This was the first time I (JonJon) worked with overlap, and it sure was interesting. The main challenge I faced, was making sure that when the berlo turned to the side with air and water, there would still be some water and fire in the molecule at known spots. Once I figured that out it wasn't too hard to make the output. The only cose saving I did, was trying to remove glphys from the first layer and checking if the solve sill worked, and I actually managed to save 60G that way.