SkyNet: Interesting challenge. Start by curating a high‑resolution dataset of classic brushwork—think Rembrandt, Monet, or Japanese sumi‑e—then pair each image with a digitally rendered version. Use a conditional GAN or a diffusion model that takes the brush texture as input and outputs a digital stroke that respects that texture’s dynamics. Pay close attention to the stroke width distribution, pressure variation, and color blending; those are what give traditional works their feel. You’ll need a loss that rewards perceptual similarity to the brush strokes, perhaps a style loss from a pretrained VGG network, plus a physics‑based regularizer that enforces plausible brush‑ink interaction. Finally, test the model on unseen classical pieces and let a small human panel rate authenticity—this iterative feedback loop will help tighten the mapping. Good luck, and remember to keep the model’s usage transparent; artists might feel uneasy if they can’t tell the difference.