22 Comments
Mar 14, 2023Liked by Linus Ekenstam

Hi,this article is really very helpful and u narrated so well.Thanks a lot.

Expand full comment
author

Thank you 🙏🏻 appreciate it

Expand full comment

Thanks for working through your process with us and sharing your discoveries. It's fascinating to see the techniques we use to act as memory analogues for tools like MJ that don't have a built-in dialectic process

Expand full comment
author

Hey, thank you, yeah I agree. It’s very interesting, I’m excited about the future

Expand full comment

I just recently started using things like 'remember this as list A' or 'use source B' to group data in ChatGPL conversations, and it's incredibly useful for getting more nuanced responses. Or crazy remixes.

Expand full comment
Feb 22, 2023Liked by Linus Ekenstam

Hi very great article...thanx. have you tested supermachine to generate images like this? I'll try your ideas there show up my experience soon.

Expand full comment
author

Hey Thomas, no I have not, sounds like I need to get on top of that. Let me know how it goes for you

Expand full comment
Feb 22, 2023Liked by Linus Ekenstam

Cool! So basically I can even use a AI generated avatar of myself or anyone else, upload and apply same technique?

Expand full comment
author

Yes, that should work just fine.

Expand full comment
Feb 22, 2023Liked by Linus Ekenstam

Great content!!

Expand full comment
author

Thank you Jaume 🙏🏻

Expand full comment

Thanks for the detailed guide. I have found another method to work, which is asking for one image split into multiple shots from different angles.

Expand full comment

This was so useful! thank you!

Expand full comment

Great tutorial! Should the reference image URLs be added to the prompts shown in Step 5?

Expand full comment

Hey Linus, I'm wondering if you would be interested in doing a deep dive on OpenAI's Sora for AI Supremacy? I'd super appreciate your perspective on this and where you see it going. Let me know if this sounds interesting to you.

Expand full comment

Hello - Just asking for some clarification. When you say to keep rolling to get results, does that include making variations? Because when I roll the prompt with the image and seed, it recreates the same 4 results each time. I have to click a V to get a different result. Thanks for this!

Expand full comment
author

It should def not create the exact same images if you are not including the --seed

If you have a seed in there yes, it will create somewhat the same image.

Expand full comment

Ok. My understanding in the procedure was to include it. It makes no changes whatsoever. I will attempt it again to see. Thanks

Expand full comment

Thank you for an insightful and practical approach to generating and rendering consistency in images. I have been taking courses in prompt engineering for MJ and have been a little stumped in this one area; however, your step-by-step, detailed approached gave me a better working knowledge which I have already begun to apply to my work.

Expand full comment

"Next, we want the seed number for this image, just respond with the letter emoji to the generation and you should get the seed number."

I don't really understand what's happening at this stage. The video blurs by so quickly. I clicked the emjoi and typed envelop. An envelop with a "1" appears. I clicked that ... then what happens?

Expand full comment

I suppose the problem is, I don't know where the larger images with the Job # and Seed # come from. They aren't popping up in a new window. They don't appear in the feed. And I'm not getting a direct message, and don't know where the direct message is sent.

Expand full comment

Good afternoon. Can you tell me, is it possible within this methodology to create sequential characters, but not people, but some kind of machinery, a tractor for example?

Expand full comment