Another thing that's really nice about making a rough version of your short film with AI as quickly as possible is that it allows you to fail fast.
I wrote a "really good" story yesterday. Then today I made a rough video of the whole thing (in under a hour), and watched it. The story doesn't work at all! Emotionally.
A few hours spent, but it's a few hours spent failing fast, instead of devoting 20 hours in making this and then only realizing that the story still falls flat.
It's always great in my head. And feels usually great even after I've written it down. But the translation process from words to video - writing a recipe vs baking the recipe and having a taste.
Failing fast is great.
Nano Banana Pro prompt:
A where is waldo image showing all Star Wars characters on Tatooine
First one to pull this off. First take. Even Waldo is there.
From screenplay to video - Sora 2
Opening of "Ukko"
I took the opening of one of my screenplays, and made a rough video of it using Sora 2 in one afternoon.
Insert obligatory soon Hollywood will be dead line here!
🧵See onwards for how I did it + compare the screenplay to the results + what are the current limitations + what will soon be possible.
I asked nano banana to get me into my favorite arcade and this is what happened...
all the details on how I made this video using visual prompting on freepik, below 🧵👇
I call this the normalization of magic.
In less than ~24 months, the world went from "video generation will never work, no way it happens" to "the audio on the generated video after the helicopter crashes is not as good as the sound of waves from the dinosaur riding the surfboard."
Today we’re introducing our latest open source video model—and it’s a big one.
This release sets a new bar for speed, quality, and control. It’s faster than anything in its class, packed with new features, and ready to run on your own hardware.
Let’s break it down 👇
1/7
@CuriousRefuge@midjourney@runwayml With animation projects, I always test 4-6 models to find out which one can animate the given style the best. In the end it never is the one I initially think it would be.
Voting is officially open for the audience favorite award for the world’s first AI Animation Competition we are hosting in partnership with @promise_ai and @LumaLabsAI
We wanted to share the phenomenal entry with you created by Markus Laukkanen.
We have been blown away by all of the phenomenal entries. Deciding on the finalist has been the hardest of all the competitions yet.
Voting ends tomorrow so head over to https://t.co/ppzwB4aqpa and vote for your favorite: https://t.co/VSUccghvqM
This is insane!
Check out this Robin Hood spec ad commercial from @fairground_tv & Glory House Studios
Youtube Channel: https://t.co/oXNQIlNN5L
made with @hedra_labs@capcutapp@elevenlabs@AnthropicAI@LumaLabsAI@midjourney
Book a consultation to turn your dream to reality: https://t.co/fNjVRj9ogi
And Portfolio: https://t.co/wgyk803rRF
Interesting bit of the US Copyright Office latest guidance on AI and copyright
- prompts alone are not copyrightable, but a combination of AI editing steps can lead to work where "the output should be copyrightable"
(!)
🚀3 weeks ago we ( @AIAnimationCom, @Uncanny_Harry, @8bit_e & me) started our AI Animation pipeline consultancy & now we have our first two gigs 🍾
Is your show 80% financed but you can’t find the rest? We could streamline your workflow, save you 30% & get you into production!
GORILLA STORY
A gorilla trying to make it.
5 minutes long. The most time I've spent on telling a story with a generative AI short movie so far. Hope you find the time to check it out.
Made using Flux, Runway, ElevenLabs, Suno, Stable Audio, Kling, Premiere.
🧵