Kate Koidan

Researcher and writer focusing on #AI. Founder @ Kiki and Mozart (#AIart newsletter). Sharing #Midjourney prompts and tricks.

[Kate Koidan @keatena]

🚀 Thrilled to share that I co-authored the 2nd edition of an award-winning book "Applied AI: A Handbook for Business Leaders".✨ Grab a FREE Kindle copy in our limited promo! https://t.co/I5sbIwWvEr
#AppliedAI #AIInnovation #GenerativeAI

[Dan Hendrycks @hendrycks]

We’ve found as AIs get smarter, they develop their own coherent value systems.

For example they value lives in Pakistan > India > China > US

These are not just random biases, but internally consistent values that shape their behavior, with many implications for AI alignment. 🧵

[Kate Koidan @keatena]

"If a politician is pro-crypto, the key question to ask is: are they in it for the right reasons? ... Do they have a good positive vision, that goes beyond near-term concerns like "smash the bad other tribe"?"

A great post by Vitalik Buterin:
https://t.co/EZYEnnVcsB

[ElevenLabs @ElevenLabs]

We were blown away by the Sora announcement but felt it needed something...

What if you could describe a sound and generate it with AI?

[Alie Jules @aliejules]

Midjourney Office Hours - Jan 17, 2024

A quick recap.

v6
- zoom, pan, vary region (features we have in v5) coming as early as next week
- style consistency will come before character consistency, able to replicate styles (Yes! exciting)
- will move from beta to alpha
- more fine-tuning to the model
- will have a style tuner - not like in v5, and not the same as the style consistency

Website
- alpha site will be opened up to more people
- a lot of work is being done on explore page - it will be overhauled (yes!)

Video
- working on video, will come before 3D

v 7
- prompting accuracy will be much better in v 7

"Stylize is like how much it listens to you or does whatever it things is good."
David

Meaning: lower stylize listens to your prompt more closely, vs. higher stylize generates images that it (the model/AI) thinks looks good.

[Kate Koidan @keatena]

Here's the newsletter. It focuses on Midjourney, but also covers other AI image and video generators as they emerge.
https://t.co/CR6A66pBc3

[Kate Koidan @keatena]

This year I've started an AI art newsletter. It's not that big yet, but I like how it goes. Check out my beehiiv rewind!

[Kate Koidan @keatena]

5. Add additional keywords for a “natural” photo. These may include terms such as "natural," "everyday life," "authentic," and so on.

[Kate Koidan @keatena]

Are you struggling with creating lifelike people in Midjourney?

Here I'll provide a few tips with examples (prompts in ALT).

If you find these tips useful, you'll also like my comprehensive Midjourney guide.

Like & Retweet & Follow, and you'll get the link to my guide in DM.

[Kate Koidan @keatena]

4. Use the “selfie” keyword

[Barsee 🐶 @heyBarsee]

AI filmmaking is almost here.

I didn't think video AI would progress this fast in 2023.

Here are the top 10 incredible examples of AI video revolution:

(THREAD 🧵) 1/13

[Kate Koidan @keatena]

@The_DailyAi

[Kate Koidan @keatena]

Prompt of the day:
a photo of a young woman with undercut punk hair and a cute dragon in a surrealist style --ar 4:5 --w 50 --s 700

I especially like how "cute" the dragon is 😆

[Kate Koidan @keatena]

For a detailed blog post on multi prompts, check out: https://t.co/Dq4817LzTd

[Kate Koidan @keatena]

Today, I want to share some examples of how multi prompts work in Midjourney.

For those who are new to the concept, note that you can instruct Midjourney Bot to consider two or more separate concepts individually by using a double semicolon (::) as a separator.

👇👇👇

[Kate Koidan @keatena]

- Negative weights can be used to reduce the impact of certain elements or remove them entirely. Weighing part of a multi prompt to "-0.5” is the same as using a --no parameter with the corresponding word.
- The sum of all weights used in the prompt must be a positive number.