I’ve been experimenting with AI image tools and am curious about Grok’s approach to revisions. A first result can be impressive, but the harder test is asking for a small change—adjust the lighting, keep the subject’s pose, or remove one detail—without losing what already works.

Has anyone tried Grok image generation edits over several rounds? I’d like to hear whether it can preserve the parts you want to keep, or whether each follow-up tends to feel like starting over.