Generative Workflow Interview - Kyle Hailey
Updated: Dec 12, 2022
Meet Kyle Hailey, a computer geek turned generative artist. With a background in fashion photography and a lifelong passion for art. Kyle was immediately captivated by the technology when he saw a friend's Facebook posts in the summer. He tried out Dall-e first, but it was MidJourney that truly hooked him and sparked his obsession with generative art.
"I think there's so much power in networkingand joining together"
In the beginning of December, Kyle organized a gallery during Art Basel, a prestigious art conference. The gallery showcased generative art made with MidJourney. He originally started posting his generative art on Instagram, which caught the attention of a friend who is the co-founder of a marketing company. This friend was looking to incorporate modern technology, including generative art, into their events at Art Basel. The friend's art department attempted to create generative art pieces, but the co-founder didn't exactly receive what he was looking for, which led to him reaching out to Kyle and to hire him instead. Kyle enlisted the help of other artists from all around the world, that he found through Instagram, to create a collective called Art Wave AI, and they collaborated on the project together.
Using Photographs as Image Prompts
When developing prompts, Kyle’s intentions are to keep things brief and to the point, considering the prompt a framework through which the art can develop. Comparing it to cooking, referencing his time working at an restaurant, it was reminiscent of how roux are made: “[this base sauce] didn't really taste like much of anything, but with that sauce you could add any spice and it would just make it sing”.
Similarly, his prompts provide a simple base that can be enhanced with additional text, style, or images, often using his own photography, to create unique and interesting art.
Once the prompt has produced its artistic results, Kyle’s goal is not to go back and tweak the prompt, but to explore the results until something catches his attention. This could be a connection between the colors, the energy of the picture, or a deeper subject matter within the photo. For example, in the case of the woman and the painting, the generation explored by the AI possessed a certain quality that added drama and suspense, making the viewer wonder about the woman’s thoughts.
“In general I like to keep the prompts as short as possible, because I like to be able to tweak it and I like to be able to understand what changes with the tweak. If the prompt is too long, it's hard to figure out what parts of the prompt are important”
This approach to exploring results aligns with Kyle’s philosophy on the use of photomanipulation tools by photographers. Often, the emotions and experiences captured by the photographer at the moment of taking a photo are not fully conveyed in the final image. With tools like Photoshop, photographers can emphasize or expose these experiences in their work. MidJourney, and generative art, presents itself as a new way of achieving his goals. Through his image prompts and the AI’s process, Kyle is able to explore these fleeting moments in greater depth and bring them to life.
Fashion Photography and Prompt Craft
Taking inspiration from the experience gained during his time in the fashion photography world, Kyle has developed an adaptable and reusable prompt template to produce captivating and varied images. The template consists of two main sections: a fixed section that provides direction for the composition and establishes the genre and style, and a variable section that allows for flexibility in the main subject, clothing, and overall aesthetic of the image.
The blue-highlighted sections function as largely unchanging brackets around the content-carrying prompt clips. The prompt begins by establishing a wide angle view, to ensure that the subject is captured in full. The “closing bracket” establishes the genre, and nudges the image towards a more consistent and repeatable style.
At the heart of the prompt, things are subject to change (highlighted in bold specifically). It starts by first establishing the main subject of the picture, choosing from a variety of known models as a guide. This is followed by a short description of the desired clothing, including the fashion designer it should be inspired by, providing guidance on the desired style and aesthetic. The last part is mainly used to help steer the AI towards the desired vibe, colors or textures. To come up with these changing variables, Kyle relies mainly on his knowledge of established figures he is familiar with in the fashion and art industry, though he sometimes gets inspired by browsing through the community tab on the MidJourney website to find new keywords for his prompts.
The prompt concludes with a series of phrases that further refine the physical appearance of the model, allowing for precise control over the final image. These phrases, highlighted in pink, are regularly updated to take advantage of the latest advancements in MidJourney algorithms and change in format and position. For example, “oversize body, full body” was required in the early stages of V4 to avoid overly skinny models, while this might lead to an overcorrection with newer versions. As a result, the template is able to adapt and produce high-quality images that are consistently visually striking.
Impact of Generative Art
Kyle concluded that he is certain that the meritocracy of art will endure because those with good artistic intentions will continue to use new tools to create great works of art. Those with good artistic intentions are still the ones who are able to create the best works of art, regardless of their technical skills or proficiency.
The democratization of the art production process through the use of new tools and technologies has made it possible for people with a wide range of technical skills to create art. This has not only expanded the pool of potential artists, but it has also made the art world more inclusive and accessible.
Previously, the ability to create art was largely limited to those who had the technical proficiency and training to use traditional art tools such as paintbrushes and pencils. However, with the advent of digital art tools and other technological innovations, anyone with a good idea and the motivation to create can now bring their artistic visions to life. The democratization of the art production process simply allows more people the opportunity to showcase their ideas and talents.
Unlocking the process of producing art has also had a significant therapeutic impact on the generative art community. The ability to create art can be a powerful form of self-expression, providing a way for people to explore their creativity and emotions. This can be particularly beneficial for those who may not have access to traditional forms of therapy or who find it difficult to express themselves in words.
Written by Tina Tiresome