Imagengine

Description

In December 2022, at the start of my first year of high school, I discovered a Discord server called MidJourney, still little known, which offered something very innovative for a time when AI was not yet a mainstream topic, apart from ChatGPT, which was only in its infancy, with plenty of mistakes that, incidentally, people made fun of. This server offered, through a subscription system, the ability to generate images with AI in chat channels. At first, I ran a few tests using the option to generate a few images for free, and I realized there was a lot of potential and that we were at a turning point where AI was about to become something more striking for the general public, because it would be more visual. People would then have a much stronger impression of understanding and seeing concrete applications of AI, since until then they had only seen a robot answering our messages in a strange way in a text interface. Now, you could generate images of things they knew and had asked for. It was not yet very realistic nor flawless, but we had passed the point where the image became understandable and of rather good quality. Of course, at that time, when you looked closely at the details, you could constantly and easily find flaws, but the image was still of good quality every time. But at that point, it was still a bit archaic: to generate images, you had to describe what you wanted to see with keywords separated by commas. Incidentally, as I write these words (in September 2025), a good share of MidJourney users still write their prompts with on-the-fly keywords. We would therefore sometimes end up with endless prompts, giving details through a chain of keywords. Example: Spiderfate, superhero, arcane magic, neon, ultra realistic, incredibly detailed, HD, photorealistic, cinematic, color grading, ultra-wide angle, depth of field, hyper-detailed, beautifully colored, intricate detail, editorial photography, tilt blur, white balance, super resolution, extremely realistic, insane detail, megapixel, ProPhoto RGB, VR, backlighting, natural lighting, incandescent, fiber optic, mood lighting, cinematic lighting, studio lighting, soft lighting, volumetric, beautiful lighting, accent lighting, global illumination, global illumination by ray tracing, optics, dispersion, brightness, shadows, roughness, ray tracing reflections, lumen reflections, reflections in screen space, diffraction gradation, chromatic aberration, GB displacement, scan lines, ray tracing, ray tracing ambient occlusion, anti-aliasing, FKAA, TXAA, RTX, SSAO, shaders, OpenGL-shaders, GLSL-shaders, post-processing, post-production, cel shading, tone mapping, CGI, VFX, SFX, elegant, hyper realistic, arcane magic, 8k, --v 4 --q 2 After 6 months of learning best practices, I mistakenly tried to monetize my knowledge by looking for people who wanted to generate specific images with AI. But of course, I was far too early for the time, and clients did not exactly rush through the door 😂. You can find my best creations below: