very nice song indeed :)
Maybe two things. The stereo image is a bit too polarised in my ears. The effects are very strongly L/R panned while the rest is almost completely in the middle. I´d spread the middle at least a little more so the L/R-sprectrum is used more homogeneously. Doesn´t have to be all over, emphasising those outstanding sideFX. Then I´d raise the level of those sounds in the middle, the FX appearing too dominant atm levelwise. Considering composition I was always waiting for the speechsample to dropout giving room for the rest of the song. I know it´s machinelike repetetion, just saying I´d like to focuse on the rest aswell. So I´d use it more like a chorus or at least having some parts without it.