Comment on Disney's Sora Disaster Shows AI Will Not Revolutionize Hollywood
mindbleach@sh.itjust.works 5 hours agoYou’re having a much more contentious conversation inside your head. Please stop projecting traits you’d rather be arguing against, when someone points out, it does the thing it’s for. It demonstrably functions. You could make a coherent moral argument about how it was made - but you haven’t. You’ve railed against an imaginary frothing psychopath, because someone politely described utility.
How it was made is addressable, by the way. It’s fixable. There will be vegan models made from bespoke, licensed, and public-domain data. Will that change your opinion in any way? Or is that complaint decorative?
supersquirrel@sopuli.xyz 5 hours ago
Stop referencing promises to prove your point, you sound like a door to door salesperson.
mindbleach@sh.itjust.works 4 hours ago
Okay, here’s reality from the recent past: some guy recreated GPT-2 for $20. Same size, similar training data, equal performance. The original required VC funding. This guy spent pocket change. That was a year ago. That’s how much efficiency has already improved, for training these models. These assholes only spend billions because it’s exclusionary, and they’re all caught in a dollar auction to see who can lose the gentlest. I’m sorry any hypotheticals about that are incompatible with your moral crusade.
Meanwhile, it does the thing.
That’s not going to change and you kind of have to deal with it. We now have programs that just do what you ask, for any output that’s text, images, audio, or video. They often fuck up in horrifying ways. But they’re usually about what you asked for. Especially if you asked for very little. That’s quite useful where small changes are wildly complex, like ‘make this guy look like another guy.’ The robot won’t do it as good as a team of human professionals, but I don’t have a million dollars to hire a team of human professionals, and I’m betting you don’t either. You can still consider projects that involve making one guy look like another.
That utility is new and it’s not going anywhere.
supersquirrel@sopuli.xyz 4 hours ago
I am an artist so I understand when I have the shallow desire to make something into a copy of another thing and my artistic capability fails me, or my lack of resources confines me from reaching my initial vision, that this is the true beginning of my artistic journey and all of that stuff before was just a way of backing myself into wanting something new or changed when I couldn’t get the perfect thing I wanted that was in my head.
I have done lots of community theater so I also understand the foolishness of thinking that the important part of making one thing look like another is aesthetic mimcry and not of capturing the elegant essence of something.
Do you think for all these years everybody watching Shakespeare plays has been getting a suboptimal experience because the two actors that are meant to play characters easily mistaken from each other aren’t actually perfectly convincing?
Image
Do you care that in Hamnet that two siblings that are supposed to look so alike that they are frequently mistook for one another, even by death itself, don’t actually look that similar? No, they are child actors who did an amazing job, to care about that in the context of the achievement Hamnet is, is to be shallow and miss the point.
My point is that even when AI is good at particular things, often the whole approach is hollow to Why? with AI. This is something artists could have explained easily to techbros if they ever listened, because the Why? is the whole point.
mindbleach@sh.itjust.works 3 hours ago
Denouncing the pursuit of verisimilitude is a novel response to hand-wave CGI. Are you this philosophical when a movie does spend a million dollars, to make two unrelated actors look exactly the same? Do you expect audiences should be happier if a no-budget sci-fi film has cardboard displays? It’s cute, certainly. But when a central complaint is that people will notice generated elements and object to low quality, I think they’re gonna notice literal cardboard.
Films are photographs. That’s why The Social Network didn’t just say the Winklevii were twins and expect people to pretend. Movies are a visual medium, whereas theater is mostly heard. Like how television has viewers but theater has an audience. You can Dogville it, and people will roll with that, but anything that looks fake is more commonly a technical failure than a stylistic choice.
So yes, you can tell people the tin can is a spaceship… but they’d rather be shown. The preference for showing over telling is so ingrained that it’s cliche. Nobody needs to announce ‘we lay our scene in fair Verona’ when you can put the mediterranean coastline onscreen, and then cut to a cobblestone village where people have pointy shoes. Folks will get it. They’ll get it on a level deeper than narration, or an overlay reading “Verona, Italy, 15° E, 40° N, June 17th 1435, 0700 hours.” They’ll get it even if the aerial shot of the coastline was bought as stock footage. Or rendered, in one way or another.