In the latest installment of UltraUpdates Technology series, you’ll discover how to effectively use OpenAI’s newly launched video generation model, Sora. This release coincides with the fifth day of the company’s Shipmas event, a 12-day showcase featuring various tech innovations and demonstrations. Here’s what you need to know and how to start using this video generation tool immediately.
What is Sora?
Sora is an advanced AI-driven model designed for video creation, capable of producing videos based on text input, animating still images, or remixing existing videos in innovative ways. Although OpenAI initially showcased Sora in February, today marks its first availability for public usage.
What’s different in this latest version?
While the fundamental capability of Sora to generate remarkable videos from straightforward prompts is consistent with what was demonstrated in February, OpenAI has focused on enhancing the model’s speed and cost-effectiveness for this broader launch. There are several new features introduced, with two particularly notable ones.
The first tool is known as Storyboard, which allows users to produce various AI-generated videos and compile them on a timeline, similar to traditional video editing software such as Adobe Premiere Pro.
The second feature is a feed that acts like a creative showcase. Users can share their Sora-generated videos in this feed, view the prompts that inspired specific videos, make adjustments, and find inspiration overall, according to OpenAI.
What can you accomplish using this tool?
It allows you to create videos based on text prompts, modify the style of your videos, and alter various elements through a feature known as Remix. Additionally, you can combine several clips using Storyboard. Sora offers a variety of preset styles for your videos, such as atmospheric film noir or a cardboard and papercraft aesthetic that mimics stop-motion animation. You also have the option to trim and loop the videos you produce.
Who is eligible to utilize this feature?
To create videos using Sora, a subscription to one of OpenAI’s premium offerings is necessary—either ChatGPT Plus for $20 monthly or ChatGPT Pro for $200 monthly. Both plans also grant access to additional OpenAI services. With the ChatGPT Plus subscription, users can produce videos lasting up to five seconds at a maximum resolution of 720p, allowing for the creation of 50 videos each month.
Subscribers to ChatGPT Pro have the ability to create videos that are longer and of higher quality, with a maximum resolution of 1080p and a length of 20 seconds. Additionally, they can request Sora to produce as many as five different video variations simultaneously from one prompt, allowing for quicker evaluation of choices. While Pro users can generate up to 500 videos each month, they are also permitted to make an unlimited number of relaxed videos, which are produced during off-peak site traffic times rather than in real time.
With both subscription tiers, users can produce videos in three different aspect ratios: vertical, horizontal, and square.
Non-subscribers will only have access to view the feed of videos created by Sora.
OpenAI has announced that it is beginning the worldwide rollout of Sora today; however, the launch will be delayed in many parts of Europe, according to the company.
How can I obtain access?
OpenAI has separated Sora from ChatGPT. To use it, visit Sora.com and sign in with your ChatGPT Plus or Pro account. (At the time of publication, UltraUpdates was unable to reach the site; a message on the site mentioned that new signups were temporarily halted due to high traffic.)
What led us to this point?
Several developments have taken place since OpenAI introduced Sora in February. Other technology firms have rolled out their own video creation tools, such as Meta’s Movie Gen and Google’s Veo. Additionally, there has been significant pushback. For instance, some artists who were given early access to test Sora leaked the tool publicly in protest of OpenAI’s use of their artwork for training purposes without any payment.
What comes next?
With the launch of any new model, it is uncertain what measures OpenAI has implemented to prevent Sora from being misused for harmful, unlawful, or unethical activities, such as producing deepfakes. Regarding moderation and safety concerns, a representative from OpenAI acknowledged that achieving perfection right from the start may not be possible.
A pressing concern is the amount of computational power and energy that Sora will consume with each video it produces. Creating a video demands significantly greater computing resources and energy compared to generating a standard text reply in platforms like ChatGPT. The surge in AI technology has already led to high energy consumption, posing difficulties for tech firms trying to reduce their carbon footprints. The widespread access to Sora and similar video generation models could exacerbate this issue further.
