After doing some testing, Imagen 4 doesn't score any higher than Imagen 3 on my comparison chart, approximately ~60% prompt adherence accuracy. https://genai-showdown.specr.net
They failed but man those snakes are cool. Awesome website!
Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
461–470 of 570 posts
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#462Earlier quoted context omitted.
There's nothing creative in having someone or something else doing the work for you. "Creating" with an AI is like an executive "inventing" the work actually done by their team of researchers. A team owner "winning" a game played by the their team. That being said, AI output is very useful for brainstorming and exploring a creative space. The problem is when the brainstorming material is used for production.
Buñuel would disagree with you: "The peak of film-making will be reached when you are able to take a pill, switch off the lights, sit facing a blank wall and project on it, directly from your eyes, the film that passes through your head."
No need for any pills. Such tech exists for centuries and is colloquially known as "reading books". It requires some lighting though
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#463Google's been coooooking
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#464An indie film with poor production values, even bad acting can grip you, make you laugh and make you cry. The consistency of quality is key - even if it is poor. The directing is the red thread throughout the scenes. Anything with different quality levels interrupts your flow and breaks your experience. The problem with AI video content at this stage is that the clips are very good 'in themselves', just as LLM result…
You don't need to make an entire movie out of this. One or two scenes that are difficult or impossible to film on a certain budget is enough to lift the production value of a movie. One can use this as CGI replacement, for example to produce a couple seconds scene of an ancient city and stretch that out with fake panning. You can also use it as a communication tool such as making a "live" storyboard to prep location,…
That storyboard idea is pretty huge. Imagine dailies go in the other direction - "here's how I want it to look"
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#465Earlier quoted context omitted.
And google is in the best possible position to detect it if they want to exclude it from their datasets.
They're never going to manage to do that, just on a technical level Plus some users might want to legitimately upload things with AI-generated content in it
I somewhat doubt that YT cares much about AI content being uploaded, as long as it’s clearly marked as such.
What they do care about is their training set getting tainted, so I imagine they will push quite hard to have some mechanism to detect AI; it’s useful to them even if users don’t act on it.
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#466I'm excited about this. Think of all of your favorite novels that are deemed "impossible" to adapt to the screen. Or think of all the brilliant ideas for films that are destined to die in the minds of people who will never, ever have the luck or connections required to make it to Hollywood. When this stuff truly matures and gets commoditized I think we are going to see an explosion of some of the most mind blowing ar…
It's already difficult enough to make a successful book adaptation, even WITH authorial intent. Can't imagine that hours of patchwork AI-generated video, with all its artifacting and consistency errors, will fare any better than "The Rings of Power".
I think not yet, but it is coming.
I can see it using some form of PEFT so that the output becomes consistent with both the setting and the characters and then it is about generating over and over each short segment until you are happy with the outcome. Then you stitch them together and if you don't like some part you can always try to re-generate them, change the prompt, ...
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#467Earlier quoted context omitted.
Arguably OLED screens and IR cameras are no paradigm shift. At least nothing comparable to "no smartphone" to iPhone 1.
There were smartphones before the iPhone. One could also describe the difference as "just a touchscreen".
The iPhone 1 featured "touch screen, GPS, camera, iPod, and internet access. Its software capabilities were a turning point for the smartphone industry" (random source: https://www.textline.com/blog/smartphone-history).
If you want to doubt that it was in fact a not a turning point you'd need to provide very strong arguments.
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#468Earlier quoted context omitted.
And google is in the best possible position to detect it if they want to exclude it from their datasets.
They're never going to manage to do that, just on a technical level Plus some users might want to legitimately upload things with AI-generated content in it
[dead]
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#469Earlier quoted context omitted.
Even if the content was 100% AI generated (which is the furthest thing from reality today) human engagement with the content is a powerful signal that can be used by AI to learn. It would be like RLHF with free human annotation at scale.
Won't the human engagement be replaced by AI engagement too? if it isn't already being replaced?
Yes it will. Soon humans will be the minority on the internet. I wrote some guesses about this 2 years ago: https://art.cx/blog/12-08-22-city-of-bots
Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
#470I'm surprised no one has yet to mention the use of the name "Flow", which is also the title of the 2025 Oscar winning animated movie, built using Blender. [1] This naming seems very confusing, as I originally thought there must be some connection. But I don't think there is. [1] https://news.ycombinator.com/item?id=43237273
For sure seems they are likely deliberately riding on the fame of the movie. I too instantly thought it is some kind of Flow movie animation collaboration similarily like Flow is represented in Blender 4.4 splash screen or is even their mascot.