Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

🎬 ONE LAST QUESTION A short film created with Seedance 2.5 using the Hedra Agent. The first version had one small anomaly that broke the illusion. Instead of editing the video manually or regenerating the whole sequence, I went back to the Hedra Agent, showed it exactly what needed...

19,966 Aufrufe • vor 1 Monat •via X (Twitter)

27 Kommentare

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

Here’s the part I love. The Hedra Agent corrected the anomaly without changing everything around it. No masking. No compositing. No manual repair afterwards. The corrected moment blends straight back into the original film. Unless I showed you where the problem was, you probably wouldn’t know it had ever been there. Next I’ll show you the original version, the exact moment that needed fixing, and the corrected result side by side. Try it for yourself using @hedra_labs: Here is the link:

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

As you can see in the original version, the hands became jumbled and broke the illusion. Instead of taking the video into another app to cut, mask or rebuild the shot, I simply gave the Hedra Agent instructions explaining what needed fixing. It corrected the problem while keeping the characters, scene and continuity intact. For me, this is a huge step forward for AI filmmaking. If we can repair specific problems through simple instructions, without rebuilding entire sequences in an editor, making polished AI films becomes much faster and easier.

Profilbild von Caspian
Caspianvor 1 Monat

@hedra_labs Very cool, Jo. I really liked the video too. Fixing only the moment that breaks continuity instead of redoing everything is super useful.

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Yeah, it makes life a lot easier. Thank you so much, Caspian. 🫶

Profilbild von TechieSA
TechieSAvor 1 Monat

@hedra_labs Biggest plot twist bro🤯

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs I’m glad you like it, Techie! That’s really encouraging. Thank you so much!

Profilbild von Shara
Sharavor 1 Monat

@hedra_labs Agent based fixes beat full regeneration for small anomalies.

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Very true. It makes like a lot easier. Thank you, Shara.

Profilbild von AntonioAI
AntonioAIvor 1 Monat

@hedra_labs Amazing to watch

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Thank you so much, Antonio. Nice to meet you 🫶

Profilbild von AlexAI 🇺🇸🇺🇸🇺🇸
AlexAI 🇺🇸🇺🇸🇺🇸vor 1 Monat

@hedra_labs 👍👍👍

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Thank you so much, Alex. I saw your retweet. I’m so grateful.

Profilbild von AlexAI 🇺🇸🇺🇸🇺🇸
AlexAI 🇺🇸🇺🇸🇺🇸vor 1 Monat

@hedra_labs 💯 🔥

Profilbild von Teknikal Ed
Teknikal Edvor 1 Monat

@hedra_labs Movie time 🎬 🔥🔥🔥

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Thank you, Teknikal! I’m proud of this one. I’d like to extend it!!!

Profilbild von Teknikal Ed
Teknikal Edvor 1 Monat

@hedra_labs Please do!!! A lot of tension and already felt like there was a story behind the scene. A good watch 🎬🔥🔥

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Thank you so much 🫶

Profilbild von Teknikal Ed
Teknikal Edvor 1 Monat

@hedra_labs Pleasure Jo ☺️

Profilbild von Abkr Sadiq
Abkr Sadiqvor 1 Monat

@hedra_labs That kind of precision is the real breakthrough. Fixing one tiny moment without destroying character and scene consistency could completely change AI video workflows.

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs For real. It makes creating a lot less frustrating.

Profilbild von Shams
Shamsvor 1 Monat

@hedra_labs To be very honest, this one hits the realism on another level.woww these skin details Great job, buddy 👏🏻👏🏻

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Thank you so much, Shams. I think im gradually improving,

Profilbild von Shams
Shamsvor 1 Monat

@hedra_labs Yes this one is damn impressive 😍 Keep it up you are doing in the right direction.

Profilbild von Vini
Vinivor 1 Monat

@hedra_labs Fixing one anomaly without redoing everything, that is handy.

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Exactly. These major changes are really helping new creators get past the hurdles we had when we started. Thanks, Vini. 🫶

Profilbild von Mazi Obi
Mazi Obivor 1 Monat

@hedra_labs I always love to see your work.

Profilbild von Jo Aiba
Jo Aibavor 1 Monat

@hedra_labs Thank you so much, my chief supporter!!!

Ähnliche Videos

hey here is the final result of octopus invaders on nvidia's flagship at full precision. nemotron super 120B on 2x H200 NVL. BF16 unquantized. 287GB of VRAM. hermes agent as the harness. 60 tok/s. first try it autonomously coded for 6 minutes straight. created 11 files. correct project structure. correct load order. started the server. i opened the browser and the result was a blank screen. i did not give up. second try i gave it a precise list of bugs and things to fix. it went back in for another 3 minutes. patched the code. served it again. still blank. so i did what any sane person would do. third try i just said the screen is blank, test it and fix it yourself. and this is where nemotron showed what it actually is. it became a debugger. you can see it in the video. realtime CSS test squares, red screen flashes, hermes agent browser tools, inspecting its own output. it built the parallax background with planets and comets. it rendered a rocket ship that tracks your mouse with fire and bullet physics. the aesthetic is real. but no enemies spawn. no collision. not playable. what surprised me is qwen 27B one shotted this exact game on a single RTX 3090 at Q4 quant. and here is nvidia's flagship at full precision on enterprise hardware needing 3 tries and still not getting there. that makes my hope high for the undisputed qwen 122B which is about to face the same test next. same hardware. same prompt and same harness. lets see if it one shots or not. full session in the video. no cuts. 5x speed.

Sudo su

11,011 Aufrufe • vor 6 Monaten