Prompt to Image: Book Reflections in Generative AI

The Extinction of Experience by Christine Rosen

Chasing Images in the Midnight Glow

From 2022 to 2024, I used MidJourney to make visual reviews whenever I read a book. At first, it was free, but around spring 2023, the free trial got super limited. I didn’t mind paying though, because it let me experiment a lot.

Insta-Brain by Anders Hansen / Prompt: primitive hunter-gatherer man, wearing fur and leather, braided hair, tribal look, standing, holding smartphone, meadow, cinematic lighting, hyper-realistic, surreal, UHD, movie photo, ultra wide angle, field depth, ultra precision, crazy detail, complex detail, beautiful color grades, Unreal Engine 5, pictorial, 25mm lens, scene depth, DOF, tilt blur, shutter speed 1/1000, F/22, white balance, 32k, backlight, VR, rays, R, dynamic pose, photograph, volume, superfine, complex details, 8K, ambient closure, volume illumination, high contrast, HDR, 8kv5 --ar 5:3 --q 2

One image takes about a minute to generate, as known. But getting the exact concept and mood I wanted usually needed at least five tiny tweaks to the prompt. If that didn’t work, I had to rewrite everything and start over. I usually went through three to five rounds like this until I landed on the images I liked. Sometimes I even stayed up all night chasing one good piece.

If I didn’t care much, an image could show up in minutes. But if I kept tweaking and retrying, it easily stretched into hours.

The hardest part wasn’t the generation itself—it was writing the prompts. I searched Reddit for good ones, collected, modified, then tested again. Even after all that, the results still weren’t “there” yet. Oh my God. I had to bring the images into Photoshop for compositing or retouching. People say AI will take jobs from designers. I don’t think so. Getting exactly what I wanted was still a huge challenge.

The Catcher in the Rye by J.D.Salinger

Still, it was fun. A great way to learn. I could play with styles, colors, and compositions way more easily than before. Writing prompts itself became a kind of design practice—I had to keep spinning my brain. All those attempts and failures sharpened my visual thinking, I believe. And at the end, the choice, the polish, the finishing touch—that was still mine.

By 2024, things shifted. I started running prompts only after checking them with a language model. Images matched a little faster. I still retried a lot, but overall it became quicker and more accurate. Loved it!

Words Are My Matter by Ursula K. Le Guin / Prompt: woman sitting on the floor, top down view, books, papers, surrounded, open books, cup of tea, painting brush stroke, Bright mint green, coffee, acrylic painting, soft pastel, warm, Ultra Wide Angle, Field Depth, Ultra Precision, Crazy Detail, Complex Detail, Beautiful Color Grades, Unreal Engine 5, Pictorial, 25mm Lens, Scene Depth, DOF, Tilt Blur, Shutter Speed 1/1000, F/22, White Balance, 32k, Backlight, VR, Rays, R, dynamic pose, photograph, volume, superfine, ambient closure, Modern, harmonious colors, intense sunlight, wide open fields and flowers, thick brushstrokes, high contrast, HDR, 8kv5 --ar 5:3

From 2025, I’ve been using Sora for free. Prompts still decide the quality and speed, but I need fewer retries. Even with all these changes, one thing is clear: AI images can surprise you, but meaning and context are always up to the designer. The sense of style, control over mood, the decision to communicate—AI can’t replace that. It’s just a powerful tool. It helps us, extends our creative work.

Small Things like These by Claire Keegan / Prompt: deep teal lighting, dark hallway, dusty floor, beige baby shoes, Bright windows, strong contrast, dirty indoor, soft light from window, cinematic atmosphere, melancholic mood, realistic style, high detail, dramatic shadows, 25mm Lens, Scene Depth, DOF, Tilt Blur, Shutter Speed 1/1000, F/22, White Balance, 32k, Backlight, VR, Rays, R, dynamic pose, photograph, volume, superfine, complex details, 8K, superfine, ambient closure, high contrast, HDR, 8kv5 --ar 5:3

Stoner by John Edward Williams / Prompt: young man in classic suit, standing, old university building, serious face, cinematic lighting, academic atmosphere, subtle, literary mood, realistic style, surreal, UHD, movie photo, ultra wide angle, ultra precision, crazy detail, beautiful color grades, unreal engine 5, pictorial, 25mm lens, tilt blur, shutter speed 1/1000, f/22, white balance, 32k, backlight, VR, rays, dynamic pose, photograph, volume, superfine, 8K, ambient closure, volume illumination, high contrast, HDR --ar 5:3

The Keeper of the Camphor Tree by Keigo Higashino

The Unbearable Lightness of Being by Milan Kundera / Prompt: two lovers face to face, surrounded by swirling smoke, intimate, poetic, cinematic lighting, soft focus, surreal and emotional mood, mid-shot, highly detailed, dramatic contrast, UHD, Future, Movie, Photo, Ultra Wide Angle, Field Depth, Ultra Precision, Crazy Detail, Complex Detail, Beautiful Color Grades, Unreal Engine 5, Pictorial, 25mm Lens, Scene Depth, DOF, Tilt Blur, Shutter Speed 1/1000, F/22, White Balance, 32k, Backlight, VR, Rays, R, dynamic pose, photograph, volume, superfine, ambient closure, Modern, harmonious colors, intense sunlight, wide open fields and flowers, thick brushstrokes, high contrast, HDR, 8kv5 --ar 5:3

Momo by Michael Ende

Homo Deus by Yuval Noah Harari — Image 1

Homo Deus by Yuval Noah Harari — Image 2

Dopamine Nation: Finding Balance in the Age of Indulgence by Anna Lembke

Nomadland: Surviving America in the Twenty-First Century by Jessica Bruder

Morning and Evening by Jon Fosse

The Midnight Library by Matt Haig