There was a time when I had a single goal: achieving perfect photorealism.
In 2004, I placed second at Miss Digital World. In 2005, I reached third place with a character I was deeply connected to: Patrizia.
She wasn’t just a model. She was an experiment. A pursuit.
A pursuit of what I used to call “the light of God.”
The impossible quest: simulating reality
Anyone who has truly worked in 3D knows this:
Modeling isn’t the hardest part.
Texturing isn’t either.
The real challenge is light.
Engines like V-Ray, Maxwell Render, and Octane made huge leaps forward. Techniques like Global Illumination and Light Dome brought us closer to reality. But here’s the uncomfortable truth:
real-world light is still out of reach.
In the physical world:
- light bounces infinitely
- every material reacts differently
- absorption, reflection, and refraction constantly interact
- the final result is complex, chaotic, and incredibly nuanced
In rendering, all of this is approximated, often through stochastic models-statistical estimations.
And yet, here’s the key point:
3D rendering remains fundamentally deterministic.
You are in control.
The breaking point: my AI experiment
A few days before Nvidia unveiled DLSS 5, I ran a test.
I took Patrizia.
Fed her image into an AI system and asked a simple thing:
“Make her real.”
The result?
Technically impressive.
A believable human. Almost photographic.
But…
empty.
The eyes weren’t the same.
The light told no story.
The emotion was gone.
That subtle magic-the feeling you could sense something behind the image-had simply vanished.
Nvidia DLSS 5: the same feeling at scale
When Nvidia introduced DLSS 5, I experienced déjà vu.
In “before vs after” comparisons-like those seen in titles such as Resident Evil-the technical leap is undeniable:
- sharper details
- cleaner output
- greater stability
But something changes.
And it’s not just technical.
Many players reported a similar feeling:
the image looks better, but feels less artistic.
DLSS 5 doesn’t just enhance images.
It interprets them.
And in doing so, it can:
- smooth out stylistic nuances
- normalize visual identity
- subtly override artistic intent
The result is a paradox:
more realism → less identity
Deterministic vs Probabilistic: the real divide
This is where the real fracture lies.
Traditional 3D Rendering
- deterministic
- controllable
- every choice is intentional
Generative AI
- probabilistic
- statistical
- converges toward learned patterns
And this changes everything.
Because art doesn’t live in averages.
It lives in deviations.
The uncomfortable question: did we really want photorealism?
For years, we chased one goal:
“Make it look real.”
Now we’re there.
And the real question becomes:
is that really what we wanted?
Because when:
- you can’t tell a render from a photo
- you can’t recognize an artist’s style
- you can’t feel the human touch
something is lost.
Something deeply human.
The hidden limitation: when perfection flattens everything
In games, art, and character design, diversity is everything:
- stylized
- cartoon
- realistic
- hyper-realistic
- exaggerated
This diversity is expression.
But when everything converges toward AI-assisted photorealism, one risk emerges:
homogenization.
And homogenization is the opposite of art.
What comes next?
This is where things get interesting.
What happens when AI:
- doesn’t just replicate
- but truly internalizes artistic styles
- develops something resembling a creative signature
What happens when it moves beyond statistics into something that feels like intent?
We might see something new:
- AI capable of generating genuine emotional impact
- systems that reinterpret reality instead of copying it
- a new hybrid form of creativity
And perhaps, at that point, the magic returns.
The light of God isn’t in the pixels
In the end, my “light of God” was never a technical problem.
It wasn’t Global Illumination.
It wasn’t ray tracing.
It wasn’t computing power.
It was-and still is-about intent.
Today, we can generate perfect images.
But perfection is not enough.
Because what moves us is not reality itself.
It’s the interpretation of reality.
And for now, that still belongs to us.
Alessandro Valori