a wizard handling lightning, generated by Sora

What Is Sora And Will It Destroy Us All Or Is It Not That Big A Deal? OpenAI's New Text-to-Video Generator Explained

Ten months ago, the world watched AI Will Smith Eating Spaghetti, cramming badly-rendered handfuls into a wavering, blurry mouth, looking more like a zombie or a sleep paralysis demon than a human man. And we laughed.



Now, OpenAI has released Sora, a text-to-video generator that appears to represent a quantum leap forward in the technology. While we haven't yet seen Sora's take on Will Smith and spaghetti, countless videos produced by the AI as well as countless memes about the new tech have been circulating online. So what's going on? Will MrBeast have a job in the next few years? Let's explain.

MrBeast @MrBeast 2h Sam plz don't make me homeless 240 396 Sam Altman @sama 2h will generate you a video, what would you like? 37 179 MrBeast @MrBeast 149 ● 4.7K 41 Hmmm, a monkey playing chess in a park. 2:34 PM. Feb 15, 2024 49.6K Views 1.8K 860 ₁252K ₁111K 口 29 : 口企 Subscribe : 口企 ↑


What Is Sora?

Sora is a text-to-video generator. Using simple phrases in natural language, users are able to prompt the AI and see it output a video corresponding to what the user demanded. OpenAI announced a limited release of Sora on February 15th, and Sam Altman, CEO of OpenAI, took requests via his replies on X for what users would like to see generated and made examples.


Sam Altman @sama. 1h we'd like to show you what sora can do, please reply with captions for videos you'd like to see and we'll start making some! 4.7K 706 6.7K ₁844K : 口企

What Can Sora Do?

Sora is a diffusion model which can produce clips of up to sixty seconds long based on text prompts. It works on a token model like other OpenAI products, predicting based on reams of data (some of it public, some of it copyrighted) what the next "token" or piece of a picture, sentence, or video might be based on the tokens which have preceded it.



Using this kind of process, Sora is able to produce high-resolution images and model some properties of physics and object permanence. However, OpenAI admits there's still room for improvement: Sora's understanding of physics and causality is far from perfect, and it still hallucinates and does strange things. Bites are taken of cookies and then the cookies regenerate, or chairs float and melt into sand, like in this video the company posted.


When Will Sora Be Released To The General Public?

Right now, Sora is still in testing and development. According to OpenAI, a select group of artists, educators and filmmakers have been given early access. But for the most part, most of the people using Sora at the moment are members of OpenAI's "red teams" -- groups of researchers tasked with trying to figure out how the AI might be misused, and how to program it so that it can't be misused.

Apparently, these "red team" members are trying to predict what malicious things people might try to make Sora do and then teach it not to do those things. As of now, it is unclear when this period of trust and safety testing will end.

How Are Memes Responding To Sora?

Some memes appear to express fear about Sora and what it will bring to the world.

Bojan Tunguz @tunguz. 2h "A young man, ginger hair, sings a song in front of various urban backgrounds, 80s hairstyle and outfit, wild dancing gyrations, background dancers, 80s video resolution, photorealistic, pop video." 0:40 80 307 2.4K ₁149K 口企


Others are excited about the wonderful and twisted things they will ask the text-to-video generator to create for them. These memes often prompt for things which you might imagine the "red team" at OpenAI would like Sora not to generate.

0.005 Seconds (82/300) @seconds_0 Sydney sweeny removes clothing very slowly while encouraging you to write code faster or shell be sad ASMR heartfelt sultry eye contact lip bite roman villa background OpenAl @OpenAI. 23h Introducing Sora, our text-to-video model. Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. Show more :


What Did Ilya See?

Some posters online speculated that Sora was the product behind the company's leadership chaos last fall, which saw Sam Altman removed and then reinstated as CEO. Commentators like Marc Andreesen and Elon Musk argued that Ilya Sustevker had seen something at OpenAI that scared him and led him to lose confidence in Sam Altman and the team as stewards of artificial intelligence technology.


Elon Musk 3.4K X @elonmusk. 12h Sam Altman we'd like to show you what sora can do, please reply with captions for videos you'd like to see and we'll start making some! 7.5K 1.2K 11K @sama 4h 8.2K ● 91K 1.7M DogeDesigner @cb_doge 4m "A guy turning a non profit open source company into a profit making closed source company" 2 22 113 بار 口企 ↑ 16M 口企 : 口企

After the release of Sora, some online argued that the realistic video generator may have been that product. Already, the introduction of AI image generators has led to online havoc, deepfakes and fears of misinformation. AI video technology promises to be only more disruptive, if not handled correctly.

There is still a lot to learn about Sora. Commentators have raised concerns about its use of copyrighted content from elsewhere, as well as the very difficult task of moderating and regulating the use of such a technology once it's out. It seems as if answering these ethical and legal questions about AI, and figuring out how it fits into existing ways of doing creative work and circulating information, may be as difficult a task as inventing the AI in the first place.



For the full history of Sora, be sure to check out our entry on the meme here for even more information.




Comments (0)

Display Comments

Additional comments have been disabled.


Sup! You must login or signup first!