‘what is intelligence’ are crucial endeavors, but they are not innocent expressions of a will to truth. They are evolutionary movements of the human will to power expressed through an expanding intelligence that is making our power of compute more and more responsible for the movement of time itself.