For a while, I assumed the biggest problem in AI image editing was obvious: the model just wasn’t good enough yet. If the output changed a face too much, use a better model. If the product logo became distorted, use a better model.
Source: [Dev.to](https://dev.to/gaven/i-thought-better-models-would-fix-ai-image-editing-i-was-wrong-5ak)