As much as I like Scenario, and am on the fence about AI art, I really do not like terminology shift from training → pre-training, and fine-tuning → training. Companies can say things like "train on your art", which still means the model was pre-trained on scraped art.
And the math & opt-in/opt-out debate aside -- I do not like the "lying by omission" aspect of it.
Like, a platform isn't more ethical than midjourney just because it provides services to fine-tune a model on a set of art samples. 🤷🏽♀️