4 ms·

Yeah, I've been saying that removing accidental similarity isn't improving code but compressing it. "Huffman coding", if you will...