So, less than 24 months after the advent of consumer AI, we’re probably the most explosive technological disruption ever seen, and the successes keep coming.
This week, Adobe announced that it would integrate artificial intelligence technologies, adding OpenAI’s Sora, into its Premiere Pro video editor, and you can almost hear the collective explosion of horror spreading over the Hollywood hills.
This is probably a risk-free ad, just a set of video editing tools, much like plugins for Photoshop. But the banal press release hides a total universe, my God. Global art is still recovering from the triple attacks on the old order.
On the one hand, the fashion art and inventory photography sectors are dissolving before our eyes. Stable Diffusion, DALL-E, and Midjourney have replaced regulations for good.
If you think I’m kidding, take a look at the new Stockcake site, which now offers millions of AI-generated inventory photographs in exchange for a public domain license. How’s it going, Getty Images?
And if you think music doesn’t happen the same way, check out Udio, which allows anyone with a pulse to create entire songs in an infinite variety of genres. It’s like they’re sitting in a music studio next to Prince. .
The last artistic bastion of video began to crumble with Open AI’s Sora preview, which proved that video can kill the radio star, if you have a computer tough enough.
It’s as revealing and surprising as we all imagined at first, but it’s also not as scary or threatening.
But this edition of Adobe Firefly puts the icing on the cake. Many other people thought that the sheer computing power required to manipulate video was an obstacle that would take years to overcome. Well, yes. Of course not. Image and music generation has triumphed over garage and CPU hurdles over the past couple of years, and we’re about to see video follow the same trajectory.
But this time, the stakes are enormous. We’re not just talking about movies, we’re talking about prime-time television, advertising, cable, and plays. If we can edit a component of the video, we can create a video. And if we can create a video, you can create a movie or a TV episode. It’s a scary sauce for some people. But that’s not going to happen tomorrow, or even in the next few years.
Tim Webber, Oscar-winning visual effects manager and CEO of visual effects company Framestore, believes AI is still primitive. “It’s not as revealing and surprising as we all imagined at first, nor is it as scary and threatening,” he says. We’re somewhere in between. “
Webber, whose credits include Gravity and The Dark Knight, estimates it will take about five years to make generative AI the number one tool for high-quality moving images.
By the way, in addition to Moore’s Law, this type of disturbance takes 10 years. We are in the seventh year since the serious arrival of LLM transformers.
He was engaged in a serious verbal exchange with film critics on the way to the Cannes Film Festival, when virtual films were on the brink of ruin with the Star Wars franchise. I was the only geek present, and those other passionate people, who knew which movie logo The Godfather was shot in, insisted that virtual filmmaking would never take off.
I knew it in a different way and patiently explained why, with some mockery. I knew that this generation walks on tiptoe, masquerading as something simple, like a device or an app. Next thing you know, he bursts in and takes over the whole party.
As we now know, the virtual didn’t destroy Hollywood. In many ways, it made it again, bigger and better. My case stood. By the way, in addition to Moore’s Law, this kind of disruption has been going on for 10 years. We are in the seventh year since the serious arrival of LLM transformers. Read whatever you want.
The important thing not to forget is that new generative AI equipment is at its most rudimentary lately. But generation moves like a high-speed charging exercise, and anyone or anything that gets in the way is very likely to end up like Wile. E. Coyote at the Road Runner Show. But is that a bad thing?
If you’re a purist, like the 35mm film enthusiasts I’ve met, then yes, AI is going to destroy the art of filmmaking. But for most people, it’s most likely a wonderful tool for exploring new and ambitious experiences.
There’s been a lot of discussion lately on the Udio Discord forums about whether the generated music tracks are true “works of art” or not. This kind of thing is as old as humanity (just ask young Tracey Emin) and the answer is. . . . I anticipate that there will be many more such debates in the next five years.
Paul Debeyec, a senior scientist at Google, laid it all out succinctly in a 2020 article in The Hollywood Reporter. “Decades from now, an AI set of rules will make your film based only on script text, even the style of a specific director. You can even ask them to make your new movie, or remake an old one, in the style of Kubrick, Tarantino or Scorsese, emphasizing the correlations between their scripts and their films. “
Now he’s a talking geek. At the end of the day, my main confidence is that Hollywood, Cannes, and anywhere else where art exists will find a way to integrate AI as just one component of this perpetual human delight we call culture.
I tried Google Gemini vs ChatGPT vs MetaAI: which chatbot generates the images?
AI avatars can now smile thanks to a new style of Synthesia, and it’s as unsettling as it gets
This Oscar-winning sports drama just snuck into Netflix’s top 10, and it’s 90% on Rotten Tomatoes.
Tom’s Guide is from Future US Inc. , a leading foreign media organization and online publisher. Visit our corporate website.