Matt Zien (KNGMKR) retweeted
Wonder by Adobe.
Real-time interactive 3D world navigation.
- minute-long rollouts from one image
- I2V, V2V
- constant latency
- Wan2.1
You can re-shoot any video from new angles.
wonder-world-model.github.io…
Matt Zien (KNGMKR) retweeted
Introducing FLUX 3.
One multi-modal model for Image, Video, Audio and Action-Prediction. Creations are truer to life in every kind of style.
FLUX 3 Video is now available in early access (link below).
Jointly trained in one unified architecture, our model can be extended to predict actions for robotics. See our work with mimic and Audi in the thread.
Matt Zien (KNGMKR) retweeted
Session Feature ✧ Swimming With the Slop
Matt Zien's first AI film didn't go viral, it got him death threats instead. Since then, the cultural impact of AI has only become more polarized, especially in the space of arts and entertainment.
In a live taping of @BW's Everybody's Business, @chafkin interviews @KNGMKRlabs on his 15-year TV journey that led to Midjourney, what the early reaction taught him about the field he stepped into, and the challenge facing creators who are simultaneously gauging audience appetite for AI storytelling and experimenting with the emerging medium.
Coming soon to the AI & Creativity Summit in New York.
The new wave of slop is tool slop. The first wave polluted feeds. This one hits balance sheets. Vibe coders selling fear back to corporations through slick demos and six-figure deals until the vapor evaporates 6-12 months later. Don’t get tool slopped.
3
9
546
Happy Friday!
We just put DeepSeek-V4-Pro up on build.nvidia.com. It’s the world’s largest open source model at 1.6T parameters, and you can run it for free running on NVIDIA Blackwell GPUs.
Try the NVIDIA NIM API → build.nvidia.com/deepseek-ai…
The silence is deafening. I have been looking hard to find one actual slop farm they took down. But it increasingly just appears to be some of the most thoughtful and unique storytellers in the medium.
I hope the team at @FlowbyGoogle understands how demoralizing this is for people trying to actually make quality work with their tools. Not engagement farms. Not fake ads. Films. And it’s happening to some of the most talented artists trying to actually make meaning with these tools.
We’re making a list if you ever care to see it.
Hey @TeamYouTube, @YouTubeCreators!
My channel @javilopen on YouTube got demonetized (wrongly IMHO) for "Inauthentic content": content that appears mass-produced / automated.
Respectfully, I'd love to show what actually goes into each video (and if you are just randomly seeing this, it works kind of like a tutorial, because I'm showing some cool tricks in the screenshots 😉):
1/ Real human scripts and narratives
2/ Prompts crafted for hours until each image/video is "perfect"... well, perfection doesn't exist
3/ Hundreds of generations per shot / video (a video is 15 seconds maximum but each generation takes minutes), and I keep less than 5% of what I generate
4/ Strong edition (that takes several weeks for certain videos) in DaVinci Resolve / CapCut + plugins
5/ Color grading and post production
6/ Sometimes: real voice actors, foley artists (for the sounds), etc
This is literally the opposite of mass-produced. Some of the piece takes weeks/months!!!
Appeal already submitted. Hoping it reaches a human who can see the work 🙏
Reposts appreciated!
Matt Zien (KNGMKR) retweeted
Matt Zien is back for our upcoming AI & Creativity Summit in New York.
Matt is a Los Angeles–based filmmaker and Cannes-selected producer whose work bridges traditional documentary and the emerging world of AI-driven cinema. Before co-founding @KNGMKRlabs, Zien was SVP of Development and Creative Director of IPC (part of Sony Pictures Nonfiction), developing and producing 200+ hours of nonfiction television.
His body of experimental and award-winning AI films, including Post-Scarcity Blues and Last Call Before AGI, are known for their emotional realism and cultural insight, reflecting his broader mission to use innovation in service of human meaning rather than spectacle. He is the co-creator of the “The Tutorial,” a collaboration with the artist Grimes.
See him on stage at @thelighthouse very soon.
Matt Zien (KNGMKR) retweeted
Today, we released Lyra 2.0, a framework for generating persistent, explorable 3D worlds at scale, from NVIDIA Research.
Generating large-scale, complex environments is difficult for AI models. Current models often “forget” what spaces look like and lose track of movement over time, causing objects to shift, blur, or appear inconsistent. This prevents them from creating the reliable 3D environments required for downstream simulations. Lyra 2.0 solves these issues by:
✅ Maintaining per-frame 3D geometry to retrieve past frames and establish spatial correspondences
✅ Using self-augmented training to correct its own temporal drifting.
Lyra 2.0 turns an image into a 3D world you can walk through, look back, and drop a robot into for real-time rendering, simulation, and immersive applications.
➡️ Learn more: research.nvidia.com/labs/sil…
📄 Read the paper: arxiv.org/abs/2604.13036
Filmmakers at the dawn of cinema had so much beef Edison literally sent goons to smash competitors cameras and chased everyone to California. And the AI video scene is no different.
9
458
Matt Zien (KNGMKR) retweeted
Seedance 2.0 is now live worldwide on OpenArt.
No waitlist. No enterprise gate. Everyone's in.
60% off Wonder Annual for 2 weeks - unlimited generations, no cap.
Matt Zien (KNGMKR) retweeted
This month, Fairground secured native placement on major connected TV and FAST channel platforms for the first time in history, putting AI creator content on television alongside traditional films and TV series.
This isn't a dedicated app or a social media experiment. For the first time, long-form narrative AI cinema is being distributed the same way as traditional Hollywood content.
Over 50 creators and studios are now live across these platforms. For many viewers, it will be their first real introduction to AI cinema and what’s possible in the medium.
This didn't happen because the industry came to us. It happened because our team Colin Petrie-Norris (@cpetrien), Alexandra Mills, Petrie, Cameron Saless, Brian Decker, Kristofer Shinn and Chris Landa built the infrastructure before anyone believed it was necessary.
It also happened because a group of creators took a genuine chance on a vision that hadn't been proven yet: Noah Wagner, Jagger Waters (@glamoureptile), Winston Mayo, HENDO AI (@johnhenderson), Nick Dorra (@ndorra), Alex Patrascu (@maxescu), Brogan Wassell (@theBW7), Eli Green, Barney Miller, Markus Laukkanen (@Markuslkknn), Tatiana Laukkanen, Henri Rantala, Jesse Koivukoski (@JesseKoivukoski), BlvcklightAI (@BLVCKLIGHTai), Tiffani Lee Joseph, Reza Sixo Safai (@rezawrecktion), Andrew Mills (@AndrewMillsZA), Aaron Bhugobaun, Michal Jan Owerczuk, Maddie Hong (@hmdehong), and Sana Hussein.
While the broader entertainment industry is still debating whether AI belongs in the room, our creators are already on the screen introducing the world to quality, narrative long-form AI cinema.
And we're just getting started. Over 13 platforms will be active by year's end, with new competitions, projects, and opportunities for creators and media partners coming next week.
If you believe AI content deserves a real stage, we're building it. Come join us!
#ai #aicinema #genai #aifilm
Today, we're announcing the winners of the Arca Gidan Prize, Edition II.
We received 95 entries and, based on 7,285 votes from the community, awarded prizes to 16.
Congratulations @calvinherbst1, @davidmaruscsak, @emmacatnip, @fabdreamai, @Jas___black, material_rabbit, psiclone, @schmede_AI, @viktoria_vw, @VisualFrisson, blglavies, Dracula4ever, nasayuaop, sagans, tom_scaria, yuvraj, as well as everyone who entered!
The top 8 will also show their work at ADOS Paris next week!
Philosophers Episode VI - Simulation Theory
What we know about the simulation:
1. It hunts accuracy. The names survive sometimes, and fragments of the ideas. The evidence does not.
2. The only way to stop them from ending everything is to keep them entertained.
3. If they decide to delete you, they don't wait until you're dead.
Generous alpha ⬇️
Here is a diagram showing how I utilize Omni-Reference in @dreamina_ai Seedance 2.0. Every project has its own unique needs, but this architecture covers my general logic. I break the process into an Establishing Prompt for the first part and Continuity Prompts for everything that follows. The main difference is the shift from Setting the Stage (Mise-en-scène) to a Continuity Protocol, though the Global Instructions (Directorial Intent) remain the same throughout.
I use a hierarchical prompting structure where I define the intent and a Lookbook before the actual screenplay. This Lookbook covers visual blocking, which maps out exactly where characters are positioned on screen. Helping the model visualize these movements is key to reducing failures. Each subsequent part must reference the final frame of the previous shot as its opening point to keep everything aligned.
Using the previous shot as a video input is essential here. It is the best way to lock in spatial information, environment, pacing, and audio. When dealing with multiple characters, this extra layer of context is necessary to prevent the model from misinterpreting the scene.
Of course, the prompt itself depends on complexity of the scene. A complex scene needs more elaboration, while a simpler one allows you to hand over more creative control to the model. The goal is to find the perfect balance in your prompt where you aren't over-specifying or under-specifying.
Stay tuned for my next post where I’ll break down the differences between Establishing and Continuity Prompts and how to effectively use this hierarchical structure.
1
5
948
NVIDIA GTC with Grimes was a fun ride. Full fireside chat should be streamable now in the GTC archive.
























