It’s pretty terrifying when you think about the possibilities of deception. And also how throwaway content is going to become. We are going to generate content at a volume orders of magnitude larger than our already current excessive volume, and finding the stuff that has real meaning and a real message is going to be even harder.
Also, artists whose work and styles fed this will be put out of business without ever being paid for their work that was used to train these models. 🫤
That sounds like hell, making money is a blast. If everything was truly equal we would all be living in extreme poverty. Global average income is $9,733 USD per year. I make that in a week, hard pass on that commie bullshit.
When I was a kid, I had seen, or at least heard of, nearly every TV show from my parent's generation. Going back probably 40 years. Like, I've probably seen every Looney Tunes, every episode of M.A.S.H., and most episodes of The Munsters, because some days there wasn't anything else to watch. My kids look at me crazy if I haven't heard of the latest flash-in-the-pan influencer, but if I bring up a 10-year old movie or TV show, they have no idea what I'm talking about.
I miss the shared culture that broadcast TV and radio gave us. Is the selection today better, with more, higher quality content? Definitely.
But all of us Millenials can quote Simpsons at each other all day even if we've never met. South park, Futurama, King of the Hill, James Bond and other corny action movies. We all saw them so many times, because that's what was on.
That shared culture is worth more than the content actually being good, IMO. Half the time now someone will ask if you've seen a show and you haven't ever heard of it.
And we will be stuck in a loop of type of art and culture that is a ouruborus feeding itself without new styles or genuine new art being fed after artists not being recognized and payed and not wanting to give more content to the machine. That dark ages are upon us and we are all singing it's praise.
If you are concerned about AI making "content" more throwaway, then you are already viewing creative works as something throwaway. Artists make works with meaning, AI doesn't have a brain, it can't make things with a meaning. That's the job of the artist.
But now, or soon, you can have one person with half an idea, like "what if The Rock had to save Shanghai from mole zombies", and they can grab a text generator to fill in most of the screenplay, and then dial in the number of synonyms for "exciting" used to describe the explosions, and come out with Day of the Living Moles, a 95 minute feature film, in a weekend. Without actually having to have had any traditional cinematography skills or breaking an artistic sweat.
There are categories of creative work that are throw-away; little sketches on napkins, improvised songs, quick sketches that an artist might think of are of no account to anyone. And the scope of what can be dashed off like that, with minimal time and effort, is growing because of more powerful tools.
Why should I watch Universal's superhero blockbuster when I can watch my buddy Jimothy's? What happens when the number of plausible films dramatically exceeds the time that movie critics have to watch them to sort out which are any good?
We spent decades depicting science fiction AIs as the key to giving humanity true freedom from mandatory labor, and now we're scared because it can do creative work too? We'll adapt. We'll be just fine. A new generation will crop up that will have no issues with AI-generated content. We're too old to see it like they will. Just like a lot of our parents and grandparents didn't understand email until they were forced to, while us kids were doing all kinds of things online.
I mean shoot, my parents still argue with me over whether electronic music is even music or not. It's just gonna be another tool in an artist's arsenal.
We spent decades depicting science fiction AIs as the key to giving humanity true freedom from mandatory labor
Very few people benefit from automation and AI. Most of us will eventually be replaced by an IA and our only freedom will be to starve (or to rebel, who knows)
Why would real meaning and messages be harder to find, does AI generated art inherently have less meaning?
Let's say I wanted to convey the message that oil companies are destroying the environment so , throwing subtlety out the window, come up with an idea of "a vampiric oil baron draining mother nature of oil", does the picture that is generated from me putting that prompt into an AI generator have any less meaning then if I actually drew it myself?
For all the advances in AI it still lacks intentionality, and always will under these current models, that has to be supplied by the person in the form of a prompt. I'd say that intention is the source of messages and meaning in art. AI just allows people without technical abilities in art to express those intentions, feelings and messages.
I can't speak for everyone, but for me personally, yes I feel like art is less interesting now. Over the past couple years or so I've found that I'm less impressed by art that I see online.
I'm not an artist, and I'm not someone who seeks out art to appreciate it. I'm just talking about art that I scroll past on the internet. I find it less interesting now. I assume that it's all AI generated, and if it's not, I figure it might as well be. It's just not interesting to me anymore. The image generated by a prompt is no more interesting or thought provoking than the prompt itself.
Now imagine that 100 oil employees make good looking ai art to show mother nature either sharing the oil with someone to help them in some way, or even make it look like oil is helping remove a cancer or something from herself. 100 different variations of this. How impactful is your message compared to theirs? Will people even see yours?
You are correct and it drives people crazy. Just consider, though, that people were saying that the web allowing anyone to publish their views as fact would undermine the averages person's ability to know what is true. It kind of did.
I don't have a hot take. I agree with you. But I also think this will change things in ways we don't fully understand yet.
We are going to generate content at a volume orders of magnitude larger than our already current excessive volume, and finding the stuff that has real meaning and a real message is going to be even harder.
It could go both ways: similar software could “compress” video (especially AI-generated video) into text prompts that could then re-create it without needing to store it. (Currently, of course, the processing cost would be higher than the storage cost for the raw video—but the scenario in which we’re cranking out excessive amounts of AI-generated content implies that the high processing costs have been eliminated.) That would also have the side effect of making it easier to find and organize videos based on their “meaning”.
I think the idea of using natural language to generate video is flawed for the vast majority of applications we want. Imagine you could give a script to one of these models and have it output a TV show episode. While we can make these models deterministic it seems like the vast majority of generative content with some amount of quality requires the addition of random noise through the process. Should we want TV episodes whose visual quality and little details shift from model to model? Why not store a plain text description infered by some model and store the video component in a medium less prone to misinterpretation? We may use deep learning compression for videos and audio in the future if there are significant advancements but I doubt the compression will be to English.
I've asked Gemini for a summary and it's pretty spot on:
This video is about AI generated videos and how they have become very realistic.
The speaker, Marques Brownlee, discusses a new AI model called Sora that can generate videos from text input. He shows examples of videos generated by Sora, including one of a woman walking down a Tokyo street, a car driving up a mountain road, and a litter of puppies playing in the snow. He points out that these videos are still not perfect, but they are much better than what was possible just a year ago.
He discusses the implications of this technology, both good and bad. On the one hand, it could be used to create fake videos that could be used to deceive people. On the other hand, it could be used to create stock footage that is more affordable and accessible than ever before. Brownlee concludes by saying that this technology is still in its early stages, but it has the potential to change the world in many ways.
TLDR: a year ago AI video was garbage. Today it’s almost as good as one that would cost a few hundred thousand dollars to pay a human production team to make (according to someone who’s professional work is creating those videos).
It’s not quite there - hands glitch out occasionally. Sometimes animation doesn’t quite line up right (e.g. walking might skip a step) but it’s 99% there and and the improvements over the last 12 months are astounding. That last 1% surely won’t take long to close.
There was a landscape drone video from a helicopter that looked absolutely real.
Note this is not publicly available yet - OpenAI said they are still working on safety features to reduce the risk of it being used to create content that they want no part in.
Red Team is a hacking term that refers to people who try to sabotage or use the system to create harmful content, as a way to test and discover problems before it is usable by any external users.
That's inaccurate. Red Team is the guys that test your security from an attacker view point. Red Teams are often contractors hired by companies. The companies are the ones paying to be "hacked", so they can fix whatever gaping security holes the red Team finds.
At least, that's usually the definition. If just talking about AI stuff, I'd call those people testers.
I'm really excited for this. This way, converting my favourite webtoons to full blown animations won't be that difficult (in the sense that it won't cost millions of dollars). Really exciting times!
That requires vastly more work to produce any results at all, to the point that most animation people might want to produce never gets made because the process is far too expensive. Mediocre animation that gets made using AI tools is better then high-quality animation that never gets made at all.
Blender and AI tools both have their place but they're not interchangeable. And just wait until Blender starts incorporating AI, which it will, because the purpose of something like Blender is to use computers to automate most of the work that would need to be done with previous generations of tools, and AI is just an extension of that. Animation will exist on a continuum from fully handmade artwork to fully machine generated artwork. Unless you think everything should be drawn by hand one frame at a time, you should be happy about everyone being able to produce animation in a way that suits their skill level and the amount of time they have available.
Not necessarily. Fine-tuning models can solve this issue to a great degree. The model's behavior is largely dependent on its training data. If it has generic training data, it's going to produce generic images.
See Corridor crew's anime experiment. They managed to solve this issue to a great degree in their second version. It's quite cool!