April 2024
Perceivings
Alan Dean Foster

Soaring

By this time you have probably heard of Sora. Developed by OpenAI, the mischievous wizards who brought you ChatGPT and DALL-E, Sora is a program that lets anyone generate videos up to 60 seconds in length simply by inputting a little text. It’s the in-motion descendant of DALL-E, which generates still images using the same kind of input. We don’t have magic lamps today. We have code.

The name Sora derives from the Japanese word for ‘sky,’ but with the implication that “the sky’s the limit.” Look for lots of ads that play on the word, ala “Sora soars,” “Go Soraing,” and so on. The app is not yet available to the general public, but expect it to happen soon. Not because its developers at OpenAI necessarily want it to be so, but because, just as happened with image generators like DALL-E and Midjourney, competition will arrive swiftly.

When you see some of the videos that Sora can produce, after the initial shock wears off you’ll begin to notice imperfections in the results. For example, mammoths and humans and cats appear to move at the same pace. The same thing happened with still picture-generating software. Sora and its like will rapidly experience similar exponential improvement. This is just Sora v1.0.

Look for a lot of 60-second movies, limited only by the imagination of those inputting text. These will be loads of fun. Think wholly artificial TikTok videos. Where the financial impact will be felt first, however, is not in motion pictures or YouTube videos, but in advertising. Television and online adverts rarely run longer than 60 seconds. Sora and its forthcoming kin are going to turn the world of advertising upside down and inside out. Activate your imagination for a moment.

Budweiser commercial. Input: “Three Clydesdales sitting on oversized couches watching the Super Bowl while quaffing Budweiser from huge steins and commenting on the game.”

Ford pickup truck off-road commercial. Input: “Ford F-150 pickup truck driving on Mars, allowing for Martian gravity, with one Martian driving and the passenger Martian tuning the radio. Both Martians dressed in American ranch-style clothing.”

Frozen TV-dinner ad. Input: “A fancy restaurant meal vibrates, then flies out of a restaurant to insert itself into a waiting microwave. After a moment, the food emerges transformed into a steaming TV dinner, ready to eat.”

Sure, every one of these commercials could be done with existing tools. Actors, animation, CGI and other special-effects techniques. That’s not what matters here.

What matters is that every one of them can now be generated with simple text input, and the only cost for the entire commercial is a computer, keyboard, and Sora or Sora-like software. Price? Essentially nothing if one already has the basic tools. Production time? Ready not in months, but in seconds. And if the first one doesn’t please, generate another, and another. At no additional cost. Maybe use editing software to tweak the final choice.

Anyone who thinks this software isn’t going to completely upend the advertising industry within the next year has not been paying attention.

The 60-second time limit does not mean film and TV won’t be affected. There’s nothing to prevent incipient filmmakers from stringing together multiple 60-second efforts for a longer production. As to short special-effects sequences, the use of Sora-style video inserts will completely transform the need for expensive and time-consuming work.

Show the Enterprise approaching a ringed planet. Show an elephant in Africa charging a safari vehicle. Bellowing and screaming to be added in post-production, although I expect that sound and music-scoring options will be added rapidly. Show an asteroid striking an American city. Show a cute cat firing a cat-sized shotgun at a mouse hole in a wall as a mouse emerges (okay, I confess I took that last one from an existing cartoon).

That’s live action. How about some minute-long animations? Perhaps some combining animation with live-action, once famous characters are licensed? Scrooge McDuck making a deal with Donald Trump. The seven dwarfs from Disney’s Snow White picketing showings of the new Willy Wonka movie (you’ll have to research that example to understand the why of it).

The trademarking and licensing of real people will become even more important than it is already. How about promoting Corvette convertibles by showing Marilyn Monroe and a young Cary Grant driving one? But — why stop with recently demised celebrities?

On YouTube you can see videos of famous personages from the past not just transformed by colorization, but lightly animated, by folks working at home. Have a look at Mystery Scoop’s laughing, blinking Abraham Lincoln. With Sora you can do anything you want with Lincoln. Or any other individual, or setting, from the past. The possibilities are endless, and mind-boggling. Add in voices and sound and, well, soon it really will be difficult to tell the real from the synthetic.

When the option arrives to put yourself in a Sora video, everyone can be a superhero. Or a fish. Though there still has to be a human brain at work to enter sensible text prompts.

For now.

Prescott resident Alan Dean Foster is the author of 130 books. Follow him at AlanDeanFoster. com.