Behind the Scenes - September Content Roadmap. (Patreon)
Content
Hello my Friends,
I wanted to sit down and tell you what I have been up to, and what is coming in the next days and weeks. You are the reason any of this happens, so you should hear it first.
I have to admit something: rebranding from AIKnowledge2Go to Into-the-latent was far more time consuming and demanding than I had guessed at the beginning. There were so many steps I did not foresee or anticipate.
For example, the website I created and the newsletter both have to be compliant with EU regulations. That alone can make a man crazy.
Another one: I had to re-record the Sage Attention and the ComfyUI base tutorial, so that from now on I can simply refer to them instead of explaining everything from scratch every single time.
But enough of that. What am I working on for you?
Public — YouTube
Cutting the *minimax h 3 video. Yes, still. Generating sample outputs that are not just okay, but instead make you go WOW*, takes a lot of time. That is exactly why I prefer to make videos on specific image or video techniques instead of introducing a whole new model.
Paid tier - only
1. Art Style LoRA training
This is something I have had in mind for a long time: a full tutorial video on training an art style.
We already did Character/Person, and Concept — like the Latex Dress. But training an art style is completely different from other LoRA training. In this specific tutorial I will teach you all the traps you can fall into, and how to get the best LoRA quality possible.
The LoRA Tutorial will be for Everyone but I have also content gated to your tier. For the full experience you will need to be tier AI Art Architect or above.
2. New Guides
I have 3 in the making:
1. A lot has changed since I wrote the original guide in 2023.
Prompting itself has actually become much easier with newer models — but that doesn't mean it has become completely self-explanatory. Quite the opposite. We now have different models with very different ways of interpreting prompts, different strengths, different quirks, and a lot more possibilities.
So I want to completely rethink the original guide and bring it up to date with everything I've learned since then.
This won't just be a quick update of the old guide. I want to make it a much better and more useful resource from the ground up.
2. A Quick Guide for better Photocompositions.
One of the biggest things I keep seeing in AI-generated images is that the individual elements can be fantastic, but the composition doesn't quite work.
And you don't need to become a professional photographer to improve that.
I want to put together a practical little guide that gives you some simple concepts and tricks you can immediately apply to your generations to make your images feel more intentional, more interesting and more like actual photography.
3. A Secret Guide. I'm not going to tell you what this one is about yet. 😏
3. A new installer… what, again?
Well… yes. 😂 And this part gets a little technical, so bear with me.
As you know, there are two things I currently have:
- Every paid tier gets my Easy Installer
- AI Art Collectors and up get Diffusion Nexus, aka Lora Helper, which helps you curate datasets for LoRA training, existing LoRAs, gallery images and much more.
When I started out, the project was one monolithic piece of software. One project. Bäm, done. For Diffusion Nexus I had to extract an SDK library that both can now use, so I don't have to implement model checks and installations twice.
I have always wanted to build software that can update itself. For the installer, that means it has to be able to deliver three kinds of updates: *Workloads (the installation logic), ComfyUI Workflows, and the software itself*. But the way everything runs currently, I cannot do this.
So that’s the pipeline. Plenty more on the way.
Thank you for sticking with me through the rebrand and everything that came with it.
Chris