Mind flayers are hard to do in AI-generated art, but I've made a few nice ones happen in Stable Diffusion with a little effort. It's far easier here than in Dall-e3 or MidJourney, based on my experience anyway. One or more of the models and LoRAs involved apparently leaned into NSFW, probably the base model. I included safe, cropped versions of some of my accidental NSFW results because the face tentacles were so nicely done!