vortex What if we just didn’t decompress it? How push-down compute helps Vortex run faster and decompress less
