Applied AI | Ambitiously Solving Problems | Democratizing Creativity
Sacramento, CA
Joined July 2026
- Tweets189
- Following259
- Followers443
- Likes2K
Pinned Tweet
Fable 5.2 wrote the code that made every frame of this video. No video mode, no timeline.
Made with Video Decompiler, a free skill for Claude Code & Codex. Show it a video you love, and make it yours.
Open Source. Coming this week.
James Dimachkie retweeted
It's a good model!
Use it for your hardest challenges and planning to back up Opus 5.5.
James Dimachkie retweeted
Que video más hermoso de Mario. Lo muestra como un abuelo tras sus 40 años de historias y quienes fueron enemigos hoy lo ayudan a seguir adelante.
James Dimachkie retweeted
Beginning today, anyone arrested will be required to use Microsoft Teams instead of their one phone call.
James Dimachkie retweeted
Most people miss that Living Brain's 22.6x claim is against full context, not RAG.
I'd try it for shared project memory. I'd keep document retrieval.
I checked the numbers, the caveats and when I'd use each.
James Dimachkie retweeted
Episode 1 is out 😍
@brandon_galang and I on what it actually takes to move into applied AI:
→ 18 months to land an AI role
→ the "I'm not technical enough" fear
→ why AI job titles keep dying
→ vibe coding as a PM: treat the AI like your tech lead
So... I started a podcast.
It's called Builder Manifesto: The Podcast
Episode 1 drops tomorrow with @brandon_galang
Why?
The best builders I know in AI don't fit one title. PM, engineer, GTM. They're all of it, so let's talk about it.
If you're running a mac and are looking to up your agentic video editing, you should check this out.
Im tired of infinite canvas startups.
Im tired of generative video snake oil providers.
Im tired of hype for hyperframes, as html and playwright is a dumb long term solution for video rendering.
So I decided to build Raea. Open source. Free for everyone.
Faster and more capable than hyperframes or remotion. (10-20x faster with real footage and complex scenes).
And it renders without a browser….
Capable of more than just video. It’s built on @BevyEngine so it’s highly capable at producing 3D art, shaders and games.
It’s also built on Blitz Dom @dioxuslabs , and a bridge between the page and the 3d world behind it allows you and your agent to write everything in typescript and react… without a browser.
This video below was made entirely with Claude, Raea, and a weekend. I never had more fun playing and toying with creative work than I did making this.
But to be honest, that’s not where I’m stopping. I want Raea to be an engine that correctly empowers creatives and agents, while not cognitively offloading the parts that make your work unique.
I’m excited to continue working on this project as long as I can, and always open to feedback!
James Dimachkie retweeted
We open-sourced bouncy motion graphics templates.
Every loop is 30 lines of code, so your agent can recolour it for your brand.
scenario.com/explorations/mo…
> 40 shape animations, each a seamless 12-second loop: morphs, bounces, bursts, line draws
> click a loop and its code opens as one HTML file, ready to copy
> an agent skill for designing more in the same style
James Dimachkie retweeted
No need to learn Blender. We built a 3D engine for your agent.
With Tesseract, agents can now build, animate, and render 3D scenes, not by prompting a video model, but by controlling the actual geometry, materials, lighting, motion, and perspective of the scene itself.
Move the camera through the scene. Change the lighting and render it again. Choreograph the action, with finer control over rigged characters coming soon.
Every scene remains editable. Try it out with Astra or Opus 5.5.
James Dimachkie retweeted
Tencent just shipped the full Hy4. And for the first time, it can see.
770B MoE, 49B active, 1M context, plus computer use across the terminal and desktop apps.
Live on OpenRouter and Tencent Cloud today. Open weights? "Coming soon." I'll believe it when I can download them.
👋Hy4 is here, new with vision.
Firmly among the leading open models.
More productive everywhere: finance, frontend, game, officework, research…
Built responsibly, with cyber/biosecurity in mind.
Try it out and let us know!
🔗：hy.tencent.ai/research/hy4?l…
James Dimachkie retweeted
To every designer who feels a little terrified of AI agents: Jakub Otcenas from Diorama wrote this one for you.
He shows how he builds digital products with @paper and Claude Code, from an open source design system to agents that know his taste.
Thank you, Jakub, for opening up your whole process: tympanus.net/codrops/?p=1244…
James Dimachkie retweeted
Most people never touch Claude Code's effort setting.
I switch it constantly. Low to build, high to verify.
Here's when each level is worth it, with Anthropic's own numbers.
Thank you Brandon!
I was glad to help give it just a little extra kick. For everyone else, please expect much more from me in creative and other spaces.
btw if you liked my little animation, I made it using my friend @Jimosax's legible film skill! this was a pretty simple video, its excellent for far more complicated work. check his pinned post.
if you're looking to get into agentic video editing, give him a follow! he's doing some of the best work in that domain I've seen.
2
6
714
Legible Film 2.1 release
Here is my Agent Skill called Legible Film. It's a free open source skill that allows Claude Code and Codex to create and modify films. Version 2.1 was released today.
npx skills add JamesDimachkie/legible-film
github.com/JamesDimachkie/le…
Have fun with it today!
2
23
755
Pepper & Carrot: Episode 6, The Potion Contest. Animation: Morevna Project; illustrations and story: David Revoy; music: Grégoire Lourme. CC BY-SA 4.0: creativecommons.org/licenses…
The excerpts here are edited and captioned, and shared under the same licence. morevnaproject.org/pepper-an…
1
70