Konrad Banachewicz’s interview

There’s a certain elegance and depth to Konrad Banachewicz’s work that sets it apart. With a keen eye for style, fashion, and the female form, he approaches his art with both technical precision and a deep appreciation for real beauty—one that is often overlooked in AI-generated imagery. While many digital artists gravitate toward flawless youth, Konrad celebrates maturity, imperfection, and the grace of aging, portraying women with wrinkles, silver hair, and the marks of time—all captured with the richness of a set photographer’s lens.

His influences are clear. Art Deco, Brutalism, and Bauhaus often shape his compositions, lending them a modernist structure that plays with contrast, form, and negative space. Whether working in stylized photography or illustration, he is meticulous in his use of greens, blacks, and reds, creating bold yet sophisticated palettes. His subjects are often seen in masks, hats, or blindfolds, adding an element of mystery, while others appear naked, smoking, or simply lying in bed, evoking an intimate, cinematic feel.

Beyond his artistic pursuits, Konrad is also an experimenter—constantly playing with prompts, testing boundaries, and sharing his findings. His curiosity extends beyond visual art; he’s even prompted Shakespearean texts (a genius idea), showing a playful yet intellectual approach to his craft. His work exudes a European sensibility (something I relate to), a certain quiet sophistication in how he frames his subjects, capturing not just their appearance, but their essence.

But art isn’t his only world. As a co-author of books on data analysis and machine learning, Konrad’s expertise bridges both creativity and technology. He also runs a Substack and weekly podcast, "This Week in Tech," where he distills the latest AI and tech news for busy minds, offering insight into the rapid evolution of the digital landscape.

From AI-generated elegance to analytical depth, Konrad Banachewicz moves fluidly between art, thought, and innovation, and it’s an honor to feature him in this interview.

Let’s discover Konrad Banachewicz together.

Without getting too personal, can you tell us a bit about yourself?

I am Konrad, I make a living crunching numbers (PhD in stats, data science -> ML -> AI). My favorite book is "Catch 22", band - Iron Maiden, color - green. I'm not my own favorite subject ;-)

Could you please tell us which country you live in?

I live in the Netherlands.

What led you to begin working with AI imagery?

Originally the first GAN models, and then - in modern era ;-) the release of Dall-E. I've always had a way with words, but drawing has never been my strong suit. One thing leading to another, I started playing with Stable Diffusion, then Midjourney (v3 at the time), and my life has never been the same.

What AI tools do you use?

Midjourney, Flux, Stable Diffusion for images, Kling / Hailuo for video. I am starting to experiment with Suno for music.

What inspires you?

Quite a few things: definitely movies are an important source (I made a whole series lately riffing off the image of the great late Gena Rowlands in "Gloria"), as well as music I listen to at the moment.

Literature classics are an amazing source: Shakespeare practically wrote movie screenplays before cinema was invented, and in the AI context, his language is a prompt ideas goldmine. Play around with the "Tomorrow and tomorrow and tomorrow" speech from "Macbeth," and you'll see what I mean.

Medusa

(One of @revelinai’s favorite images by Konrad Banachewicz)

In which other medium, if any, do you practice art?

Do martial arts qualify? I used to do karate, but these days it's Krav Maga - which, while related, is way too practical to be considered "art".

Would you consider AI-generated art true art?

I think so. Yes, it does build on things done before - but it's not exactly a new phenomenon. Shakespeare took from Greek classics; Jimmy Page borrowed rather freely from Robert Johnson; you can hear Dvorak and Holst in the soundtracks of Zimmer and Williams, the list goes on... Yet all those people are - rightfully - recognized as having delivered something new to the world. So yeah, I think AI-generated art definitely *can* be art.

Please share one or a few of your favorite images with us.

Here are seven of my favorite images:

Black and white art

Black

Warrior art

Angel art

Do you title your AI-generated art? If you do, what inspires you to come up with these titles?

Not as a rule - every once in a while I have a clear idea how to title something, but it doesn't happen very often.

When do you tend to be the most productive, and do you work in long sessions or short bursts?

Evenings definitely - I've been an owl for as long as I can remember. I don't work in sessions, it's more like a burst of an idea, and then I stay on it until I like the result.

What type of prompts do you prefer: text or AI-generated descriptions through fed images?

I oscillate between both, along with quotes from plays or lyrics - I frequently start with one, make a first pass, then try to get a description of this image, and use the result to improve on my original prompt.

When prompting with text, do you write simple text-based prompts or complex ones?

Both - simpler ones are fun with personalized styles in Midjourney, but with tools like Flux (that can handle longer text thanks to their architecture), it's fun to let the words flow sometimes.

Do you think text-based prompts should be shared within the AI art community?

Absolutely: I learned a lot from others sharing theirs when I was starting out, so I try to pay it forward.

What is the most unconventional method you have used to create an image?

I'm not sure it qualifies as "unconventional," but a not uncommon workflow of mine goes like this: take a real picture (from my phone), get its description with a Vision Language Model (like Llava or Idefics), use this as a first iteration of a prompt, improve on the prompt (text-to-text) with Claude or ChatGPT - e.g., to sound in a certain style.

How many images have you generated using AI technology?

Over 15 thousand.

Please share your X (Twitter) handle with us.

@UniBananStates

Tell us a bit about your X handle and X username.

I've had a nickname "Banan" (Polish for "banana") since kindergarten, and the phrase "United States of Banan" occurred to me years before I had any idea where to use it.

How many X followers do you currently have?

1,714

How many X accounts do you currently follow?

859

How long have you been on X (Twitter)?

Since June 2023.

Tattoo

Is there any other online platform where we can view all of your previous work apart from X?

Not at the moment, although getting stuff to Wirestock is on my to-do list.

In this world of cryptocurrencies and non-fungible tokens (NFTs), are you currently selling your digital art on any platform? If so, which platform are you using?

I'm not into NFT, so that's a no.

Are you selling tangible AI-generated artwork, such as high-quality prints like Giclée prints?

Not at the moment, no.

Could you suggest some AI artists you like that we could follow?

Is there anything else you would like to add or share with the AI community?

X is a crazy place, especially with the algo changes - keep the faith, everyone, and remember: we are not the first ones to find ourselves in such a predicament.

“It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of light, it was the season of darkness, it was the spring of hope, it was the winter of despair.”