Some @Grok Imagine testing.
Well, little Nic is officially entering his "villain era." I threw a desaturated, cinematic overlay on the latest test, and honestly? He looks like he’s plotting to overthrow the entire kingdom. I don't normally use cinematic overlays because it is overdone and technically being used incorrectly. I feel more comfortable now using the effects because the anchor is coming along better. Most of these big budget AI things people see can be slapped on in 5 seconds and they call it a masterpiece. I feel the anchors or characters need to be examined far closer than outputting high quality renders before you go off and gloat. This example has slight drift, but I'm fine with that as long as I'm teaching and curating the anchor.
I’ve been obsessed with mouth and facial movement watching frame by frame until the movement stops looking like an AI render and starts looking like actual human tissue.
Most define realism by resolution or skin texture. That’s just makeup. I’m looking for the signature.
Every language is a lifetime physical workout. The way you round your lips for French versus how you flatten them for English isn't just an accent; it’s a physical habit that carves the resting tension of your face over time.
I’m currently stress-testing my 9-year-old anchor against French, Italian, and English phonetics. A younger face is more supple, letting me see the imprint of the language form in real time. If I aged him up a decade, the muscle resistance would be entirely different. You can't just slap a wrinkle filter on a face and call it aged; you have to fundamentally shift the biomechanics of how that face handles language.
I used the desaturated overlay to strip away the shiny veneer. When you remove the color noise, you can’t hide the biomechanics anymore. You’re left with the raw physics. If the coupling between the thought and the muscle isn't there, the illusion breaks.
It’s not just about the lips moving to audio. It’s about the lips knowing how to move because the character is actually thinking the language, not just performing it.
So @BytedanceTalk released Seedance Mini, now available in @dreamina_ai and @capcutapp.
Our tests confirm that it’s definitely faster and cheaper, but the quality differences were also pretty noticeable. Faces looked more synthetic, there was more artifacting, and more issues with continuity and physics.
Overall, Seedance Mini was impressive, but if you're choosing between "good" and "really good," we'd still lean toward the original model, especially for professional work.
I wish Elon Musk posted more about AI again. Especially about cool things happening at xAI, like upcoming new models. We haven't heard a lot about Grok 5 lately. 🤷♂️
I made a 1m50s sci-fi short entirely with an AI workflow.
Visuals generated in Kling 3.0.
Editing, color, and music by Codex.
It’s called THE LONG WAY HOME.
A film about space, memory, and the one place every journey leads back to.
This is my Dad’s reaction. Pure joy.
One thing I’ve learned with AI is that the technology itself isn’t emotional. The emotion comes from the context you bring to it.
Tap in.
What would it feel like to remember from inside a family portrait?
Not just look at an old photo, but step back into it.
This is my family in the late 1980s. My Black dad, my Korean mom, my sister, and me. A military family living in West Germany.
When you're a kid, memories don't feel cinematic. They feel ordinary. Riding in my dad's black Cutlass on the Autobahn. Family trips to Holland. Playing on military base playgrounds. Falling asleep in the backseat on the drive home. Monopoly on the patio. Summer cookouts. Moments you don't realize you'll miss someday.
One thing I've learned with AI is that the technology itself isn't emotional. The emotion comes from the context you bring to it.
These images aren't powerful because they move. They're powerful because they mean something to me. It tells my story.
Somewhere between a dream and time travel, I was able to revisit a moment that only existed in a fading photograph and in my memory. Before the aging. Before the diagnoses. Before life became more complicated.
Just a happy family together.
As a kid, it felt like we wore the same outfits every day and that summer would last forever.
Now I have children of my own, and I understand why my parents took these photos in the first place.
This video became a Father's Day gift to my dad, but also a gift to myself.
A chance to spend a few more minutes in a memory I thought was gone.
This was created from my real family portrait.
Happy Father's Day ❤️
Anime is experiencing a renaissance and filmmakers are making cuts like never before.
We have entered the new era, fusing AI technology with animation. Are you ready?
@kojomojoworld#animeAI#AIfilmmakers#FilmFire
How about some good old 90s action chase scene in New York City? This was really fun to make. Super Cheerleader chased by a Helicopter. Save the Cheerleader, Save the World. It took me 7 days to shoot and cost me around a million dollars. Thank you New York City for letting me shoot this scene there. You guys are awesome.
As a thank you for 1000 followers, I give you epic slop with my dear friend SLY. Thanks to all of you, guys. You guys are the best. lmao everything is so bad, the acting, the reloading, but this is part of the ai charme.