Huh. That's one of the parts I enjoy most! There's so many ways to approach it. One I haven't seen documented widely: Use a slow belt holding two science type

Huh. That's one of the parts I enjoy most! There's so many ways to approach it. One I haven't seen documented widely: Use a slow belt holding two science types (one each side) to feed the left side of a faster belt (which then has a left lane that alternates between the two colors). Repeat for the right side (your fast belt now has four colors). You now have a four-color belt to feed your labs, but it backs up when you use more of one color than another, so at the end you add splitters to extract the individual colors into four belts, and feed them back into the start (with priority, so they don't block).