Visual effects went through this same development issues as the industry matured. What took this industry decades to advance it taking months in AI. Think the spaghetti Will Smith and now this. Another one people don't mention here but is specific to video is higgsfield ai.
Unsure if it's just the way they prompted it / coded it, but the output is far too much a literal direct copy of the lyrics. The best music videos have a story arc on the theme of but often not litearlly the lyrics, and start with obscurity and reveal something (following all the literary/story mechanisms)
Consider Amber Run - Found lyrics versus the video, and the story arc of the video
The entire thing was cringeworthy to the core. I kind of enjoyed it though because it perfectly epitomized "AI slop" in the first 30 seconds so wonderfully. "Michelle Pfeiffer, that white gold" - show a blonde woman in a gold sequined top! "Livin' it up in the city" - show a shot of a big city!
If anything, the absurd literalism of the video contrasted so perfectly with the (IMO) brilliant clever originality of the lyrics. E.g. "Michelle Pfeiffer, that white gold" is actually a not-so-subtle reference to cocaine. Imagine if the lyrics were as stupidly unoriginal as the video ("Now we're all snorting cocaine!!").
Wierd Al videos are a parody of an existing property. "White and Nerdy" is a parody of "Riding Dirty by Chamillionaire," but the lyrics are about nerd stereotypes (as an intentional contrast with black culture as presented in the original,) and a great deal of creative effort is put into making those lyrics humorous while also fitting to the original theme. Nothing about Wierd Al's videos are "totally literal," certainly not in the sense of these AI videos, which are "literal" in the sense of "literally showing what the lyrics are describing."
Curious how much time in addition to tokens this costs. If you have to spend $25 and wait 45 minutes to get a basically unwatchable video, I'm not worried about indie film makers being replaced just yet...
This wasn’t possible even a year ago, with the speed of things changing and how much money is spent on movies is there really a doubt that someone will be able to make $100 million movie for less than $1 million in token spend?
I’m curious (admittedly skeptical) what you mean by this. Are you talking about a world where director’s just like…don’t actually make movies and create AI media?
There's definitely still room for innovation in custom use cases with AI. I like to write and draw comics, but it's very time consuming to make a finished product. Working with tools in default and you'll have a bad time. You have to really guide it and if you're using a longer story to adapt, it'll compress things and lose context during Thinking. Luma labs was the most interesting tool I've seen so far, but there's still a lot of room for growth
It is jarring to me that most of the dancing seems slightly out of sync with the music. It is like a music video uncanny valley - images look good, but the lack of sync to the sound shatters the illusion entirely.
To me nearly all of the real world dancing looks like that. The dancers could literally be olympic level and I still don't see the connection between their movements and the sounds of the music. Music videos sometimes do sync for me but only on very simple motions, and for whatever reason shuffle dance on video almost always looks great.
Even their demo clips are soundly in the uncanny valley, and I assume normal folks using it and having to pay for each clip will have generally worse results.
The GPT ones are strange. The $25 fable one to me is subjectively better than the others. The $100 fable one is too literal and robotic.
The jevons paradox is you need auteurs to curate vignettes or effects and cut or mask them in etc. That's not really different philosophically when software entered art in other ways. I could see errors/glitches lowering in time but I doubt there will be much acceleration.
> need auteurs to curate vignettes or effects and cut or mask them in etc
The problem is that reliable, repeatable professional-grade commercial art and design sensibilities happen in full-time careers. It’s entirely different than fine art, where intense self-exploration and experimentation are a very viable option.
These tools are exacerbating an already difficult creative job market so there’s no reasonable path to get those skills. Our creative professional pipeline is fundamentally broken.
The same thing is happening in software, I see the ladder pulled up and don't feel vulnerable as senior staff. If anything, we face a massive and increasing competency crisis in computing because there is a cult dumb enough to believe acceleration and doomer cases for LLMs.
As someone who has marketed music, shot music videos, directed music videos, cut music videos together from stock footage: you don’t need auteurs.
You did back when MTV made songs big.
No one actually pays attention to the details of music videos any more. It’s visual chewing gum at best. The reality is that now, if you have something half decent, nice colours, nice lighting and a wee bit of a story, no one is going to care.
The only other route is a huge budget spectacular - but you only get the huge budgets if you label lends you the money to make a huge budget video because they think it will increase the amount of money they make - while extending the amount of time it takes you to recoup.
Ultimately, now, it is just another social media asset, so promo videos are built with that in mind.
> No one actually pays attention to the details of music videos any more. It’s visual chewing gum at best.
Hilarious to hear someone in industry blame their audience for the commodification of the medium. Is every industry like this? Surely nobody goes into creative fields thinking “I can’t wait to feed the masses slop!” Who’s killing our spirit?
He's not wrong though. In the 80s I'd watch video shows and over the course of a week I'd probably see some videos 10 times. And it wasn't background filler - I'd actually be sitting in a chair/couch and watching the videos. Kids don't do this anymore. First many/most songs are made popular through TikTok memes, not videos. And videos really are mostly just played in the background as they do other stuff. No one is just tuning into Yo MTV Raps or Headbangers Ball anymore.
Video killed the radio star. Streaming killed the video. Sure, lots of people use youtube for consuming music, but how many of them truly are watching the videos or just have them on while they do something else without seeing the images? With that in mind, putting anything on screen is just checking the boxes
Seems like if you build some more scaffolding around it, it wouldn't be bad. I think AI video isn't quite there yet so you probably would want to lean into that. For example you could ask for an animated or cartoon music video so the real shots don't look weird. Also if you gave it some guidance on what a good music video is like it would probably help as well. But yeah idk may be that's not the goal here.
They are terrible. Perhaps if one on purposely prompts for a weird imaginary, the result could be less awful even acceptable, like that famous Gucci ad. These videos instead are common scenes with the super cringey AI artifacts.
But why spend the same amount of money on AI instead of humans? My guess is that shooting a music video is probably fun for a lot of artists. And with AI the result is not predictable and might be inconsistent in the dumbest ways.
My guess is that an AI music video would have to be a lot cheaper for artists to consider it outside of making one just because you want to make an AI music video.
Though we're finding the studios contracted to do this can bill $50k. I know several studios that previously billed clients six figures for ad campaigns (P&G, HBO, pharma, etc.) are now charging five figures and winning a lion share of the bids now.
Not sure why Wan is the focus of this article and Seedance is a footnote. Wan/LTX/open models are significantly behind Chinese closed source models. (And the Chinese have left the Western models in the dust.)
These are pretty terrible, but for me there were at least a few moments where they genuinely became "so bad they're awesome". They actually re-enforced an idea I keep coming back to: sometime soon a real artist is going to use AI to make something amazing, not by aiming for flawless "realism" or some kind of pastiche slop, but by leaning into the weirdness that often comes out of AI.
Video can be good if you stick to a garden path of simple scenes with tons of examples in the training material and not a ton of motion between overlapping objects in a scene, and don't really care too much about specifics.
As soon as you want something very specific, or something novel, or anything with a lot of moving objects/people, it falls apart.
As a clanker-apologist I gotta say this is the strawman of AI slop brought to life. Letting the machine do literally everything without even supervision and see how it turns out? No surprise at all the results are so bad.
My fav AI video is still Post-Scarcity Blues from a year ago https://www.youtube.com/watch?v=q_t3h2AZ0KY There have been others I've enjoyed since then. But, that one stands out in memory. Work warning: it is occasionally just a bit spicy.
Spice 1 released an AI video for Jealous Got Me Strapped. Got a young AI Spice 1 and Tupac. It's interesting
https://youtu.be/j36hjNuAIWQ
These are awful. It’s like Suno music. Seems convincing if you half listen. As soon as you pay attention you notice all the cracks.
Unless I'm misunderstanding the article, or your comment, the models were responsible for generating the music _videos_.
The music itself is Uptown Funk... which was a very successful song in 2014 (https://www.youtube.com/watch?v=OPf0YbXqDm0)
The videos are indeed awful though.
They're just saying that like Suno Music, if you look closely the cracks show. They're not saying the music is AI generated.
Visual effects went through this same development issues as the industry matured. What took this industry decades to advance it taking months in AI. Think the spaghetti Will Smith and now this. Another one people don't mention here but is specific to video is higgsfield ai.
It is remixing what the industry created, not advancing past it. OpenAI even ended Sora in under a year. Anthropic doesn't even bother.
Those old movies with old visual effects are watchable and still enjoyable. This is never good, interesting or enjoyable.
Spaghetti Will Smith was 3 years ago, so it’s taking a little longer than months.
Unsure if it's just the way they prompted it / coded it, but the output is far too much a literal direct copy of the lyrics. The best music videos have a story arc on the theme of but often not litearlly the lyrics, and start with obscurity and reveal something (following all the literary/story mechanisms)
Consider Amber Run - Found lyrics versus the video, and the story arc of the video
https://www.youtube.com/watch?v=Yj6V_a1-EUA
Literal music videos are still fun and a valid creative direction, e.g., Vance Joy's "Riptide": https://www.youtube.com/watch?v=uJ_1HMAGb4k
The thing about art is that literally everything is a "valid creative direction." But that doesnt make everything immune from derision.
Coincidence that both songs reference Michelle Pfeiffer or was that free connotation at work?
Free association?
That’s not really the same thing.
Riptide gets away with it by being so utterly on the nose it reads as ironic.
In an interview, an adult actress was asked about the things she says during scenes. She said she describes what is happening literally at any moment.
This is what LLM models do.
The entire thing was cringeworthy to the core. I kind of enjoyed it though because it perfectly epitomized "AI slop" in the first 30 seconds so wonderfully. "Michelle Pfeiffer, that white gold" - show a blonde woman in a gold sequined top! "Livin' it up in the city" - show a shot of a big city!
If anything, the absurd literalism of the video contrasted so perfectly with the (IMO) brilliant clever originality of the lyrics. E.g. "Michelle Pfeiffer, that white gold" is actually a not-so-subtle reference to cocaine. Imagine if the lyrics were as stupidly unoriginal as the video ("Now we're all snorting cocaine!!").
Fun fact (if you care): Back in the '90s, pretty much every music video produced in Azerbaijan literally matched the lyrics.
Weird Al videos are often totally literal and extremely fun as a result.
https://youtu.be/N9qYF9DZPdw?is=tU_8p-hDZv9gjAJ6
Wierd Al videos are a parody of an existing property. "White and Nerdy" is a parody of "Riding Dirty by Chamillionaire," but the lyrics are about nerd stereotypes (as an intentional contrast with black culture as presented in the original,) and a great deal of creative effort is put into making those lyrics humorous while also fitting to the original theme. Nothing about Wierd Al's videos are "totally literal," certainly not in the sense of these AI videos, which are "literal" in the sense of "literally showing what the lyrics are describing."
Pretty neat that the gallery of videos plays from where the previous one was paused. Nice UX.
> None of the music videos were great
Glad they acknowledge this.
Curious how much time in addition to tokens this costs. If you have to spend $25 and wait 45 minutes to get a basically unwatchable video, I'm not worried about indie film makers being replaced just yet...
"not great" is a huge understatement
This wasn’t possible even a year ago, with the speed of things changing and how much money is spent on movies is there really a doubt that someone will be able to make $100 million movie for less than $1 million in token spend?
Do you... Have a lot of professional experience with making $100 million-budget movies?
Directors and editors using Seedance can fire the film studio.
This is a fundamental shift in how storytelling is funded and made, not in who does the driving.
Same as is happening with code.
>funded and made
I’m curious (admittedly skeptical) what you mean by this. Are you talking about a world where director’s just like…don’t actually make movies and create AI media?
I'm classically trained and I honestly can't really tell how these are worse than the human produced ones. They all look kinda the same to me.
[delayed]
How well do these models do producing a proof-of-life video? Asking for a friend.
Did it misunderstand "don't believe me just watch" and put a bunch of washing machines in on the Fable 5 $100 at about 1:20?
There's definitely still room for innovation in custom use cases with AI. I like to write and draw comics, but it's very time consuming to make a finished product. Working with tools in default and you'll have a bad time. You have to really guide it and if you're using a longer story to adapt, it'll compress things and lose context during Thinking. Luma labs was the most interesting tool I've seen so far, but there's still a lot of room for growth
It is jarring to me that most of the dancing seems slightly out of sync with the music. It is like a music video uncanny valley - images look good, but the lack of sync to the sound shatters the illusion entirely.
To me nearly all of the real world dancing looks like that. The dancers could literally be olympic level and I still don't see the connection between their movements and the sounds of the music. Music videos sometimes do sync for me but only on very simple motions, and for whatever reason shuffle dance on video almost always looks great.
These videos are far from good, but it's possible to create a decent AI music video if you use a human-in-the-loop workflow: https://www.youtube.com/watch?v=52Km88uAZ4M
Embarrassing!
You can still delete this, there is time.
What's embarrassing?
The outputs
I did let Claude Code Opus + OpenRouter API Key (limited to 25EUR) create an an Arthouse Video about "Is AI art or can it go?"
https://m.youtube.com/watch?v=E6XofKqFYeQ
I like the scene with the hands.
Please repeat this with models from MiniMax:
https://www.minimax.io/news/minimax-hailuo-23
(No affiliation, I just want to see how well they work).
Even their demo clips are soundly in the uncanny valley, and I assume normal folks using it and having to pay for each clip will have generally worse results.
Please stop using AI for creative work. This is utterly abysmal.
The GPT ones are strange. The $25 fable one to me is subjectively better than the others. The $100 fable one is too literal and robotic.
The jevons paradox is you need auteurs to curate vignettes or effects and cut or mask them in etc. That's not really different philosophically when software entered art in other ways. I could see errors/glitches lowering in time but I doubt there will be much acceleration.
> need auteurs to curate vignettes or effects and cut or mask them in etc
The problem is that reliable, repeatable professional-grade commercial art and design sensibilities happen in full-time careers. It’s entirely different than fine art, where intense self-exploration and experimentation are a very viable option.
These tools are exacerbating an already difficult creative job market so there’s no reasonable path to get those skills. Our creative professional pipeline is fundamentally broken.
The same thing is happening in software, I see the ladder pulled up and don't feel vulnerable as senior staff. If anything, we face a massive and increasing competency crisis in computing because there is a cult dumb enough to believe acceleration and doomer cases for LLMs.
As someone who has marketed music, shot music videos, directed music videos, cut music videos together from stock footage: you don’t need auteurs.
You did back when MTV made songs big.
No one actually pays attention to the details of music videos any more. It’s visual chewing gum at best. The reality is that now, if you have something half decent, nice colours, nice lighting and a wee bit of a story, no one is going to care.
The only other route is a huge budget spectacular - but you only get the huge budgets if you label lends you the money to make a huge budget video because they think it will increase the amount of money they make - while extending the amount of time it takes you to recoup.
Ultimately, now, it is just another social media asset, so promo videos are built with that in mind.
None of these would cut it.
> No one actually pays attention to the details of music videos any more. It’s visual chewing gum at best.
Hilarious to hear someone in industry blame their audience for the commodification of the medium. Is every industry like this? Surely nobody goes into creative fields thinking “I can’t wait to feed the masses slop!” Who’s killing our spirit?
He's not wrong though. In the 80s I'd watch video shows and over the course of a week I'd probably see some videos 10 times. And it wasn't background filler - I'd actually be sitting in a chair/couch and watching the videos. Kids don't do this anymore. First many/most songs are made popular through TikTok memes, not videos. And videos really are mostly just played in the background as they do other stuff. No one is just tuning into Yo MTV Raps or Headbangers Ball anymore.
Video killed the radio star. Streaming killed the video. Sure, lots of people use youtube for consuming music, but how many of them truly are watching the videos or just have them on while they do something else without seeing the images? With that in mind, putting anything on screen is just checking the boxes
AI videos as in remotion based videos look much better imo since it can code much better than it can prompt for videos with a coherent narrative
https://youtu.be/uDAeAuYyl0E (parody of Claude announcements) https://youtu.be/cSsVNtGPOIg (recreating a fireship video)
This is incredible! Did you make these yourself? I am familiar with remotion but this is a lot more than just splicing images, etc.
Did you use a skill library to make this?
So AI could generate a great lyrics video.
Seems like if you build some more scaffolding around it, it wouldn't be bad. I think AI video isn't quite there yet so you probably would want to lean into that. For example you could ask for an animated or cartoon music video so the real shots don't look weird. Also if you gave it some guidance on what a good music video is like it would probably help as well. But yeah idk may be that's not the goal here.
They are terrible. Perhaps if one on purposely prompts for a weird imaginary, the result could be less awful even acceptable, like that famous Gucci ad. These videos instead are common scenes with the super cringey AI artifacts.
Regular music videos (including the writing/recording) can easily go into 6 figures. I wonder what the $200,000 AI music videos looks like.
But why spend the same amount of money on AI instead of humans? My guess is that shooting a music video is probably fun for a lot of artists. And with AI the result is not predictable and might be inconsistent in the dumbest ways.
My guess is that an AI music video would have to be a lot cheaper for artists to consider it outside of making one just because you want to make an AI music video.
Okay, then a $50,000 music video.
The idea is to reduce production cost and therefore more plentiful/accessible.
You don't need to spend $200k, because results can be bad for cheap.
https://youtu.be/HDdsKJl92H4
Though we're finding the studios contracted to do this can bill $50k. I know several studios that previously billed clients six figures for ad campaigns (P&G, HBO, pharma, etc.) are now charging five figures and winning a lion share of the bids now.
Not sure why Wan is the focus of this article and Seedance is a footnote. Wan/LTX/open models are significantly behind Chinese closed source models. (And the Chinese have left the Western models in the dust.)
I wouldn't do anything production grade in Wan.
3 of the 4 agents used Wan, the other one used Seedance.
When the line was "don't believe me just watch"
And then the clip was literally just an arm wearing a watch!
That's freaking hilarious!
It's like someone playing charades
the giant cartoon dragon with a sign that says "retired" made me guffaw
Skip the Claude Fable 5 $25 video to 1:42. The disembodied Adams' Family hand is on the job.
It's Wan video, which is a shitty video model.
The OP would exclusively be using Seedance 4k if they were serious about this.
It was the agents job to pick the model.
These are pretty terrible, but for me there were at least a few moments where they genuinely became "so bad they're awesome". They actually re-enforced an idea I keep coming back to: sometime soon a real artist is going to use AI to make something amazing, not by aiming for flawless "realism" or some kind of pastiche slop, but by leaning into the weirdness that often comes out of AI.
Wow. These are horrible. Sort of refreshing. I thought video was better than this now, but I guess not.
Video can be good if you stick to a garden path of simple scenes with tons of examples in the training material and not a ton of motion between overlapping objects in a scene, and don't really care too much about specifics.
As soon as you want something very specific, or something novel, or anything with a lot of moving objects/people, it falls apart.
Video is better. The models just chose older video models which are not SOTA
Exact same impression
The fable $25 version was the best.
claude’s is bad and chatgpt’s is horrific
Wow. These are all terrible. Music video producers can breathe a sigh of relief.
As a clanker-apologist I gotta say this is the strawman of AI slop brought to life. Letting the machine do literally everything without even supervision and see how it turns out? No surprise at all the results are so bad.
My fav AI video is still Post-Scarcity Blues from a year ago https://www.youtube.com/watch?v=q_t3h2AZ0KY There have been others I've enjoyed since then. But, that one stands out in memory. Work warning: it is occasionally just a bit spicy.
These are getting really good. Much more interesting than the average music video already.
can't tell if sarcasm or just zero taste...
I only glanced at the video thumbnails and was repulsed. But then again, I'm not the target audience for this garbage.
This is my favorite AI generated (or assisted?) music video: https://www.youtube.com/watch?v=ctBpVI6lRyo