The controversy angle is what stands out to me. Medium built its reputation on human voices, so folding AI assistance into the writing process itself feels like a real shift in identity, not just a feature addition. I’m curious how they’ll handle disclosure when a draft starts with machine suggestions. As a reader, I mostly want to know whose words I’m actually reading. If the tool stays clearly optional and transparent, that seems more defensible than quietly blending the two.