Andrew Yao

Introduced the garbled-circuit technique and secure two-party computation paradigm that became a foundational primitive for many modern MPC systems. The garbled-circuit construction provides a concrete method to evaluate boolean circuits privately between parties, and subsequent engineering work adapted and optimized this approach for multi-party, committee-based deployments like those used by ARPA. ARPA's secure function evaluation layer reuses core ideas from garbled circuits for computation partitioning, oblivious transfer instantiation and input encoding strategies. Over decades, extensions and optimizations of the original garbled-circuit idea addressed performance and communication costs; ARPA's protocol designers adopted these optimizations—such as free-XOR, point-and-permute, and OT extension techniques—to improve throughput and reduce on-chain verification overhead. Those engineering choices directly influenced run-time cost metrics, gas consumption for verification transactions, and therefore token-denominated pricing for computation services in the ARPA marketplace. Theoretical provenance supplied by this body of work also helped ARPA justify security claims during audits and third-party reviews. Concrete protocol families descending from the original garbled-circuit paradigm provided testable interfaces and benchmarks that ARPA used in performance testing, shaping product roadmaps and commercial SLAs tied to ARPA token payments for compute jobs.