Raw LLM Responses

So, all in all, I agree with you here. But hear me out. Try to use GenAI to create some *specific*. You probably won't be able to iterate without either paying (no different than having to pay adobe for digital art supplies - protip: fuck adobe) but even using some free generative models getting *exactly* what you want out can be very challenging. Driving GenAI is almost an entirely different skill, where the transferrable skills from your traditional or digital art is more in understanding compositions and accurately describing them - probably more like doing Art History than Art classes. Naïvely dropping in a prompt and taking the first result gives you the appearance of using AI - but this is the baseline, and is not the true power than "we" the artists can extract from AI. Just yielding control to an AI provider and accepting slop back is absolutely going to give the soulless output, because nothing in - nothing out. But build your own model based on open source tech, use your own corpus of work to train it, become skilled at what prompts result in good output for your model - its like having an apprentice, you've imbued it with the accumulative efforts of your love of the medium, and in time you can learn to work collaboratively. My dream for the future is that we can all have truly personal assistants that are like out digital children - the humdrum creation can be left to them, and we can work solely on pushing the boundaries of what is possible.