Once a stopover for starving artists and gin-blossomed locals, Fanelli Cafe in Manhattan has become the choice meeting place for micro-celebrities brandishing hot dogs and Miista-heeled shoppers craving mid-spree martinis. In its 178 years on the corner of Mercer Street and Prince Street, the SoHo standby has never been hotter. The diners there, however, are more like a 4.1 out of 10.
At least, according to a GeoCities-esque website called LooksMapping.
LooksMapping is a digital heat map that claims to show “which restaurants have the most attractive diners — according to AI.” Visitors to the site can toggle among 9,800 restaurants in New York, Los Angeles and San Francisco to see how an artificial intelligence model rates the hotness of diners on a scale of 1 to 10.
“This website just puts reductive numbers on the superficial calculations we make every day,” its homepage reads. “A mirror held up to our collective vanity.”
LooksMapping was created by Riley Walz, a 23-year-old programmer in San Francisco with a penchant for using Google review data to make sardonic observations about the restaurant industry. Mr. Walz was also one of three people behind Mehran’s Steak House, a fake restaurant with a near-perfect Google rating that opened for one night in 2023.
To create LooksMapping, Mr. Walz left his laptop open over a weekend as an A.I. model scraped 2.8 million Google reviews. From 1.5 million unique accounts, it identified 587,000 profile images with distinguishable faces. Mr. Walz then prompted the model to extrapolate whether those pictured were young or old, male or female, and, to put it in vintage internet terms, hot or not.
The model was fairly accurate at detecting apparent age and gender, Mr. Walz said, as the A.I. gave estimates of age and gender on a probability between 0 (younger; male) to 1 (older; female). Enough photos were ambiguous that he opted to round up or down to save time, but a more thorough assessment could have treated values near the midpoint as indeterminate or representative of middle age or other gender identities.
The way it scored attractiveness was “admittedly a bit janky.” It favored seemingly arbitrary details to gauge hotness, like whether a profile image depicts a person wearing a wedding dress (hot), and if a photo is blurry (not). “The model isn’t just looking at the face,” Mr. Walz said. “It’s picking up on other visual cues, too.”
While LooksMapping is more cultural commentary than practical resource, its premise speaks to a growing trend of diners prioritizing a restaurant’s clientele over its food or atmosphere. For three years, the influencer Alyssa La Spisa has hosted a series called “Where Hot Guys Eat IRL” on her Instagram. “I decided to go straight to the source and ask the hot guys directly where they go!” she wrote in a caption.
The influencer Jordan Helms shares videos dedicated to “what all the hot girls in New York are currently obsessed with,” and recently featured the Italian restaurant Campagnola on the Upper East Side. And the dating influencer Tiff Baira often posts about where to meet bachelors, claiming in one video to have found “the best restaurant to meet rich eligible men in L.A.:” the old-school red sauce restaurant Dan Tana in West Hollywood.
Ratings on LooksMapping pay no credence to legacy. The Manhattan location of Hillstone, the low-lit chain popular for its French dip and charming suburban familiarity, scored a sensible 5.7, while the six businesses opened by the aesthetics-minded New York restaurateur Keith McNally averaged a 4.97, just shy of beautiful.
Proximity makes little difference: Kiki’s, the perpetually overflowing Greek restaurant in Manhattan’s Chinatown, scored an 8.2, but its neighbor Bacaro, a slightly more upscale Italian restaurant that attracts virtually the same clientele, scored a 3.7.
In Los Angeles, Gjelina, the sun-dappled hot-spot on a scene-y stretch of Abbot Kinney Boulevard, scored a 5.9. The Chick-fil-A about five minutes away narrowly beat it out with a 6.2.
The Berkeley-based food writer Soleil Ho instantly found flaws in LooksMapping’s rankings of restaurants across San Francisco. “The algorithm seems to have a thing for Asians, and a bias against places that are Black-owned and/or in Black neighborhoods, like the Bayview,” they said in email.
The project puts A.I.’s flaws, including its racial biases, front and center. When Mr. Walz first posted the project on X, 20,000 users visited the site on its first day. One of the first replies criticized the “red to blue gradient across the city,” as New York’s “hot” restaurants (marked by red pins) are mostly concentrated in largely white, affluent neighborhoods downtown, and businesses grow less attractive (marked by blue pins) as you move uptown and toward the Bronx.
Erwann Millon, a programmer who specializes in image-generation, said that racial biases in artificial intelligence are baked into the programming like original sin. “When you’re automating data collection at scale, the biases in your A.I. model tend to reflect the biases of the humans who created the data it was trained on,” he said.
“It’s making fun of A.I.,” said Mr. Walz, who has not eaten at any of the San Francisco restaurants rated for top hotness. “One of the ugliest restaurants is a country club.”
Ho also noted that in San Francisco, the top two “hot” spots (Ararat Kebab & Gyros in the Tenderloin, and Himalayan Cuisine SF in Pork Gulch) and the top two “not” places (Pizza Zone N Grill in Bayview and Mandarin House SF in Bernal Heights) are largely takeout joints. That proves the data is “not an accurate gauge of what the human customers in the dining room look like,” they said.
Whether by mistake or design, LooksMapping occasionally hits the nail on the head. Sissi Lu, a photographer, lives in the West Village of Manhattan and is a regular at the neighborhood Italian restaurant Malaparte. She came close to guessing the restaurant’s 6.6 rating, estimating a 7. “It’s not a particularly over-the-top place, but I see some really attractive people,” said Ms. Lu, 29, from her perch at the bar. “This place is a little tucked away, and it’s a very date-y spot.”
The SoHo outpost of Raku, a Japanese restaurant with three locations in Manhattan, was among the 47 awarded a 10 in New York. But even at the slightly lower-rated East Village location (9.5), a group of 20-something regulars slurping noodles seemed surprised to hear that the restaurant scored so highly.
“It’s a younger crowd,” said Sophie, 28, who asked that only her first name be used because of concerns about her work. The dining room is a cozy and minimal gathering place popular among the area’s college students. “It’s a very chic ambience, but I can’t say I’ve ever thought about the clientele like that.”
Back at Fanelli Cafe, a recent trip to more accurately gauge the 4-rated crowd found no off-duty models. But good manners (and good sense) would stop most onlookers short of calling anyone ugly.
“Actually, that score makes sense to me,” said Jasmine Baker, 24, a server there. “Eighty percent of our clientele are tourists, so they may not be ‘New York hot.’” She then deliberated with her colleagues at the host stand outside and returned to confirm what we all know: Beauty is in the eye of the beholder.
“OK, so we actually do think that’s low,” she said. “It should be over a 5, at least.”
Follow New York Times Cooking on Instagram, Facebook, YouTube, TikTok and Pinterest. Get regular updates from New York Times Cooking, with recipe suggestions, cooking tips and shopping advice.
The post The Map Rating Restaurants Based on How Hot the Customers Are appeared first on New York Times.
