Christian Szegedy posted that whoever thinks AI did not transform mathematics is ridiculously delusional. The researcher, who co-invented Inception and batch normalization at Google before co-founding xAI and moving to Math Inc., made the statement on X. Leo Boytsov replied that he remains somewhat skeptical yet sees the development as the start of a paradigm shift. Boytsov added that specific optimistic forecasts may not arrive as fast as expected, and he drew the same point for coding. The exchange centers on the extent and timing of AI's role in mathematical work according to the two researchers.