They're the ones that released Stable Diffusion (the ML model that can generate images from text prompts).