$100 AI Music Video: Claude Fable 5 vs. GPT-5.6 Sol
tryai.dev344 points by hershyb_ 20 hours ago
344 points by hershyb_ 20 hours ago
Like most of this stuff, it's obviously impressive technology compared to what existed a few years ago. But the end product has zero artistic value. It's a grey goo of the average of every concept picked up from the concept of the song.
A talented creative with a vision could make something more interesting and enjoyable in an afternoon with a $0 budget.
As much as I don't want jobs taken away by AI, a creative person with a vision that would usually required $1 million budget, could create such a video using AI by being very specific in what they require which would not be a " grey goo of the average of every concept "
eg. "scene 32, same 2 characters from the previous scene, now dressed in the garb of Egyption middle kingdom high priests, though with cat faces, dancing on the back of an elephant that is running fast with blurred feet on the surface of a river with licks of water that burst from the water in time with the music that I gave you before, length of song for this clip 2:21 to 2:31"
Your mind is your limit. AI is a tool. If you tell it "create me a music video for this song set in egypt" then it may turn out AI sloppy.
Art isn't just the screenplay, or a prompt; it's the clash between reality and the intent (whether a location or practical set constraints), the individual skill of everyone involved, costume and prop department. And it may change during production. And effort. If it's easy, it's easily reproducible and not worth that much. And novelty, that's something current models aren't capable of. That's why Jarimoquai's Virtual Insanity video became popular. That's why practical effects are more impressive than CGI, even if they're jankier and hide behind clever camerawork.
If you've ever tried to get around Gemini or ChatGPT's guardrails, stock footage poisoning, and just general tendency to produce the most frustratingly banal version of your prompt possible, then you would understand that prompting AI images and video absolutely already involves a clash between reality and intent (in the sense that this is Google and OpenAI's world, we're just living in it).
I would love to get around the "individual skill" part, but all of my artist friends would excommunicate me if they knew the AI projects I've been working on, let alone if I asked them to collaborate.
> it's the clash between reality and the intent (whether a location or practical set constraints), the individual skill of everyone involved, costume and prop department
That's not art, that's the limitations that everyone working in this industry have been working night and day to overcome for hundreds of years
A notable chunk of criticism GenAI receives has to do with output irreproducibility and semantic instability relative to the input, so this is a bit entertaining to read.
If there's a way to do something better and worse, then one can absolutely talk about added and absent value. I browse AI generated images a ton, and the difference between beginner / low effort submissions and high effort / advanced submissions is very stark.
I think you are making the same mistake as everyone else with respect to art.
Art has nothing to do with _mechanical_ difficulty. I see this misconception all the time. Examples
- Kumail Nanjiani roided up for his next movie. This has mechanical difficulty, sure. But what does it add to the artistic element?
- Dream Theatre guitarists play olympics with their guitar and play solo's that are mechanically impossible for normal people. Yet we still find Beatles have more artistic value, why?
I hope this popular misconception will die down. I don't want mechanical difficulty in art being praised. I feel these are things people hold on to because mechanical difficulty has some moat and people don't want to give it up.
I don't think that was fair comparison, earlier comment was that creative person can make million dollar art without having million dollar budget, then again creative person can splash paint on canvas while being dead drunk and make most valuable painting.
Most people find it more impressive if you learn to play Dream theater song on guitar than if you program it in daw.
I partly agree with you - I used to think of mechanical difficulty as a kind of 'proof of work' for art. Nowadays I am less interested in the idea of what counts as art or not.
However, consider works like photorealistic drawings of eyes (usually at a very large scale). A lot of people like these, and consider them to be fine art, while others (art critics?) consider them bland. Contrast that with a Picasso flamingo drawn with a single line.
There is artistic value in high-effort, low concept works as well as low-'effort', high-concept works and other combinations. Perhaps what we are all looking for is effort of some kind (I prefer human effort) whether that is conceptual effort or practice in drawing lines on paper.
I’m curious to understand why you think drawing photorealistic eyes is different from a person beating a Guitar Hero time record.
Both are pushing boundaries in mechanical dexterity. Yet one is considered art and another is not.
If I was feeling unfair I would say there is no difference! However, if a friend drew me an eye on A3 paper would I frame it? Probably yes...
Others have said this better, but art is an qualitative/emotional thing, not a simply measurable quantity of an object. We tend to use the effort put in as a proxy measure for quality, but when photocopiers and 3D printers and - yes - genAI get involved, that becomes harder or impossible.
Beating Guitar Hero is emotional, but not artistic. Photorealistic drawings could be emotional, and may or may not be artistic. Putting a text prompt into a generative model to make a (frankly terrible) music video for $100 is ... something else.
Anything can be artistic, but some things are very hard to make artistic.
Beating Guitar Hero could artistic. I don't even know how, but I'm sure some artist could beat Guitar Hero in a way that was somehow artistic.
This terrible $100 music video which I'm not even going to watch, is surely not art. Also I'm totally sure some person is going to spend $100 with an AI and make art. It's very hard, but they day will come. The slop-to-art ratio will be astronomical.
[flagged]
The problem with this discussion is we are conflating artistic value with economic value. A music video can be valuable as art, but it can also be valuable as a tool for promotion or for generating revenue. If the goal is pure economic efficiency then an AI has the potential to create a music video more efficiently than a human. If the goal is to produce something that makes people feel some kind of emotion then AI will work against you. The same goes for code. I use AI to write code that I'm not precious about but if I want to feel proud of my work, you can be sure I'll write it by hand.
[flagged]
The physical limitations work with you, not against you. Have you ever read a book? The physical limitations is the blank paper, you need to be creative enough with the words you need to express for your vision to come to life.
Given that all serious artists seem to be extremely anti-AI, I strongly suspect that you may be wrong.
I've been trying Fable for coding in the past week and while it's incrementally better than Opus sometimes, what it's churning out is still far from art - and I do think it's possible for code to be art.
If you're defining "serious artists" as some approximation of "people who have made money with their art" then it's clear that there's a conflict of interest here.
Now I happen to be in the camp that believes that AI on its own isn't really capable of producing art (but like a human creating a collage, there can be art created by humans using AI... there is artistic merit in the arrangement).
I still despise that AI is going to be used to strip creators of their labour value, but that's more related to my objection to capitalism than it is to AI.
I'm simply defining serious artists as people who are serious about their art, whether they're award winning filmmakers or amateur photographers. Roughly nobody out of all those people wants to use AI.
(If you want to define it as professional artists instead, then I don't even see much of a conflict of interest. Professional programmers tend to be cautiously optimistic about the new tool and are finding uses for it, why wouldn't professional artists? But, they largely do not see it that way.)