For several years now, generative AI has been actively present in our lives, and we still haven’t decided how to treat it. Some are incredibly excited to test each new model, others are desperately afraid that machines will soon take away their jobs, and some, like Elon Musk, threaten to not let employees into their offices with ChatGPT on their newest iPhones.
But so far, the most striking manifestations of AI are numerous variations of modifying already existing content. For example, “the behind-the-scenes life of Hogwarts,” as we’ve told you here. Well, the advent of magical moving photographs, like in the books about Harry Potter, is not far away… Oh, no, wait – they are already here!
More info: X
#1
#2
To briefly describe what happened, this week Luma presented its new neural network for generating video from text queries, called Dream Machine. According to the developers, the model understands well how people, animals and various objects interact with the world around them, so that the videos are realistic, with smooth and seamless physics.
Before the advent of Dream Machine, the absolute leader in the field of generative video models was the famous Sora from OpenAI, but Dream Machine has an undeniable advantage – it is accessible to a wide range of users, here and now. And it even offers to generate 30 videos per month for free.
#3
#4
“Luma’s Dream Machine was just released this week and has immediately become the most loved AI video generator. Why? Other than OpenAI’s hyped Sora or Google’s Veo, Dream Machine is available to everyone. And it is qualitatively superior to the previous leading models of Runway or Pika,” says Boris Eldagsen, the German photographer and AI expert, with whom Bored Panda got in touch for a comment here.
For example, just yesterday, right after reading the news about the model presentation, I started generating a video using the simplest prompt: “Panda typing on a typewriter sitting at the table near the sea.” The result is here and it’s damn impressive. At least to my untrained eye, the video looks just as good as Sora’s. And whenever there is a new free tool, there is always a desire to do something cool with its help.
#5
#6
Why not, for example, revive classic memes of the past – from the girl on fire to the Pope in a down jacket? To find out what happened after the distracted boyfriend turned to look at the girl in the red dress? To try to understand what the woman was trying to prove to the cat, and why Harold still hid his pain? No sooner said than done!
“Memes are considered an important part of Internet culture that spreads by means of imitation,” Boris Eldagsen comments. “Representing an idea, they have no fixed form and circulate as mutations, parodies, remixes. As such, the animation of a meme image is a logical development.”
#7
#8
It’s funny, by the way, that the animated meme about the Pope looks, let’s say, like fractal AI art – after all, the original picture was also generated by artificial intelligence. And, well, this is perhaps the first example in human history of an AI-made image becoming a global meme. Or not only humanity anymore?
#9
#10
“Is an animated meme ethical? The Greek word “ēthos” originally meant “custom, habit” – which shows that ethics depend on a particular group, tribe or society. That means, each group must decide for itself. But all groups must adhere to their respective national laws,” Boris Eldagsen ponders.
“So, the better question is: Is an animated meme legal? A long time, it has been debated if memes deserve fair use protection or if they are copyright infringement. One argument pro “fair use” was that memes are “transformative” (that is, giving novel meaning or expression or a new context to the original work).
Animated memes fulfill this criteria for sure. From a 4s (25 frames per second) animation, only the first frame is taken from somebody or somewhere else. The remaining 96 frames are AI-generated and as such “hallucinated”. Is this legal or not? We don’t know. The area of legal uncertainty caused by AI is growing constantly.”
#11
#12
Be that as it may, the videos that you can watch in this collection are still the results of a prompt made by a living person, made of flesh and blood. Skin and bones as well. So it’s too early to sound the alarm that AI has entered the territory of humor, which was originally inherent only to us humans. But the most creative among us have received an excellent apprentice…
