How to Generate Deepfakes of Yourself In Midjourney to Make Anything Your Heart Can Imagine...Including Spicy YouTube Thumbnails!

And why I'm going to train either a Stable Diffusion model or a paid competitor and show you if that is a more effective method for tomorrow. Because yeah...these deep-fakes are pretty dang deep!

As long as you’re okay with the deepfake looking kind of, er, deepfakey…you’re golden! Note: I really did fight this bear, though.

Y’all.

The things I do for you.

I obsess over how I look. Hate looking at millions of pics of me. Hate it!

But to train Midjourney, it was necessary. So dive into self-study I did.

And while these deepfakes definitely look, um, fake, I think for the purposes of YouTube thumbnails, it’s about on par with the scroll-stopping bait-and-switches I’ve seen from the big YouTube guys who are for sure using AI in their image generation.

The impetus for this post? I had a reader write me to ask me how to generate an image of herself in a fun and novel scenario, and I thought: Oh sure, man, that’ll be no problemo.

Boy did I learn differently. Hours upon hours upon hours figuring out what works.

But! I’ve now got the key parameter you need to use to make it as realistic as possible and I know how to introduce weird scenarios like my buddy the bear up there.

Ready to enter uncanny “Yassification” valley with me?

If I can stand it, so can you!