"Record Seed" is a collection of 404 abstract paintings made using a style-based generative adversarial network known as StyleGan for short. The collection was generated based on 500 of my paintings made on decrepit record sleeves collected from record shops around Southern California that were being tossed out due to severe weather damage. I ended up selecting 404 because the main tool I use when I travel for music is called the SP404. The sampler sparked my interest in record collecting which eventually led me to making the paintings and later this body of work. I also felt as though the machine learning process could be viewed as a sampling process of my paintings' data giving it a similar feeling to my SP404. The pieces here are made from 4 different iterations of the same dataset to provide extra variety in the machine's "seed" output. The paintings date back to 2009 and the machine learning took place from April 10th 2021 to September 1st 2021. Each one is unique.