I was totally ignorant towards AI. I did not intend to touch it with a 90 foot pole.
But... It's either you ride it, or get ridden !!
Plus I am also noticing that more and more tools are becoming free, but you pay for the AI part in the tool, which was trained on your work. So this became a mission. Get rid of Adobe, Microsoft and all other tools that chain my through subscriptions to their AI tools, let me use the part that's "analog" in those tools and build up my own AI infrastructure. Where I can render whatever I want and not get a group of diverse woman, when asking for a picture of the founding fathers.
The fact remains, that I think that AI is the Pandora's-Box this society, this version of apes walking this planet, should never ever have opened.
So, I dove into AI for the past few months.
My most important guides were:
- Run local
- Run uncensored
Find a tool that let's me investigate, research and play
So I decided to work with ComfyUI, because it seemed to me that this is the tool that gives me most freedom.
It runs locally, it is flexible and let's me build the workflows myself and there's a large community that created a lot of tutorials for it.
As a developer at heart, off course I got to the point where I needed my own custom nodes quickly.
What if I created tools for AI using AI? So I told an LLM what I need and guided it to the point where the node did what I expected and worked stable. I am at the very beginning of my AI journey. While image creation is the driver behind this, there's the professional in me that want's something I can work with.
Here's some nodes I released:
- A node pack that helps with prompting / Video
- A node that let's one find nodes in a workflow
- A node that enables html inside a node (inline documentation and UI)
Stay tuned for more postings from me about this. While the first reason for me posting is for myself, so I can go back and find out what I was doing, secondly, maybe someone else is on a similar journey and we can share our experience !
Peace !!
