Last year OpenAI wowed everyone with their new AI video creation tool - Sora. Much to the disappointment of everyone, they did not release it to every country immediately. But, finally, we in the UK can play with this tool.
Sora currently sits separate from ChatGPT (similar to the image generation tool DALLe originally, which is not integrated to ChatGPT). To use it, you have to be a paid user of ChatGPT, and even still usage is heavily restricted (for those of us on plus, only 50 videos a month).
Something which made my spidey senses tingle
I had to give OpenAI my date of birth, which struck me as odd. Is this because they are data mining or because there is a risk of inappropriate videos being created? Either way, this left me feeling a bit uneasy with this technology.
Video Sharing and Learning
One thing I really like is being able to see everyone’s videos - it’s helpful to see other people’s creatives and you actually get to see the prompt and storyboard used by other people. Learning how other people prompt and seeing their results is a great way to learn how to do so better.
This is similar to how Midjourney showcasing images created on their software and the prompts.
I’m particularly taken by a video of “Terrier's Joyful Sprint” - as this is actually the dog I want! And it looks very happy.
The prompt to get this video was:
A high-energy pure white Wire Fox Terrier sprinting at full speed through a narrow dirt path surrounded by endless sunflower fields. The dog's entire body is pure white, from head to paws, with no black or brown markings. Its wiry fur bounces with every powerful stride, ears flapping slightly in the warm summer breeze, and its sharp eyes remain focused forward. The golden sunflower fields stretch endlessly on both sides, their bright petals swaying gently in the wind. The sky is a brilliant shade of blue with a few scattered white clouds. As the terrier dashes forward, small dust particles lift from the dirt road beneath its paws, enhancing the sense of speed. 📷 Cinematic Camera Motion: A low-angle forward-facing tracking shot, where the camera is positioned slightly ahead of the dog, smoothly moving backward while keeping the dog perfectly centered. The golden sunflowers create a stunning contrast against the dog's pristine white fur, making the scene visually striking.
Wow - two thoughts:
- The video is pretty accurate for this prompt
- The prompt is very long and really detailed
How to create a video on Sora
- Sign up, date of birth, username (and unfortunately be paying for ChatGPT)
- Write a prompt
- Options
- Upload or choose an image or other video
- Choose
- a preset (eg Film Noir)
- size (16:9, 1:1, 9:16)
- resolution (480, 780, 1080 - not available on plus)
- length (5s, 10s, 15s, 20s - only 5 / 10 available on plus)
- Variations (1, 2 or 4 - this is a bit sneaky as set to 2 on default which costs some of your limit)
- (Optional) Storyboard (where you can actually direct the video)
My Attempt
In comparison to the lovely terrier, my first attempt at a video for Taught by Humans was not great. I didn’t even prompt, I just played with the storyboard feature. Definitely a learning curve for me (which I will share!)
From frustrated to the future of learning
Note - I was deeply confused why Sora decided to add wine glasses to the video. As per usual I made a typo “everyone wines” was supposed to be “everyone wins”.
Let me be clear - while the tech is incredible, it only highlights the need for creatives. I’m not a video designer, I have no clue how to storyboard, or how to craft a compelling video (and using Sora made me feel like the candidates on The Apprentice who never quite nail the role of creative director).