WooouTK
patreon
AI girls in stocks
🕑 Added 2023-11-04 12:18:15 +0000 UTCLately, I've been trying to see what I can achieve with the Bing image generator based on DALL-E 3. The truth is, for every good image, that is anatomically correct and has the correct number of toes, I have to make between 10 and 15 failed attempts. This is not to mention the sensitive content filter that the application itself has, based on certain words like "feet" or "tickling". Sometimes it surprises me by triggering even when changing the prompt from "cartoon-style girl" to "Pixar-style girl." For now, the AI feels a lot like a slot machine to me, where most of the time, you don't get a good result, but occasionally, there's a jackpot. Here's the best I've managed to get, and I'm also attaching the prompt for your reference.
PROMPT: "Mujer sonriente estilo dibujo animado realista, los tobillos y rodillas de la mujer estan emplazados a traves de un pesado cepo medieval de madera y acero que los cubre y va sobre una peana, la mujer se halla sentada, las piernas de la mujer estan rectas, la mujer no lleva calzado, la mujer tiene cinco dedos en cada una de sus extremidades visibles."
As a Spanish speaker, you can imagine that I did it in my native language since it's easier for me to rephrase the text using synonyms or asking for the same thing in a different way so that the AI understands what I'm requesting without triggering the content filter.
