Credit: Unsplash
OpenAI is about to launch its latest creation, Sora, a video generation tool that promises to take AI-powered content creation to the next level. While it’s not available to everyone just yet, YouTuber Marques Brownlee got an exclusive first look and shared his initial thoughts in a recent 15-minute review. Here’s everything you need to know about Sora, the future of AI-driven video content.
What Is Sora?
Sora is OpenAI’s new video generation platform that’s launching on Monday — though, for now, only a select group of users will get early access. According to Brownlee, Sora isn’t integrated with ChatGPT (OpenAI’s chatbot), but instead stands on its own as a separate video generation tool. To get started, users can head to Sora.com, where the homepage displays a feed of recently generated videos, curated by OpenAI. These videos are all created using user prompts or uploaded images.
i’ve been watching tons of openai’s sora videos on threads. it appears the creative folks are on that side of the world. pic.twitter.com/sT6jg0OhJ3
— David Dosu (@thedaviddosu) December 10, 2024
Features That Make Sora Stand Out
Sora isn’t just a tool for creating random video clips. It offers some really exciting features that make it versatile and interactive:
- Video Generation from Prompts and Images: Users can input text prompts to generate a video, or upload images that Sora will turn into a full-fledged video. This allows for endless creative possibilities.
- The “Re-mix” Feature: One of the most interesting features is the “Re-mix” option. If you’ve generated a video and want to make tweaks, you can describe the changes you’d like, and Sora will regenerate the video with your adjustments. There’s also a “strength” setting, allowing you to control how drastic the changes are — higher settings yield more creative and artistic modifications.
- Storyboard Feature: If you want to create a more complex video, Sora lets you use a “Storyboard” tool to string together multiple prompts. This helps create consistent video sequences, a known challenge with AI video generation.
How’s the Quality?
Sora can generate videos in up to 1080p resolution. However, as with many AI tools, there’s a tradeoff. Higher-quality videos take longer to generate — 1080p footage can take eight times longer than 480p footage, which is the fastest setting. In Brownlee’s tests, a 1080p video took a “couple of minutes” to generate, but he also noted that the speed could slow down as more people start using the platform.
Got access to Sora today, OpenAi's text to video generator.
Pretty cool.
Prompt for my first video.
a hacker in a cyberpunk dystopian future using a neural interface to jack into a virtual reality mainframe to take over the internet and replace it with an ai of his brain pic.twitter.com/dqvaJxyHCA
— Chris M. Walker (@cmwalker) December 10, 2024
What’s the Catch?
Sora Sora Sora! I'm sure you have seen some of the Sora videos posted and have not been impressed compared to @Kling_ai or @Hailuo_AI.
You are prompting wrong! With Sora, just tell it what you want. Here are a few examples (with prompts) to show you how it's done. @OpenAI pic.twitter.com/uzib4Oo5Xs
— Flyover Base (@FlyoverBase) December 9, 2024
As with all AI-driven video tools, Sora isn’t perfect. Brownlee pointed out a few major issues that are common in generative video tools. For example:
- Object Permanence Problems: In Sora videos, objects sometimes pass through one another or disappear unexpectedly, which can make the videos look jarring and inconsistent.
- Legs Are a Big Challenge: One of the most noticeable issues Brownlee encountered was with legs. Whether it’s people or animals, Sora had trouble keeping the front and back legs correct when characters moved around. Legs would often swap positions or look anatomically incorrect, making walking sequences look off.
While Sora shows incredible promise for the future of video creation, it’s clear that the tool still has a few kinks to work out. However, the ability to generate videos from simple prompts, remix existing content, and create full video sequences is a big leap in generative AI. As more users get their hands on Sora, OpenAI will likely fine-tune these issues and introduce new features.
The rumors are true – SORA, OpenAI's AI video generator, is launching for the public today…
I've been using it for about a week now, and have reviewed it: https://t.co/jII49vkuHN
THE BELOW VIDEO IS 100% AI GENERATED
I've learned a lot testing this, here are some new… pic.twitter.com/uA1EhRuK7B
— Marques Brownlee (@MKBHD) December 9, 2024
For now, it’s exciting to see how AI is changing the landscape of video creation, and Sora might just be the tool to make professional-quality videos accessible to everyone. If you’re one of the lucky early users, get ready to experiment with this cutting-edge technology — but keep your expectations in check as it continues to evolve.
