Verifiable Computation Without Trust Anchors

Most verifiable computation systems sneak in some trust. Maybe it is set up ceremonies or trusted hardware. What happens if you yank those out completely? If the verifier cannot make any assumptions about the prover, then the only thing left is pure math. That means you need protocols that are agnostic to the environment, and that gets tricky fast. Zero knowledge proofs are close, but even they sometimes rely on setup or cryptographic assumptions. The dream would be a proof system that stands on its own, no bootstrapping required anywhere. Nobody has cracked this in practice yet. Every attempt reveals new tradeoffs. It's like trying to build a castle on sand and refusing to import any bricks.