AI can’t do this. Seedance 2.0 can’t do this. Prove me wrong!

This actually looks very similar to an audio reactive comfyui workflow. Yours is better IMO But honest question. It looks like you may have used a computer and not keyframed by hand. Whats the difference if someone uses slightly different math to make something like this?

the comfyui workflows look like ass lolol but I'm sure AI will be able to sync visuals nicely at some point. It's just not a priority yet as the race is to "cook hollywood" and to answer your question, yes I used a computer but I also keyframed this by hand. The difference is simply this is still very hard to do right now and soon it won't be because AI will take the wheel.

Honestly I think the workflows are a skill issue. I'm not comparing this to yours, but I managed to get this in about 30 mins. Glitch app and some keyframes into the workflow. I'm sure someone like you could do way better. It's really just creative keyframe interpolation (zero prompts btw)

Interesting result! Thanks for sharing

I'll just say it looks awesome. But don't tempt me on AI capabilities.

Music by the great sound wizard (🎧🧙‍♂️) @Abel_Okugawa

Had a few minutes to see what I could come up with

Oh also the way the energy gradient / texture shifts through the lines is pretty damn close to my work, closest I’ve seen in that aspect!

Doesn't hold a candle to your work. And would take much more effort to get audio reactions happening. Just wanted to see what I could manage in a few minutes.

The history of AI is a vast graveyard where every ‘Artificial intelligence will never…’ has been laid to rest.

AI can’t do that , right now? Give it six months. This pattern repeats itself with unsettling precision — the same delays, the same cycle, the same outcome. Time passes, circumstances shift, and somehow the window always reopens. It never fails. Six months later, the conditions align again as if they were always meant to.

bro give ai like 20 mins to figure it out, stop coping

We used to talk about "You have an extra frame in your loop at :24", but now the imperfections might just be more evidence that it's not AI. It's a strange time for art.

actually I think that's a glitch in twitter plackback lolol... it's not looped! But I totally know what you mean and yes very strange for sure.

@chelseaevenstar Wow

Even if it can… Nothing beats the artist

This looks exactly like something ai could do.

this is cope

I tried, looks like you're right! I was kind of expecting it to be able to do some form of audio-reactive effects but not really... (this is with it making the audio itself, I'll still have to test with audio input)

Sexy 3d art!🔥👀🔥

seedance 2.0's limitations depend on the task complexity and data input quality

but you're adding to the dataset when you post 😬😬

hahah nothing I can do about that...

you forgot to mention "yet" , AI can’t do this YET !

there is a way to get Ai to generate visuals based on sound wavs so yes sorry to tell you ai can do this

human touch still wins

sometimes limits show more than the wins do

If you mean an audio‑reactive visualizer. LTX‑2 can take audio as input to drive the generation. It’s not very good yet, but I think it’s just a matter of time. You could improve it by training a LoRA on an audio visualizer dataset. Anyway, here are my two attempts with it.

So AI can't use reactive noding system in comfyUI to make visualizers now?

give it a try. it looks like ass!