Machine learning puts your dicks in a row — but is it art?

Creative Coding Amsterdam co-founder Sabrina Verhage made an extraordinary visualization of 1372 vibrators scraped from sex toy webshops. The image was put together using ofxTSNE, ofxCcv and ofxAssignment; machine-learning algorithms designed to identify faces.

The vibrators were identified and then arranged according to t-distributed Stochastic Neighboring Embedding, or tSNE. This is a complex method of approaching AI where machines learn to differentiate between different items within a category.

The rubber dicks are then visualized according to what makes them similar – and what makes them different. The result is an uncomfortably- beautiful rendering of vibrators that looks like a pornographer’s vision of the world’s most fucked-up beaded curtain.

The real magic is in the arrangement. Verhage’s method doesn’t rely on her to decide what order to place the devices in, the coding does it for her. The visualization is the result of her work as a programmer.

Tech sites are talking about machine-learning, but not many are taking sex tech seriously. Sabrina might be one of the few people who has more than just a curious interest in making technology that titillates. She first got the idea to spice up her work after attending Sex Tech Hack – a real event designed to get coders interested in technology that improves people’s sex lives.

Fine art featuring hundreds of tiny renderings of sex toys is in short supply these days. My hat goes off to Sabrina Verhage for pushing technology and art into satisfying new directions.