There a lot of video models available.
* Google veo3
* Flux Kontext
* Seedance Video
* Hailuo Video
* Kling Video
There a lot of video models available.
* Google veo3
* Flux Kontext
* Seedance Video
* Hailuo Video
* Kling Video
All are paid subscriptions, I can't generate a one off video on them right?
You can. They exist. Forget the name of the big one people use now
Anyways, the generation and delivery mechanism are still not the same, which is the case here, I'm not trying to say you guys are wrong, just that there's a hint of uniqueness here you are refusing to acknowledge, it's alright, Im stoked about creating an inline webtorrent client with magnet links because of this, and that's enough for me π
As long as people are having fun π
And hopefully people make videos worth watching to put on that client.
There's only as much I can do π
I'll definitely be using it to post copyright video content π€£
The one nostr:npub1rtlqca8r6auyaw5n5h3l5422dm4sry5dzfee4696fqe8s6qgudks7djtfs made was pretty good.
I think thereβs some opportunity in promotional types or videos, maybe even product announcements etc. Iβm sure itβll even get good enough for entire movies.
For now itβs semi interesting when it works well but the majority of the slop is not appealing to me at least. That could change given the pace of progress in this field.
π
A whole feature film from a prompt is probably unrealistic. But creating scenes and editing them together into larger creations is already happening. But that takes talent, and a vision. There a lot on YouTube already doing exactly this. Most current models top out at ~10 seconds, so you end up with a lot of quick cuts. It will get better though.
Exactly. I used to do this professionally which is why my clips seem better than others. Iβm putting a lot of editing work into them.
This is costing Ray Buni real money in API fees he is paying to a centralized, commercial AI service, running the very models I mentioned.