Scalable Transparent ARguments of Knowledge (STARKs) are a kind of succinct zero-knowledge proof which do NOT require trusting any party to generate a Common Reference String (CRS). In this work, we examine the applicability of STARKs to improving Counted-as-Collected verification in the homomorphically tallied elections. In particular we are interested in using STARKs to allow very efficient tally verification while providing everlasting privacy to the information made available for public verification. This work provides a useful reference for the computational and verifiability trade-offs of using STARKs.