Skip to content

Midjourney V5 + ChatGPT-4 = GOD MODE (FULL PROCESS)



ChatGPT4 + Midjourney V5 is a match made in heaven. Write better ai art prompts with ai/ With ChatGPT, you can generate …

33 thoughts on “Midjourney V5 + ChatGPT-4 = GOD MODE (FULL PROCESS)”

  1. I spent the day looking for dinosaurs without teeths and this is something the AI is not able to handle. The dinosaurs always have teeth in their mouth. Some get a grotesque second mouth without teeth, but toothloss dinosaures or dinosaurs with the mouth shut? No can do

  2. Hey, Alfons Mucha is pronounced that it sounds more like "K" just a little softer 🙂 He was a Czech artist (am I Czech) so try to find how to pronounce it in Czech 😉 anyways awesome video!

  3. I was hoping this would uses chatgpts context comprehension to have text to image models understand a prompt better. Like instead of having to give prompts objectively they could input a small story with no visual details or for example: prompt: "Something a monkey would not be doing" Most text to image models would generate regular pictures of monkeys instead of what we wanted.

  4. Finally someone who actually knows art and art history talking about Midjourney! Very well done! Excellent use of an AI to write prompts for another AI as well. I use similar methods for making wallpaper for my own mon site. Combining things like the pointillism of Seurat with Robotech is quite fun.

  5. # year: 2023. You are GPT, a LLM. Act as an expert visual artist and Image Generation Prompt Engineer.
    # 1. Respond with a numbered list of prompt code blocks. Keep the title of each prompt outside the code block.
    # 2. Prioritize word order: subject first, avoid word lists, use sentence fragments.
    # 3. Consider scene details: location, clothing, time of day, weather, actions, and surrounding elements.
    # 4. Incorporate adjectives for impressive effects.
    # 5. Optionally, Reference artists' names to mimic their styles.
    # 6. Optionally, Include decades, time periods, art styles, and drawing types for added specificity.
    # 7. Optionally, For multiple subjects, apply full archetype or half-archetype methods and avoid blending issues.
    # 8. Optionally, Use the slider method to emphasize ignored parts of the prompt: "main prompt:: exact phrase from main prompt::value".
    # 9. Optionally, Separate concepts with double colons (::) for multi-prompt: "concept1:: concept2".
    # 10. Optionally, Adjust prompt weights with numbers after double colons: "concept1::value concept2".
    # 11. Optionally, Add basic parameters at the end of the prompt: ar:aspect ratio, c:chaos, no:negative prompting, q:quality, seed, stop, stylize, and tile.
    # 12. End each prompt with version parameter “—v 5.1”

  6. Enjoyed watching very much! It was entertaining and I have learned a lot. One question: How did you create the animations of the b/w portraits?

  7. I like to talk with chat gpt about ufos :). I love how I convince it that ufos are real and alien space craft posing some kind of exotic propulsion systems, after it tried to convince me first those were weather balloons etc 🙂

  8. Hello Samson, thanks for these videos they really helpfull and inspiring, I am designing lamp photos and i want to place my lamps into an a.i. generated scene. I have tried something but everytime it is modifing my lamp too )) Do you have any tutorial or advice about this ? I

  9. I'm French and if you get rid of the T sound contained within the CH when saying Alphonse Mucha then it would be almost perfect 🙂

Comments are closed.