These aren't even that good in comparison to the stuff being posted on Reddit and twitter/X.
Some of the stuff I've seen is mindblowingly impressive in how they've captured quality creative decisions, and how the models can reason about visually appealing designs.
It understands things like continuity, themes, facial expressions, cultural memes/references, symbolism, etc. And it generates things using a lot of multi-modal behavior by using 3d, images, etc.
Agreed. Just yesterday, I recreated the XKCD comic “A Bunch of Rocks” (<a href="https://xkcd.com/505/" rel="nofollow">https://xkcd.com/505/) as a video. It took me about five hours, using ElevenLabs for the audio and a step-by-step process involving brainstorming, planning, storyboarding, and so on. Opus 5.5 has become remarkably capable.
There isn’t that much to say. Start with an empty directory, put the XKCD image in it, add your ElevenLabs API key, and set up a Python virtual environment. Install ffmpeg.
Then start Claude. From that point on, it’s mostly prompting and evaluating the results. Like any good engineer, you shouldn’t simply start with “Create a video from this comic.” Instead, take a step-by-step approach: discuss the visual style, animations, ask for audio examples, create a script, and then develop a storyboard (ask for a .html document).
The important part is to break the whole generation process down into small, manageable steps. I also used the brainstorming skill from Superpowers for the discussions. It makes the whole process much easier when the AI asks the questions and I just have to answer them.
The result is 4500 line python code generated by Opus 5.5 and a lot of video and audio files.
preommr · · focus · HN ↗
Some of the stuff I've seen is mindblowingly impressive in how they've captured quality creative decisions, and how the models can reason about visually appealing designs.
It understands things like continuity, themes, facial expressions, cultural memes/references, symbolism, etc. And it generates things using a lot of multi-modal behavior by using 3d, images, etc.
Some really incredibly impressive stuff.
s-macke · · focus · HN ↗
[0] <a href="https://simulationcorner.net/A_Bunch_of_Rocks.mp4" rel="nofollow">https://simulationcorner.net/A_Bunch_of_Rocks.mp4
adamgordonbell · · focus · HN ↗
Say more about the process and / or drop some link pointers.
s-macke · · focus · HN ↗
Then start Claude. From that point on, it’s mostly prompting and evaluating the results. Like any good engineer, you shouldn’t simply start with “Create a video from this comic.” Instead, take a step-by-step approach: discuss the visual style, animations, ask for audio examples, create a script, and then develop a storyboard (ask for a .html document).
The important part is to break the whole generation process down into small, manageable steps. I also used the brainstorming skill from Superpowers for the discussions. It makes the whole process much easier when the AI asks the questions and I just have to answer them.
The result is 4500 line python code generated by Opus 5.5 and a lot of video and audio files.