After a Year
I never thought about AI before ChatGPT came out in 2022. Not once did I think, "Hey, you know what would be great? A digital assistant that helps me get things done fast, can do my taxes, can make Hollywood-quality movies for 8-second clips." But here we are—it's 2025 and it's all on the table now.
I'm not a doomsday scenario type of person. I kind of always pictured myself going out in an airplane, but that's a weird thing to talk about in therapy.
Believe it or not, in addition to struggling as a freelance artist this past year, I've been deep-diving into AI. I guess you could call it vibe coding, but reading dense academic papers on neuroscience to understand AI better doesn't strike me as particularly vibes-based. When I first started coding Python with AI, I thought, "Wow, this never works," and now it's so good that—well, if you're reading this, chances are you already know.
I've used Python to help me code a whole host of things, but mainly SideFX Houdini-related projects. This is, I guess, a blog post with mostly words, but sometimes you just need to explain yourself.
The past six months or so, I've been creating frameworks in and for AI. Started small with a random Twitter/X post. I have to thank Greg Brockman for the idea, if I'm being honest. Greg Brockman is the co-founder of OpenAI (heads up). Just to clarify, we don't know one another. Greg posted maybe six months ago this sketch of a framework concept, which ended up being so useful for AI research that I was using it like Old Bay seasoning—sprinkling it all over different research materials.
Then I started my own version of Brockman's framework, literally adding different leaders in AI and mapping where they focused their energies. Then I started trying to push those concepts further and further. The thing no one talks about with AI is that the vibe coding in 2025 is so good that you can build yourself into a lot of trouble. But the real skill is getting out of that trouble and making something worthwhile from the effort.
While I'm waiting for a patent on some of these concepts, I can't post the full framework here yet, but I can share some examples of what I've produced during my research time.
Before I go on about frameworks and AI, this definitely relates back to visual effects. In fact, my goal—or obsession lately—is trying to figure out a way for artists and students to quickly get up to speed on these technologies that aren't being taught at universities in 2025. It breaks my heart when my mentees ask about AI as if it's the first time they've thought about it, because in my mind, isn't that what they're paying for? Relevance in a super competitive industry?
I'm attaching a couple of samples * of what I've used it for so far. Please feel free to contact me with questions and comments.
*SAMPLES (Plans for Houdini/Unreal/Comfyui Tools)(WIP):
1. HDK tools for Houdini 20.5 lighting:
https://docs.google.com/document/d/1ICfxyZ9vz6ouAazx9LsLqJcDvFMeWfrJXzcNwz_nj4g/edit?usp=sharing
2. Houdini Fabric Generator for Sustainable Fabrics:
https://docs.google.com/document/d/1Cq0s2EhEjAu2oQ9XwEP2e3f1K6LswgZCGubP602lhlY/edit?usp=sharing
3. Comfyui Houdini Bridge:
https://docs.google.com/document/d/1DkkW-1AQBfuTpvNPMBbPYzXsKgIt1-mor70GjiezVXk/edit?usp=sharing
4. Advanced Cinematic Rig Camera for Unreal Engine:
https://docs.google.com/document/d/1H1wvJveBcTKZ83o1mczynfqVXPyKR1N4v4K9sFj5q5w/edit?usp=sharing
5. Bacteria Knitting HDK Solver:
https://docs.google.com/document/d/1wQpENP8skAJc6NmiD9fYu7CigHjHtKPuatso59K2cHo/edit?usp=sharing
