AI Validates Nietzsche

“Who’d a thunk it?” (Hairspray) One of Nietzsche’s most controversial propositions has been validated by recent developments in the science of Artificial Intelligence.

AI has graduated from Algorithm School and moved on to the Academy of Creative Design, finding novel solutions to novel problems. But this raises a more fundamental issue:

True creativity cannot be reduced to the mere reshuffling of defined variables, the rearrangement of words on a page, for example. I arrange words on a page. So did Shakeseare. See the problem. To be useful, AI needs to be able to distinguish the melodious prose of David Cowles from the tortured poetry of William Shakespeare. You can do it…but can AI?

To be clear, it is not just a matter of noticing a difference between Shakespearean poetry and my prose; it is a matter of assigning a ‘quality assessment’ to each. Cowles – A (of course); Shakespeare – C- (keep trying).

A great grandchild finger paints. “Is this good, Boka?” Innately, she understands that all ‘arrangements’ (if I may use that word for it) of paint on paper are not of equal value. 

She knows she likes Sally’s painting more than Josh’s. But why? And what about her own painting? Is it as good as Sally’s? As bad as Josh’s? 

To confuse matters further, Josh says he really likes her painting while Sally just said, “Mine’s better!” But is it? 

Josh, Sally and my great grand all have an innate sense of Beauty. How they apply that sense IRL to actual paintings is another matter entirely. The concept of Beauty is universal and innate; the appreciation of Beauty is the work of an individual’s lifetime.

(Sidebar: According to the Baltimore Catechism, RCC, c. 1955, our existential purpose is to know, love, and serve God.  We know God by pursuing Truth, we love God by curating Beauty, we serve God by behaving justly and demanding Justice for all. These values are universal and felt innately by each of us.)

But where does this innate sense of Beauty come from? Is it a function of evolution? Is it a product of human physiology? Is it an electrochemical phenomenon? Here’s where Nietzsche weighs in:

“…No one gives a human being his qualities: not God, not society, not his parents or ancestors, not he himself…The fatality of his nature cannot be disentangled from the fatality of all that which has been and will be…

“One belongs to the whole, one is in the whole – there exists nothing which could judge, measure, compare, condemn our being, for that would be to judge, measure, compare, condemn the whole… But nothing exists apart from the whole!” (Twilight of the Idols) 

The fundamental unit (quantum) of Being is the Whole. The Whole is not an assemblage of Parts; there are no ‘parts’ per se. What we call ‘parts’ are just aspects of the Whole. In fact, “Nothing exists apart from the whole.” 

Therefore, there is nothing in position to judge the whole. Judgment implies a transcendent perspective and according to Nietzsche, there is no such thing as Transcendence. Everything is immanent, the world is ‘ontologically flat’, so nothing is objective and there can be no valid judgements.

Of course, no one wants to be told that their ‘Dear Leader’ is naked. Even today, perhaps especially today, people (e.g. A. J. Ayer and Albert Camus, infrequent bed fellows as they may be) argue that objective judgement does not require a transcendent POV. Their bumper sticker of choice reads, “Good without God.” 

So who’s right? Nietzsche or the Secular Humanists?  

A recent article by Rohit Kumar Thakur (7/27/2025), It’s Game Over,  settles the score: Nietzsche – 1, Humanists – 0. First, Thakur asks the question, as old as Plato and as fresh as my great grand, “But How Does It Know What’s ‘Good’?”

He elaborates: “How do you judge art? How do you judge creativity? ASI-ARCH’s creators knew that just chasing a high score on a benchmark would be a disaster. That leads to ‘reward hacking’, finding cheap tricks to boost a number without creating a genuinely better design.”

If evaluative standards were hardwired, our non-linear creative impulses would be straightjacketed by arbitrary, linear ‘objective criteria’. Our so-called ‘evaluations’ would be nothing more than our subjective biases projected onto the material world.

“So, they built a ‘Fitness Function’…They use a separate LLM to act as an expert judge. This ‘LLM-as-Judge’ looks at the new design and scores it on things like innovation, structural complexity, and elegance… So, an AI is literally judging another AI’s creativity.”

But does that really solve the problem? Relatively, yes; absolutely, no! One ‘Fitness Function’ can be better than another where ‘better’ means ‘more independent’ vis a vis the system it is set to evaluate.

In an Ideal Universe, the Evaluative Function would be 100% divorced from its subject; it would be truly transcendent. In a less perfect world, the Evaluation Function can be designed with an eye toward ‘relative’, albeit not ‘absolute’, autonomy.

The further an Evaluative Function is removed from its subject, the more its judgements are free to identify true Beauty – and the less they are constrained by the opinions, tastes and biases of its engineers. The probability of any event occurring exactly as it does is miniscule, perhaps even infinitesimal, ε. 

Now build a second system just as fine tuned as the first and use that second system to evaluate the first. The probability of any one event-evaluation pairing is infinitely less than the infinitesimal probability of the event itself, i.e.  ε². If we’re not there yet, we’re getting awfully close. 

It is frustrating to live anywhere other than in an Ideal World. It seems like we can approach anything but reach nothing. Take Absolute Zero for instance (0° Kelvin). We can get remarkably close, but we’ll never reach it.

But in one respect, we do live in an Ideal World. Beauty, Truth, and Justice are Eternal Values that precede (logically and ontologically) the material world. Every Actual Entity (event) reflects in some way and to some degree these Eternal Values.

Best of all, these Values are entirely Transcendent. There is no trace of personal bias. Only Beauty evaluates ‘beauty’, only Truth evaluates ‘truth’, and only Justice evaluates ‘justice’. Logically (not ontologically or temporally), the Eternal Values even ‘precede’ God. 

The Eternal Values absolutely transcend the material world. They constitute the Essence of God who by definition is perfectly and totally Transcendent. A transcendent God is the sole guarantee that the Eternal Values are objectively normative, that they qualify as a Categorical Imperative. 

Without transcendent values, judgements can only be ‘relatively valid’ and ‘provisionally normative’. Ayer, for example, identifies ‘kindness’ as his core value. He’s certainly entitled to live his life according to that value. But attractive as it might be to us, it is only Ayer’s opinion. I doubt that Nietzsche, for example, would agree. 

Nor does Ayer’s designation of ‘kindness’ as ‘core’ entitle him to prosecute someone alleged to be ‘serially unkind’. Bottom line: Ayer and Camus shop for values the way I shop for flavors at Baskin Robbins…and according to the law in Massachusetts at least, I am not entitled to impose my flavor choices on you. Bummer!

The rapidly maturing science of AI clearly supports Nietzsche’s intuition that Judgement is valid only to the extent that it is independent of the matter being judged. Complete independence requires a transcendent POV – something not allowed in Nietzsche’s cosmology but hard wired into most ‘post-pagan’ theologies.

So, who’d a thunk it indeed? Nihilist Nietzsche joining forces with Orthodox Theists (like Pope Leo XIII) and 21st century computer scientists to combat the ‘humanist heresies’ of the 20th century.  Ain’t metaphysics marvelous!

***

Artist Gustave Courbet, The Painter’s Studio, 1855, oil on canvas, 361 × 598 cm, Musée d’Orsay, Paris. This painting mirrors Nietzsche’s idea that judgments of creativity are never absolute but always shaped by perspective, context, and the interplay between individual and whole—just as AI’s evaluations of art reflect both independence and relational dependence.