Back to the 90s: Stanford professor Dan Boneh is fond of using an academic paper from the 1990s to make a point – today's “cutting edge” zero-knowledge technology is an old concept born of necessity, originally called "proof-of-compute." The paper states that "… a single reliable PC can monitor the operations of a herd of supercomputers…." Or, more practically, that a growing collection of low-powered mobile devices could perform lightweight proofs, showing that powerful central computers had done their jobs. This breakthrough was the root of zero-knowledge, and for thirty years, it languished in obscurity. Why? Because unexpectedly, Moore's Law put a supercomputer in every pocket. The problem of low-powered mobile devices disappeared. Today, the challenge has reemerged in a new format. Decentralized blockchains offer super-powers – trustless computing and internet-scale value transfer. But they can't keep pace with anticipated demand at the scale of mass adoption. To integrate decentralized blockchains with the velocity of global financial transfers, micro-transactions, identification checks, access passes, and more, we need centralized computers capable of quickly handling the flood. The math behind zero-knowledge allows us to bundle these high-volume transactions and submit them to a decentralized blockchain in efficient packages that can be proven correct by an independent user. Decentralization, at scale, is what zero-knowledge math helps us achieve. It is a thirty-year old idea whose application has finally come. Like so many technologies fetishized as "cutting edge," it isn’t new. The future is back to the 90s.