All-in-one AI video app to storyboard, create, & collaborate using generative AI models: text-to-video, image-to-video, video-to-video. #ImagineTogether
@Kyrannio This seems like it might be the result of forcing non-words (instruments) into a word based format. That's not well represented in speech to text training and so the "best fit" might be rather odd. Consider looking at the data itself (without LLM)? e.g. ffprobe
@rainisto@henrydaubrez The challenge I find here is someone who feels they "know the full process" and "you know the AI part", creates rigidity of past structure. On the other hand showing up and telling someone "the new way to do things" is also false and overstated. There are many new ways.
@henrydaubrez I'm working on a hands-on educational event with machine cinema. I open with a clip from the Junkyard King (thank you!) and discussion of process - "how we make key frames." Right now, I think key frames sit at the intersection of idea and story. Defining the next agentic wave
@phillip_isola@sainingxie Hmm. What forces the thicket structure? From comp Neuro it makes me think of the pressure from that an inhibitory predictive model could be used to de-correlate the known redundant shared output. Any intuition about mutli-headed attention pushing similarly?
@pt5films@henrydaubrez I'm happy to give you a tour. (DM me) Our platform is live but it's a work in progress adapting to the changing landscape. Since both primary lead and community engage with AI, I think a key part is how much control to cede and constraint to impose on your community.
@henrydaubrez I think I remember this character from your earlier tests! And the glowing sprite is a theme too. So your familiar spaces and themes, and the logic of the magic in your world gets to grow along your own journey. Even across different tools / models. (E.g. Seedance)
@henrydaubrez@pt5films I am actually very interested in the small team approach. Or "one primary lead plus community". I think of it as "asymmetric collaboration" where there is clarity that one person holds narrative authority, and there is a clear contract for participation.
Selection is now closed. Anyone who sees that and thinks it might be interesting to them or their Twitter community, chime in with a short message and I can mention you in our next posted opportunity.
Lol, the difference between other channels and the crickets here. I am glad to have linked in. Here is a teardrop for shouting into the void when you build in public and have no following on twitter.
Included perks for next week:
NanoBanana Pro, Sora 2.0, Kling 3.0
Grok – surprisingly strong and cost-efficient
Seedance 2.0 – drops Feb 24
Plus 30+ of the best AI video models — all in one production workflow.
What makes 10zebra different?
Agentic quickstart workflows.
Storyboard → AI video generation → curation → trim → 1-click line edit
Images, video, or sequences.
text2vid, image2vid, vid2vid. Camera compatible. Hybrid production.
Use any video clip as reference across models