© IE Online Media Services Pvt Ltd
Tags:
Journalism of Courage
It’s been less than a week since OpenAI introduced its much-awaited video-generation model, Sora. Ever since, the Internet has been buzzing with sample videos created by eager users. Nearly a year after OpenAI teased its video model capable of creating hyper-realistic outputs, the AI powerhouse has finally rolled out the model.
With Sora, users can generate videos up to 20 seconds long in up to 1080p resolution. The videos can be in widescreen, vertical, or square aspect ratios. Moreover, OpenAI is also allowing users to try their own assets to “extend, remix, and blend or create” new content from text. The company has also developed new interfaces, making it easier to prompt Sora with text, images, and videos.
Before we go on about Sora, below are some examples of videos generated by the model, as shared by users on X.
The First Humans
X user @KNGMKRlabs created a 4.18-minute-long video about the first humans on Earth. The output using Sora is combined with an AI-generated narrator. Reportedly, the user took a week to prompt and stitch together the video outputs. The video footage begins with a Neanderthal braving through a hailstorm in a snowy landscape. The montage later shows childbirth, cavemen sitting around a fire, and landscape shots of the early terrains.
Pelican on a bicycle
Another X user Simon Willison, who shared the news of Sora being available to all ChatGPT Plus and Pro users in non-European countries, also shared an amusing video of a pelican riding a bicycle. Willison used the prompt: “A pelican riding a bicycle along a coastal path overlooking a harbor.” The video output is realistic and shows the bird casually riding a bicycle on a path flanked by foliage on one side and a seacoast with boats on the other.
AI-generated news clip
Popular tech influencer Marques Brownlee shared his creation using Sora on X. Brownlee, who has been using the AI model for over a week, shared a news clip that he claims is 100 per cent AI-generated. The influencer went on to share a thread of his learnings from the AI model. The video, which is essentially footage of news, has some telltale signs of AI-generated videos, such as garbled text. According to Brownlee, Sora created the cutaways, moving text ticker, and news style shots, and the newsreaders looked real.
The Golden Eye
Justin Hart, another X user, shared a unique video created using OpenAI Sora. The video is essentially a close-up of an artificial eye that has an iris that resembles the circuits on the motherboard. The close-up further reveals metallic petals adorning the eyelashes and a swirl of colours—blue, black, and silver.
Bringing history to life
AI entrepreneur Pietro Schirano shared a video of a Roman soldier resting after a battle. The video shows the soldier sitting on a log of wood in a deep and dark forest, with the camera panning around him. The visual captures the soldier’s outfit and helmet with accuracy.
Anime on the go
Hidekazu Furukawa used Sora to create an anime. The short clip he shared shows a cat-eared maid greeting and smiling. He used the prompt: “Cute cat ears maid of Japanese anime style.”
Travelling to the Alps
AI influencer Rowan Cheung shared a short clip showing a man sitting in a train and gazing at the Swiss Alps through the window. For his first-ever video using Sora, Cheung used the prompt, “A young man looking out of the window on a train overlooking the Swiss Alps.” Cheung also remarked that the site was laggy as millions were accessing it for the first time, making the queue longer.
Another influencer, Blaine Brown, used Sora to create a video rendition featuring multiple shots. While the primary video shows an aerial view of a castle, the second video shows it from a different angle, and the third video shows the castle from the perspective of two joggers marching towards the structure.
Aditya Ramesh, whose bio claims that he is vice president of research at Sora, shared POV footage following an ant into its nest. The video shows an ant walking into its nest through a tunnel-like path with the camera following it. Ramesh used the prompt “POV footage of an ant navigating the inside of an ant nest.”
https://twitter.com/model_mechanic/status/1759068809867166129
Sora will be available to users subscribed to OpenAI tools. So far, users in the US and most countries internationally have access to the tool. Sora is yet to be available in the UK and Europe. When we attempted to log in to Sora, the site showed a message that read, “Sora account creation is temporarily unavailable.”
According to the site, it is facing heavy traffic, and account creation has been temporarily disabled. “If you’ve never logged into Sora before, please check back again soon,” said the message on the site.
