Author Topic: Open AI Sora text to video - unbelievable new AI video tool (Read 13477 times)

Andrej.S. · « **Reply #50 on:** February 29, 2024, 11:46 »

@stocker2014
Would not bet on that in 3 years.

I wonder if Kodak thought similarly about digital photography back then. Just a passing trend and fantasy.

Of course, there are already cloud-based solutions.
But they have the current disadvantage that they are not yet user-friendly and you need some IT knowledge (at least basic phyton or other programming skills) to set up your own AI models.

Just mark my words that the downfall will start with mature multimodal models in the next 3 years.

As soon as users start using them en masse and can generate almost free content for their needs in real time with easy prompts, no one will want anymore trawling through millions of generic videos in agencies libraries if they don't adopt their own AI assistants.

Do you really believe that users will then still buy generic videos of for example moving waves at the beach during a sunset for 50 bucks and have to search a matching one for their needs when they can generate this content by prompting "create a blog article of travel tips to the maledives and generate a matching video".

Time is money.

stoker2014 · « **Reply #51 on:** February 29, 2024, 14:22 »

Quote from: Andrej.S. on February 29, 2024, 11:46

Do you really believe that users will then still buy generic videos of for example moving waves at the beach during a sunset for 50 bucks and have to search a matching one for their needs when they can generate this content by prompting "create a blog article of travel tips to the maledives and generate a matching video".

Yes, that's exactly how it will be. I don’t know about $50, but now most buyers buy videos on subscription services. There they will buy by subscription in 3 years, 5 and 10 years.

stoker2014 · « **Reply #52 on:** February 29, 2024, 14:26 »

Quote from: Andrej.S. on February 29, 2024, 11:46

As soon as users start using them en masse and can generate almost free content for their needs in real time with easy prompts, no one will want anymore trawling through millions of generic videos in agencies libraries if they don't adopt their own AI assistants.

And spaceships will fly to Mars and other galaxies! Everything will be that way!

All these technologies are not made so that people can use them for free. The buyer will have to pay money either for cloud power or for a powerful PC, and all this will cost money. Therefore, anyone who is not a fool will go to the subscription stocks, pay $16 a month for a subscription, and download everything he needs.

cobalt · « **Reply #53 on:** February 29, 2024, 14:30 »

The problem is getting something very specific with only your prompt.

And most people don‘t even know what they are really looking for, when they start creating.

I think most agencies will offer you a hybrid search - you add a prompt or description for ai and at the same time you will be offered the ai creation on your prompt as well as general results from search terms.

then you can pick a few files from either ai or the database and refine your prompt again.

Which gives you new hybrid results…etc…

Already today some of my best ai images come from using one of my own files as a starter image.

the need to look at a huge database for „inspiration“ will not go away.

it will probably also be combined with prompt description service. if you choose a file from the stock database you will be given prompt suggestions for creating something similar for yourself.

ai creation and a stock database are not exclusive opposites, they will be used together to make life easier for the customer.

stoker2014 · « **Reply #54 on:** March 01, 2024, 09:37 »

https://youtu.be/TWSaI9twba8

Will New AI Video Tool Sora Kill Microstock? When Will Sora Release and Be Available?

Andrej.S. · « **Reply #55 on:** March 01, 2024, 15:14 »

@cobalt
The hybrid approach is a best-case scenario if the agencies manage to integrate an AI assistant that is able to find the right image or video for the user's purpose in a very short time.

The AI assistant should completely replace the current searchbar with keywords.
Ideally, you would then enter a text description, voice message or screenshot of your website, product, article, etc. as input.
It must be a time saver.

But the crux however is, when AI image and video generators will be so much advanced as a part of multimodal models, that they would be still more time-efficient and above all more precise for the most generic stuff.

And the second problem is that agencies will probably have to invest much money for own development of AI assistants. Perhaps they will try to license ChatGPT oder Gemini, who knows. So agencies will have more expenses while the revenues will drop and they can't cut the contributors' commissions that much anymore.

Believe me, it is a naive idea to think that prompting will remain it's current complex structure. Take a look at the videos of Gemini. The multimodal model can already interpret complex visual context almost in real time.
And that's much more complex than just a text description.

@stoker2014

I don't say that the big agencies will completely vanish but just that the demand of generic content will probably drop by about 90%.

I think you underestimate the possibilities of multimodal models that for sure would be integrated into an all-in-one suite as a subscription.

Midjourney as a single tool for generating images is not a viable business model for the future. It is far too expensive for a single purpose. Most current users just try to generate and sell images via agencies that's the current irony.

But if an all-in-one solution for texts, images, videos, coding, etc. will be placed on the market for perhaps 49 or 79 USD/month, which is on every content type at the same level or above like Midjourney or Stable Diffusion for image generation, then you can think twice about whether someone will still buy a subscription with the agencies.

stoker2014 · « **Reply #56 on:** March 02, 2024, 09:05 »

Quote from: Andrej.S. on March 01, 2024, 15:14

But if an all-in-one solution for texts, images, videos, coding, etc. will be placed on the market for perhaps 49 or 79 USD/month, which is on every content type at the same level or above like Midjourney or Stable Diffusion for image generation, then you can think twice about whether someone will still buy a subscription with the agencies.

The key word here is if

Artist · « **Reply #57 on:** March 03, 2024, 22:05 »

This is getting crazy... I have seen some new clips which are way too good.
Stock video industry is at risk.

SpaceStockFootage · « **Reply #58 on:** March 04, 2024, 02:02 »

Quote from: stoker2014 on February 29, 2024, 10:06

Quote from: Andrej.S. on February 29, 2024, 09:27
I can therefore imagine that in 3 years the AI videos will have high quality.
Yeah, and any buyer will have a PC (probably worth $1000 and no more) to quickly generate such a video. Probably the generation time will be several seconds, maximum minutes.
It's all fantasy. It will be easier, cheaper and faster to shoot or buy a video than to generate it.
Even now you can use cloud computing power, but it costs money. It’s cheaper to buy a ready-made video, which is what buyers do.

This AI software generates content in the cloud, you only need a very basic device to generate images and footage. You've been told this multiple times, but you never listen. Yes, there are AI models that use your computer to generate content, but they're usually more experimental models that are still in testing, the kind you might get from Github or somewhere... most AI content generated today does not use your computing power to produce.

stoker2014 · « **Reply #59 on:** March 04, 2024, 03:35 »

Quote from: SpaceStockFootage on March 04, 2024, 02:02

This AI software generates content in the cloud

This means you will pay a lot of money to the cloud and pay constantly for every video created.

It will be cheaper for a buyer to buy a video on stock than to pay to the cloud.

Evaristo tenscadisto · « **Reply #60 on:** March 04, 2024, 04:21 »

AI is transforming all the industry. The issue with microstock is its video limited options, and shots that require filming still cost considerably more than any generative video AI solution.

I recall that 24 years ago, it took around 16 people to create a small corporate film. Today, it is essentially achieved with 1/4 of the team (scriptwriter, camera, and editor/motion graphics) using a combination of videos and After Effects (AE) templates purchased from microstock, along with some recorded material. Consequently, in the upcoming years, the role of the editor/animator will evolve to include new functions in generating AI videos to complete films.

I have observed this change in various agencies involved in producing magazines, books, events, or advertisements for social media. These agencies allocate less budget to traditional photo shoots with cameras and more towards cost-effective generative AI platforms. It's interesting to witness professional photographers already generating images for agencies using tools like Midjourney. However, it is somewhat peculiar, as the act of clicking the camera button is sometimes entirely replaced by keyboard and mouse clicks.

Adopting AI technologies will undoubtedly make a difference in the realm of video production. Still, I believe that Microstock will persist and likely adapt its selling strategies.

This can currently be observed: Adobe is taking a commercial and generic approach by incorporating AI as an output option in image creation (selling AI images for commercial purposes). On the other hand, SSTK is now following more an editorial path. Their images are real and documentary, serving as input for AI (machine learning). The lifespan of a commercial image differs from that of an editorial one. The editorial image gains value over time, as it becomes impossible to reproduce, while the commercial one depreciates in a short period.

For us contributors, the most prudent approach appears to be investing in both types of markets (Editorial and Commercial).

SpaceStockFootage · « **Reply #61 on:** March 04, 2024, 18:04 »

Quote from: stoker2014 on March 04, 2024, 03:35

Quote from: SpaceStockFootage on March 04, 2024, 02:02
This AI software generates content in the cloud
This means you will pay a lot of money to the cloud and pay constantly for every video created.
It will be cheaper for a buyer to buy a video on stock than to pay to the cloud.

Where are you getting this information from? Please provide comparative costs for purchasing a stock video versus generating an AI video.

Andrej.S. · « **Reply #62 on:** March 05, 2024, 02:39 »

Quote from: SpaceStockFootage on March 04, 2024, 18:04

Quote from: stoker2014 on March 04, 2024, 03:35
Quote from: SpaceStockFootage on March 04, 2024, 02:02
This AI software generates content in the cloud
This means you will pay a lot of money to the cloud and pay constantly for every video created.
It will be cheaper for a buyer to buy a video on stock than to pay to the cloud.

Where are you getting this information from? Please provide comparative costs for purchasing a stock video versus generating an AI video.

C'mon he is either delusional or just a web troll.
You can't expect to get valid arguments.
Haven't found any details for the generation costs but there are rumors for a subscription plan of 60 minutes generation time for 20 bucks and 500 minutes for 50 bucks.
Aswell a pay to go model with costs about 0.01 to 0.10 USD per second.

But I am quite sure the prices will drop fast since NVidia is investing billions in new AI chips for real professional demands like in health, chemistry, etc. industries. So you could rent in few years very advanced GPUs, which should shorten heavily the generation time.

stoker2014 · « **Reply #63 on:** March 06, 2024, 07:21 »

Quote from: SpaceStockFootage on March 04, 2024, 18:04

Quote from: stoker2014 on March 04, 2024, 03:35
Quote from: SpaceStockFootage on March 04, 2024, 02:02
This AI software generates content in the cloud
This means you will pay a lot of money to the cloud and pay constantly for every video created.
It will be cheaper for a buyer to buy a video on stock than to pay to the cloud.

Where are you getting this information from? Please provide comparative costs for purchasing a stock video versus generating an AI video.

My friends are using the cloud to create artificial intelligence models. They pay a lot of money for the use of computing power.
How much will it cost to create a video? How do I know. You probably think that all this will be free for you or buyers.

I am sure that it will be cheaper for the buyer to buy a video on subscription stocks than to create it themselves in the cloud. I think it will be cheaper for the buyer to buy the video and on classic stocks.

Andrej.S. · « **Reply #64 on:** March 06, 2024, 12:06 »

The big question will be if big agencies like Adobe really want to sell generic content from the contributors in the long term. They would have still to pass on contributor's commissions of approx. 30 to 40%.

I doubt it. Look at Adobe's Max sneaks. They have developed so much AI stuff and already are integrating it in their tools.
They have developed even an own LLM.
The invested for sure a huge amount of money. They have to rearn this amount quickly to get joyful investors.

Firefly is getting better and better.
They stated they had already 3 billion generations with firefly in couple months.

And look what they are trying with offering "Mission" contests in the community to finetune their image generator for just 50 to 80 bucks.

When their models can replace most of the current content, they will just recreate the majority and take contributor's content down from their servers.
They will find people, who will contribute some stuff they need to finetune their models for some bucks.

Or they will only allow inpaint modifications in firefly for only with their tools generated images and movies.

This will happen in the first step with photos and then movies.
Just watch the developments in the following years.

opal · « **Reply #65 on:** March 06, 2024, 14:47 »

My concern is that I don't believe what I see on Social Media anymore because of AI, what is true and what is false.

SpaceStockFootage · « **Reply #66 on:** March 06, 2024, 22:03 »

Quote from: stoker2014 on March 06, 2024, 07:21

Quote from: SpaceStockFootage on March 04, 2024, 18:04
Quote from: stoker2014 on March 04, 2024, 03:35
Quote from: SpaceStockFootage on March 04, 2024, 02:02
This AI software generates content in the cloud
This means you will pay a lot of money to the cloud and pay constantly for every video created.
It will be cheaper for a buyer to buy a video on stock than to pay to the cloud.

Where are you getting this information from? Please provide comparative costs for purchasing a stock video versus generating an AI video.
My friends are using the cloud to create artificial intelligence models. They pay a lot of money for the use of computing power.
How much will it cost to create a video? How do I know. You probably think that all this will be free for you or buyers.
I am sure that it will be cheaper for the buyer to buy a video on subscription stocks than to create it themselves in the cloud. I think it will be cheaper for the buyer to buy the video and on classic stocks.

If they're your friends then just ask them how much they're paying, instead of just making a complete guess as to the cost. And what do you mean by creating artificial intelligence models? Are they creating videos... or are they creating software, models and platforms that can generate these videos? There's a big difference!

Anyway. DALL-E, Midjourney etc... they vary from $10 a month for unlimited images to about $0.02 per image. And there's plenty that are free, albeit probably not quite as good. So what's a video going to cost? It's not been announced yet, but Andrej has heard $0.01 to $0.10 a second which is only going to decrease as they get more users and cheaper GPUs. Sure they could get a $30 a month subscription and download a ten second video that is pretty much exactly what they're looking for... or they could pay $1 or less and get a ten second video that is 100% exactly what they're looking for.

Now... you're probably right. I mean, you're very wrong about the whole computing power thing... but you could be right about the prices and whether it's easier to go for a subscription at Envato, Storyblocks or Motion Array. We don't know how much Sora is going to be and the quality will become more apparent when it's released to the public. But 6 months to a year? GPU pricing vs performance has halved, bugs have been ironed out, better models have been developed, adoption increases... the situation will probably be very different.

MicrostockGroup Sponsors

Author Topic: Open AI Sora text to video - unbelievable new AI video tool (Read 13477 times)

Andrej.S.

stoker2014

stoker2014

cobalt

stoker2014

Andrej.S.

stoker2014

Artist

SpaceStockFootage

stoker2014

Evaristo tenscadisto

SpaceStockFootage

Andrej.S.

stoker2014

Andrej.S.

opal

SpaceStockFootage

Related Topics

Sponsors

Microstock Poll Results

Sponsors