San Francisco’s tech scene never fails to surprise, and the latest buzz is about a quirky new website that’s turning heads, and restaurant reviews, into an AI-powered “hotness” map.

Riley Walz, a local tech innovator, has created LooksMapping, a platform that uses artificial intelligence to rate the attractiveness of restaurant patrons based on their Google Maps profile pictures. The site, which covers San Francisco, Los Angeles, and New York, provides a color-coded map where red indicates “hot” clientele and blue suggests the opposite.

Walz is transparent about the project’s limitations, openly acknowledging the AI model’s potential biases and flaws. The methodology involves scraping millions of restaurant reviews and using an AI algorithm to assign “attractiveness” scores to reviewers’ profile pictures.

The results are as entertaining as they are questionable. Unexpected restaurants like Fondue Chinoise, Hot Pot Champ in Chinatown, and Ararat Kebab & Gyros in the Tenderloin top the “attractive patrons” list. Meanwhile, the iconic Zuni Cafe receives a modest 4.4 rating, while Son & Garden and Bansang boast high “hotness” scores of 8.8 and 9.1 respectively.

While the website might seem like a playful experiment, it raises interesting questions about AI, privacy, and the increasingly creative ways technology can be applied to everyday experiences. Walz, known for creating nostalgic and innovative web projects, seems more interested in exploring technological possibilities than making serious social statements.

This isn’t Walz’s first unique project, he’s also created sites like Bop Spotter, which continuously tracks music in the Mission District, and an archive of newspaper front pages that allows users to explore historical news.

As with many AI-driven projects, LooksMapping serves as a reminder of both the potential and the limitations of artificial intelligence in interpreting human characteristics. It’s a quintessential San Francisco tech experiment: part serious innovation, part playful exploration.