I let my kids have access to the app in the hope they would be inoculated against being obsessed with AI video and it actually worked. They got bored in like 2 days.
It simply doesn't compare well with handcrafted short form videos that are already plentiful on TikTok (which I absolutely don't let my kids watch).
I was talking to other people re: difference between code & other domains. Code is, for customer, what it does.. not how it does it. That is - we can get mad about style, idioms, frameworks, language, indentation, linting, verbosity, readability, maintainability but.. it doesn't really matter for the customer if the code does the thing its supposed to do.
Many things like entertainment products don't work that way. For a good book/movie/show, a good plot (the what) is table stakes. All of the how matters - dialogue, writing style, casting, camera/sound/lighting work, directing, pacing, sound track, editing, etc.
For short format low stakes stuff like online ads, then the AI slop actually probably works however.
Same for say making a power point. LLMs can quickly spit out a passable deck I am sure. For a lot of BS job use cases, that's actually probably fine. But if it is the key element of a sales pitch, really it's just advanced auto-formatting/complete, and the human element is still the most important part. For example I doubt all the AI startups are using AI generated sales pitches when they go to VC for funding.
A promotional flyer for an event could work perfectly well in plain text. The art is pure social signal - this event is thrown by the type of people who put art in a certain style on their flyers. Your eye is caught and your brain almost immediately discards the art.
Same with power point - you make a power point so that everyone knows this decision was made by the type of people who make power points. A txt file and a png would have gotten the job done.
Same also with memes - you could just _say_ a lot of these jokes, but they're funnier with a hastily-edited image alongside.
What happens when other platforms start trying to get people to pay? I think there's a race to find a revenue stream for this stuff. As soon as one company can find a way to monetize it, they'll all end up doing it. Right now, we're in a place where companies are losing so much money, they have to decide how much they can lose before they pull the plug.
OpenAI just proved you cannot burn money indefinitely.