PHOTOGRAPHY
The Most Attractive People Have the Worst Dating Profiles
I’ve run 200+ profiles through a 9-signal AI engine. The pattern stopped surprising me around profile 80.
My father was a photographer. He spent 40 years teaching people to see photographs the way strangers see them.
Not the way the subject sees themselves. Not the way the photographer sees them. The way a stranger sees them. In the first two seconds. Before they form an opinion. Before they know they’ve formed one.
He died in February 2025. Two weeks after I brought him home. That is a longer story. The short version is: the AI engine I built to read what photographs communicate — I built it because of him.
He gave me a Da Vinci notebook. He would have found what the engine found completely unsurprising.
Here is what it found.
The people who are objectively the most physically attractive — by conventional metrics, by symmetry score, by the standards any photographer would confirm — consistently produce dating profiles that perform below their potential. Sometimes far below.
Not because they are doing something obviously wrong. Because they are doing something very specifically wrong that their attractiveness makes harder to see.
I want to show you the Mona Lisa first.
I ran La Gioconda through the engine because it seemed like an obvious test. Composition score: 97. Mystery score: 98. Symmetry score: 4.1. Dating composite: mediocre.
Leonardo da Vinci spent four years on that face. He used sfumato — a technique that takes a lifetime to master — to soften the transitions between shadow and light. He painted her smile ambiguous on purpose. He understood that mystery holds the eye longer than clarity does.
She would underperform on Bumble.
This is not a flaw in the engine. This is the point of it.
Da Vinci was optimizing for gallery walls. For a viewer who stands still for minutes. For 500 years of contemplation. He was not building for a 4-inch screen, a 1.2-second attention window, and a thumb moving left because the bio said “I love to laugh.”
The signals that read as extraordinary in the Louvre are not the same signals that read as “swipe right” at 11pm on a Tuesday.
Same face. Different context. Different rules.
Conventionally attractive people have spent their lives being the Mona Lisa. Admired. Looked at. Appreciated from a slight distance. Dating apps don’t reward being looked at. They reward looking like someone you can talk to.
Those are different jobs.
The Vinci engine runs 9 AI agents over a photograph simultaneously. Each reads a different dimension of what the image communicates. Not how you look. What you project.
Warmth. Trust. Power. Mystery. Threat. Scroll-stop. Archetype. Composition. Texture.
High symmetry scores. High power scores. High composition scores. Low warmth. Low approachability. Elevated threat signal.
The last one is the one nobody expects.
Every human visual processing system runs a threat assessment on every face it encounters. It is not a choice. It happens before conscious thought, in the amygdala, in milliseconds. The output of that assessment influences whether the viewer feels comfortable proceeding.
Here is how attractive people accidentally activate it.
They try to look powerful. They look directly at the camera with intensity. They lower their chin slightly, which is a dominance signal. They do not soften this with warmth because they are focused on looking good, not on looking safe.
The result: a face that reads as dominant, symmetrical, conventionally attractive, and slightly unsafe to approach. On a dating app, slightly unsafe to approach is functionally the same as unattractive. The viewer swipes left and doesn’t know why. They call it “something about the eyes” or “just a vibe.”
It’s not a vibe. It’s a signal.
The problem has a name. I call it the passive signal trap.
Get Napoleon Beltran’s stories in your inbox
Join Medium for free to get updates from this writer.
When you are conventionally attractive, you learn early that your appearance does the work. People approach you. Conversations start. The photograph is enough, or close enough, that you never develop the habit of thinking about what else it needs to communicate.
So you post the photo where you look your best. Which is usually the photo where you are most aware of being photographed. Which is usually the photo that communicates: I am aware of being photographed.
That awareness is the trap.
The camera-conscious expression. The carefully chosen angle. The controlled environment. The lighting that flatters. These are all choices that optimize for one variable — how good you look — at the direct expense of three signals that matter more for dating than raw attractiveness does.
Warmth. Trust. Approachability.
A photograph optimized for “I look my best here” is a photograph optimized for a magazine cover, not for a conversation starter. Magazine covers are for looking at. Dating profiles are for deciding whether to reach out.
There are four places where the gap shows up in the data, and none of them require a new photo shoot. They require understanding what your existing photos are communicating that you didn’t know they were.
The first is what photographers call the Duchenne problem. A genuine smile activates two muscle groups. The zygomatic major pulls the mouth corners up. The orbicularis oculi raises the cheeks and creates crow’s feet around the eyes. A posed smile activates the first group only. The visual system knows the difference. It processes this in the same millisecond window as the threat assessment. Posed smiling reads as performing warmth rather than feeling it. In a dating context, performing warmth is almost worse than no warmth, because it reads as calculated.
Cover the lower half of your best profile photo. Look only at the eyes. Is there warmth there independent of the smile? If not, that photo is performing, not communicating.
The second is body orientation. Warmth communicates through the body before it communicates through expression. A body angled slightly toward the camera, or toward another person in the frame, reads as engaged and open. A body squared directly at the camera reads as presenting itself for inspection. People who are comfortable being looked at tend to present themselves for inspection. People who are comfortable with other people tend to angle toward them. The difference is subtle in the photograph. The signal it sends is not.
The third is the most important one, and the hardest to manufacture. The most attractive dating profile photo I have seen through the engine was not a professional headshot. It was a slightly blurry photo of a woman caught mid-laugh at something outside the frame. Symmetry score: average. Warmth score: 91. Scroll-stop: 88. She was not aware of the camera. And that unawareness, that proof that something more interesting was happening than being photographed, communicated more about what it would feel like to be around her than any posed photo could.
The best photographers know this. Pablo knew this. The photo that communicates presence is rarely the photo where the subject is most focused on the photograph.
The fourth is where most attractive people leave the most on the table, and it has nothing to do with the photo.
Because their photos do more work than average, they write less careful bios. Short. Minimal. “Looking for someone to adventure with.” “Lover of good food and bad puns.”
These bios are not bad. They are empty. And emptiness in a bio communicates something specific: I didn’t think the person reading this was worth real effort.
The engine scores bios on four dimensions. The single most predictive difference between high-performing and low-performing bios is what I call you-orientation. Does the bio speak toward the reader, or just describe the writer?
“I love hiking on weekends” describes the writer.
“If you know any trails that still feel like a secret, I want to hear about them” speaks toward the reader.
Same information. Different signal. Completely different response rate.
Attractive people rarely think they need to speak toward the reader. Their photo is already speaking. What they don’t realize is that the reader is running two separate assessments: the photo assessment (are they attractive) and the bio assessment (are they a person I want to talk to). Winning the first and losing the second is a very common outcome. It is also entirely fixable.
This is not a problem of attractiveness. It is a problem of feedback.
Conventionally attractive people receive positive feedback on their appearance from an early age. That feedback trains them to optimize for how they look. The optimization is rational given the feedback they receive.
Dating apps give different feedback. Not “you look good” but “did someone reach out.” The gap between those two signals is where the paradox lives.
The people who perform best on dating apps are not necessarily the most attractive. They are the people who have learned — through failure, through deliberate effort, or through someone telling them exactly what I am telling you now — that a dating profile is a communication event, not a photograph.
You are not trying to show someone how you look. You are trying to communicate what it would feel like to be in a room with you.
Those are different jobs.
Pablo spent 40 years teaching people this difference. He built a photography school around it. The AI engine I built to finish what he started is called Vinci, and it says the same thing he did, in a language made of numbers instead of film.
A photograph of a person is not a record of their appearance. It is a record of what they project.
That difference is where all of this lives.
The Vinci engine is live at vitruviano.app. You can run your own photo through the full 9-signal analysis and get a Date Ready Score for free. The paid report adds bio analysis, a breakdown of exactly where your signals are leaking, and specific edit suggestions. It costs $9. If you’ve already spent money on dating apps without knowing what your profile was actually communicating, it is probably the better investment.
Napoleon Beltran is the founder of Vitruviano / Vinci, an AI platform that reads what photographs communicate to strangers. He was a cruise ship photographer for four years across 100+ countries. He lives in Manila.
