I built an AI slop machine over the weekend. I will never launch it, after seeing what I created. You put in a story prompt, and it generates a more structured story with stakes, cliffhangers, a beginning, middle, end. Splits it into episodes. Storyboards the whole thing, makes a cast, music, poster, voiceovers. Then generates the visuals (different styles available from anime to cinematic) and you can play the episode as a slideshow, or render it as a video. Uses all the latest models that you would expect. A few observations from this exercise: 1) it was trivial to build There's no moat here. This is just "AI glue", building an app that wraps several AI models. The moat is the GTM strategy, which leads me to... 2) this space is insanely competitive there's tons of AI movie makers all using the same models with some thin layer of prompt direction or UX as the differentiation. totally saturated. 3) AI video courses I noticed that some of the biggest advertisers in this space are people selling "how to create your own AI movie" courses. There's something off about this, almost dropshippy. It suggests that there's more value in hyping the thing and selling courses about the thing, than there is in the thing. 4) this shit is really expensive All of those cool AI videos that you see using the latest video models, are expensive to make. A 60 second "mini episode" of a story built using my platform would cost $5 in raw inference. There's no business here, whatever margin you put on top of that can just get eaten by a competitor. And imagine if a user buys $100 of credits (of which $90 is passed through to your AI provider) then decides they hate the movie that was created, and asks for a refund or charges back. Enough of those in a month will bankrupt you. 5) just didn't feel right It's a fun little demo that got some laughs from my friends but ultimately it's contributing to the never ending pile of AI slop. Is that something I want to be part of? Not really.