Might be true, recently there have been highly boasted improvements in the RTL-to-GDSII pipeline, using AI, as one does. Both from Cadence and Synopsys.