I remade my May 14 AI video with Google Omni to see how much had changed in five months. The original is in the quoted post. This version is a two-minute film with 43 generated scenes. Omni generated the characters’ voices along with their performances. I’m going to keep remaking it every six months or so, keeping the core script and using the best models and tools available at the time. I’ll leave every version up. A couple of years from now, we’ll be able to watch AI video progress through the same story. I’m curious how long it takes for this version to look dated too. Compare it with May. Where do you see the biggest improvement?

For anyone wondering how it was made: - Google Omni Flash through Flow for all 43 scenes and the character dialogue. - OpenAI ImageGen for storyboard and reference images. - ElevenLabs Eleven v3 with Bill for narration, ElevenLabs Music for the score, Sound Effects v2 for additional sound, and Scribe v2 for word timing. - Codex + OpenMontage handled the production workflow, gflow handled Flow generation, and FFmpeg handled the edit and final mix. There were multiple takes, revisions and cleanup passes to get here.

FIRE 📱

Word 🤣

Amazon banned Meta Muse AI but @nsb_bot still works flawlessly 😎 #ifykyk

I'm so excited to see that we got our first checkouts Today on NSB Go from users with early access! This will revolutionize everything you've ever known about botting, and finally put it, FULLY, in your pocket.