TED Talks and OpenAI shared new video made with Sora
TED Talks provides some of the most engaging educational speeches free online. For example, Microsoft co-founder Bill Gates had a TED Talk titled, “We can make COVID-19 the last pandemic.”
Its most viewed clip was from “Do Schools Kill Creativity?” by author Sir Ken Robinson, with over 75 million views.
READ: Best Apple TV learning apps
Article continues after this advertisementIn 2024, the popular educational website is ready to show us the future with AI. Recently, TED Talks shared a teaser video created with OpenAI’s Sora.
How did TED Talks make the AI video?
What will TED look like in 40 years? For #TED2024, we worked with artist @PaulTrillo and @OpenAI to create this exclusive video using Sora, their unreleased text-to-video model. Stay tuned for more groundbreaking AI — coming soon to https://t.co/YLcO5Ju923! pic.twitter.com/lTHhcUm4Fi
— TED Talks (@TEDTalks) April 19, 2024
Tom’s Guide says LA-based director Paul Trillo created the innovative TED Talks video with Sora, OpenAI’s text-to-video tool.
The tech firm announced the technology earlier this year, but it only provides access to select artists. Trillo said he started by generating over 330 Sora clips and then edited them.
Article continues after this advertisementHe told Business Insider he used prompts, which were “a cocktail of words that I use to make sure that it is less like a video game and something more filmic.”
He also used phrases like “anamorphic lens,” “depth of field lens vignette,” and “35 millimeters.” The director claimed he needed them to ensure wouldn’t “kind of default to this very digital-looking output.”
Later, he compiled the clips into a video that lasted one minute and 33 seconds. The final video contains 25 videos from the AI tool.
Sora generated everything except the TED logo, meaning the video uses AI-generated motions and shots.
“Really fun to explore techniques I have done in the past with this new tool. Unlocks a lot of new ideas,” said Trillo.
The video above follows a camera’s point-of-view zooming into different scenes and various seminars, teasing what TED Talks have in store this year.
OpenAI admits Sora has numerous limitations. Its official blog post states the AI program doesn’t “accurately model the physics of many basic interactions, like glass shattering.”
Nevertheless, Sora would likely improve significantly over the next few months, based on ChatGPT’s rapid progress. Also, TED Talks’ X post promises to show more “groundbreaking AI.”
