Do objects in your home or from your childhood remind you of other figures, animals, or natural phenomena? Here’s how Tim Yarzhombek uses visual association and textures to create fantastical sculptures on exactly.ai

1. When I was a child, objects would often remind me of other things: car headlights that resembled a face; or twigs I’d play with as though they were guns.

I often recall these visual associations when creating illustrations. Sometimes, I apply these in exactly.ai to generate completely new images. Today, I’ll show you several models where I used these associations as a method.

2. At the beginning, I don’t usually start with a specific association. I choose an interesting object, create a model, and start experimenting.

Here's a model made out of broken pencils. There's drama in how the pencil material is chiselled, and an interesting texture in the multiple layers of paint, wood, and graphite (as we explored previously, exactly.ai loves texture). After testing different sculptures, I thought of using prompts of cone-shaped objects. This led to the creation of magnificent birds and astonishing mountains.

3. In the next example, the association appeared immediately. Without any particular goal, I trained a model based on photos of cabbages.

One of the first prompts was the single word, "skull". The result was terrifying and impressive because both skulls and cabbage are organic. If you like horror, neural networks excel at this! Given the backdrop of current wars – where lives are treated as expendable – this result was especially poignant.

4. My next model was inspired by a meme drawing from the artist Borya-Spec. The meme depicts how, with a few simple tools, a loaf of bread can be turned into a trolleybus. However, the caption humorously asks, "But why?"

These images say it all:

5. Next, I decided to work with a local Moscow phenomenon – poplar cotton. Every June, the city is filled with clouds of fluff, made of seed hairs coming from poplar trees.

I created a model and tried to generate some ideas, but none of them could be realised. A year later, I thought, ‘cats are also fluffy’ – so I revisited the idea. The result was crazy and, because the model was based on photos from Moscow courtyards, felt like it was happening just around the corner.

6. Lastly, a model based on brooms, another object from my childhood. I really like the formation of tied material.

Initially I created animals. The results were interesting but lacked a metaphor. When I asked it to generate a woman, exactly.ai imagined a solution where the skirt turned out to be shaped like a broom. This immediately added some depth to the idea, touching on gender roles and the hardships of women's lives in domestic spaces.