Thematic Spotlight: OpenAI’s New Tool Turns Your Text to Video

OpenAI, creators of ChatGPT, announced a new artificial intelligence model named ‘Sora’ which can seemingly create photorealistic videos from user-inputted text prompts. While little is known about the technology behind the software, OpenAI claims Sora is capable of generating ‘complex scenes with multiple characters’ and ‘understands not only what the user has asked for, but also how those things exist in the physical world’.1

Despite the barrage of sample videos released by OpenAI circulating social media right now, text-to-video technology is not a new concept. Big tech rivals to OpenAI (and by extension Microsoft) have also created versions of the technology. Just a few weeks prior, Google announced ‘Lumiere’ a similar program which promises to generate videos with ‘seamless motion’ at 80 frames per second.2 Meta platforms released ‘Make-A-Video’ in 2022 which not only accepted text inputs but could also accept image prompts.3 Nonetheless, it is undeniable that OpenAI’s latest iteration of the technology has been the most impressive – the graphical fidelity of its video is shockingly realistic, capable of fooling much of the internet.

Meta platform’s ‘Make-A-Video’ is currently the only text-to-video program that is fully available for public use. Both Lumiere and Sora remain in early testing and have yet to announce an official release date.

As OpenAI has continued to cement itself as a leader in the artificial intelligence race, the firm has also ballooned in valuation. Last week, OpenAI signed agreements which allowed employees to sell their shares in the company at a valuation of US$86 billion – more than 600% of its estimated value of US$14 billion in 2021.4,5

Access AI and Big Tech with FANG.

Looking for more ETF Express? Check out this week’s Market Moves and Commodity Calls.