ChatGPT opens its API, Leonardo's New Image Prompt, MidJourney's V5, and Meta's Controversial Leak – Volume 2
All the AI news happening this week – plus I'm giving away my AI toolkit for free!
AI Highlights this Week
Meta’s LLaMa (Large Language Model Meta AI) model leaked on 4chan, resulting in debates about how GPT tech should be shared with the public.
Open AI released ChatGPT API using
gpt-3.5-turbo
and priced 10x cheaper than existing GPT 3.5 models.Open AI’s Whisper model is available as the large-v2 model through their API as whisper-1.
Salesforce and OpenAI announced the beta ChatGPT app for Slack to summarize conversations, point to resources, and speed communication with AI writing tools.
Midjourney announced v5 is coming soon for next-gen images, and paid subscribers can “rate” images to help the system learn preferred styles.
AI News from Twitter
1. Orbofi introduced AI factories to build and train game & media assets:
2. Leonardo adds the “Image Prompt” feature to combine multiple images for interesting results:
3. Scenario partners with The Sandbox to combine virtual gaming with an asset-creation platform:
Luminogram Technique
I’m always looking for new and interesting techniques for AI-generated images.
One of the best resources I’ve encountered for Midjourney is Andrei Kovalev’s Midjourney Style Guide. I appreciate reference sites that include images as examples of each style, technique, keyword, or parameter used in image prompts.
I noticed a technique called “Luminogram” and decided to try it with one of my favorite Midjourney prompt tricks of using emojis as prompts.
First results:
Prompt: “🐚 luminogram”
I posted one of the images above and asked people to try it and reply with their results. We got fabulous images and tips to try!
Rob Lennon explained how he likes to use “chaos” in his Midjourney prompts:
The Midjourney Documentation defines the “chaos” parameter as this:
The--chaos
or--c
parameter influences how varied the initial image grids are. High--chaos
values will produce more unusual and unexpected results and compositions. Lower--chaos
values have more reliable, repeatable results.
--chaos
accepts values 0–100.
The default--chaos
value is 0.
I added the “chaos” parameter to a prompt with 2 emojis and the “Luminogram” technique. The results are more varied, with “chaos” added in the image below.
Prompt: “🐚 🌊 luminogram --c 10”
More amazing luminogram images in the thread and the comments:
Another Community Favorite is the “infrared photography” style:
Get it free: The AI Visual Creator’s Toolkit
I’ve spent the last few weeks curating a database of visual AI tools anyone can use to level up content – and now I’m giving it to you free!
I’ve broken tools into these categories:
Visual AI Generation Tools
AI Image Generation Tools
AI Editing Tools
AI Animation Tools
AI Gaming Assets
Prompt Generation Tools
Styling Resources
Want access? Grab it here.
Thank you for reading, David! I'm glad you found it helpful.
I started writing online last year and it's still a new experience for me. I try to learn how to find creation tools and techniques for people with no experience , like I was. I have a lot more to show you!
Great post. As a newbie, it's really helpful to see the techniques explained!