The speaker only recently (a few weeks before this conversation) understood the deep physical reason why the minimum bound in assembly index is so important, despite intuiting it was correct earlier.