Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

blog.google

Post reply on HN
Print view View original

271–280 of 570 posts

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#271

Post by alickz »

alpaca128 wrote: post #138
ahmedfromtunis wrote: post #88

Earlier quoted context omitted.

The amount of gatekeeping I see when this topic is brought is outstanding! Why can't people be happy that more individuals would be soon able to create freely in a more accessible way? Personally I can't wait to see the new creative doors ai will open for us!

How is the requirement to use a computer and maybe pay a cloud subscription in the long term more accessible than other kinds of art? Which individuals are gatekept exactly? Before you bring up disabled people (as often happens when the term accessibility is used), know that many of them are not happy to be used as a shield for this without ever being asked and would rather speak for themselves. I've tried AI image g…

>How is the requirement to use a computer and maybe pay a cloud subscription in the long term more accessible than other kinds of art?

Because many other kinds of art require thousands of hours to learn before getting to the level of current AI

The real gate keeper to art isn't the cost of a pencil, it's the opportunity cost of learning how to use it

Some people have creative ideas they cannot realise and tools like AI help them do it. The more people that can realise their creative ideas the better it is for everyone.

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#272

Post by fragmede »

danpalmer wrote: post #236
woolion wrote: post #99

Earlier quoted context omitted.

The winning image entry for "The Yarrctic Circle" by OpenAI 4o doesn't actually wields a cutlass. It's very aesthetically pleasing, even though it's so wrong in all fundamental aspects (perspective is nonsensical and anatomy is messed up, with one leg 150% longer than the other, ...). It's a very interesting resource to map some of the limits of existing models.

In my own testing between the two this is what I’ve noticed. Imagen will follow the instructions, and 4o will often not, but produces aesthetically more pleasing images. I don’t know which is more important, but I would say that people mostly won’t pay for fun but disposable images, and I think people will pay for art but there will be an increased emphasis on the human artist. However users might pay for reliable to…

People pay for digital sticker packs so their memoji in iMessage are customized. How much money they make on sticker packs is unknown to me, but image generation platform Midjourney seems to be doing alright.

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#273

Post by a2128 »

Legend2440 wrote: post #120
imp0cat wrote: post #105

Earlier quoted context omitted.

Would Google really stoop so low and try to use the success of the movie to prop their AI video generator tool? But then again, the do no evil motto is long gone, so I guess anything goes now?

It's a common word. There are like 50 things named Flow. It's unrelated.

Before picking a name it's necessary to Google it and make sure you're not squatting on anything important. It's hard to believe that they didn't find they're about to squat on an Oscar winning animated film less than a year after its release. They decided to roll with it anyway, for a tool that basically aims to eliminate animators and filmmakers

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#274

Post by themacguffinman »

tootie wrote: post #220
lenerdenator wrote: post #94

I do find myself wondering if the people working on this stuff ever give any real thought to the impact on society that this is going to have. I mean obviously the answer is "no" and this is going to get a bunch of replies saying that inventors are not to blame but the negative results of a technology like this are fairly obvious. We had a movie two years ago about a blubbering scientist who blatantly ignored that to…

Remember when they fired Timnit Gebru for publishing on AI safety?

Quite a narrow view to interpret what happened there as firing Gebru for publishing on AI safety. Google still conducts and publishes research on AI safety, just without Gebru who helpfully offered to resign if Google didn't name her critics.

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#275

Post by alickz »

gamblor956 wrote: post #197
andoando wrote: post #124

Earlier quoted context omitted.

I don't think that's true. Is a film director not a creative? You could come up with your own story and direct the AI to generate it for you.

In your example, the "come up with your own story" part is the creative part. But you're not "directing" the AI to generate it for you. You're just giving it a command. You're selecting from the results it outputs, but you're not controlling the output. A film director is a creative. Ultimately, they are in charge of "visualizing" a screenplay": the setting, the the design of the set or the utilization of real locati…

>But you're not "directing" the AI to generate it for you. You're just giving it a command.

That's what direction is though. Film directors prompt their actors and choose the results they like best (among many other commands to many other groups)

>You're selecting from the results it outputs, but you're not controlling the output.

The prompt controls the output (and I bet you'd have more control over the AI than you'd have over a drunk Marlon Brando)

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#276

Post by alpaca128 »

Clamchop wrote: post #173
alpaca128 wrote: post #138

Earlier quoted context omitted.

How is the requirement to use a computer and maybe pay a cloud subscription in the long term more accessible than other kinds of art? Which individuals are gatekept exactly? Before you bring up disabled people (as often happens when the term accessibility is used), know that many of them are not happy to be used as a shield for this without ever being asked and would rather speak for themselves. I've tried AI image g…

There's a pretty standard argument that creating artworks should or must require the hardship of developing the skill to bring vision into reality, or paying someone who can. That can be debated, but the position is textbook gatekeeping. Other disapproval comes from different emotional places: a retreading of ludditism borne out of job insecurity, criticism of a dystopia where we've automated away the creative experi…

Gatekeeping means restricting power or opportunities. Is it gatekeeping when a skill can be learned for free, and on top is also optional? Is it also gatekeeping to forbid students from using ChatGPT for homework because it would force them to go through the hardship of earning an education?

I have seen what 99% of people are doing with this "clear and present value". Turns out when you give people a button to print dopamine they probably aren't going to create the next Mona Lisa, they're just going to press the button more. Even with AI, creating compelling art is still a skill that needs to be learned, and it's still hard. And why would they learn a skill when they just decided against learning a skill? Incentives matter, and here the incentives massively favor quantity over quality.

> Whether AI has worked well for you isn't just irrelevant

My point was that it's creatively restrictive, and that current models tend to actively steer away from creative outputs. But if you want to limit yourself to what corporations training the models and providing the cloud services deem acceptable, go ahead.

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#277

Post by vunderba »

xixixao wrote: post #56
vunderba wrote: post #22

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

Awesome showcase! Fun descriptions. Are there similar sites?

Thanks! There are definitely other GenAI image comparison sites out there - but I found that the majority of them were more concerned with visual fidelity which IMHO is a less challenging problem than prompt adherence.

This is probably one of the better known benchmarks but when I see Midjourney 7 and Imagen3 within spitting distance of each other it makes me question what kind of metrics they are using.

https://artificialanalysis.ai/text-to-image

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#278

Post by Klonoar »

ahtihn wrote: post #185
duped wrote: post #164

Earlier quoted context omitted.

> Why can't people be happy that more individuals would be soon able to create freely in a more accessible way? The gates are wide open for those that want to put in effort to learn. What AI is doing to creative professionals is putting them out of a job by people who are cheap and lazy. Art is not inaccessible. It's never been cheaper and easier to make art than today even without AI. > Personally I can't wait to se…

> The gates are wide open for those that want to put in effort to learn. Why is effort a requirement? Why should being an artist be a viable job? Would you be against technology that makes medical doctors obsolete?

> Why should being an artist be a viable job?

Because we as a society have valued it as one for eternity.

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#279

Post by fragmede »

viraptor wrote: post #268
tintor wrote: post #140

Earlier quoted context omitted.

More difficult examples: - wine glass that is full to the edge with wine (ie. not half full) - wrist watch not showing V (hands at 10 and 2 o'clock) - 9 step IKEA shelf assembly instruction diagram - any kind of gymnastics / sport acro

What's the reason to test the "not showing ..."? I've never seen anyone make that kind of request in real life. They ask for what they actually want instead. You'd ask for a clock showing 3:25 rather than "not 10:10". I mean, it's a fun edge case, but I'm practice - does it matter?

> I mean, it's a fun edge case, but I'm practice - does it matter?

*in practice, not I'm practice. (I swear I have a point, I'm not being needlessly pedantic.) In English, in images, mistakes stick out. Thus negative prompts are used a lot for iterative image generation. Even when you're working with a human graphics designer, you may not know what exactly you want, but you know that you don't want (some aspect of) the image in front of you.

Ie: "Not that", for varying values of "that".

Re: Veo 3 and Imagen 4, and a new tool for filmmaking called Flow

#280

Post by cynicalpeace »

Basic principles:

1. People like to be entertained.

2. NeuralViz demonstrates AI videos (with a lot of human massaging) can be entertaining

To me the fundamental question is- "will AI make videos that are entertaining without human massaging?"

This is similar to the idea of "will AI make apps that are useful without human massaging"

Or "will AI create ideas that are influential without human massaging"

By "no human massaging", I mean completely autonomous. The only prompt being "Create".

I am unaware of any idea, app or video to date that has been influential, useful or entertaining without human massaging.

That doesn't mean it can't happen. It's fundamentally a technical question.

Right now AI is trained on human collected data. So, technically, It's hard for me to imagine it can diverge significantly from what's already been done.

I'm willing to be proven wrong.

The Christian in me tells me that Humans are able to diverge significantly from what's already been done because each of us are imbibed with a divine spirit that AI does not have.

But maybe AI could have some other property that allows it to diverge from its training data.

Post reply on HN