Why do people assume to maintain privacy you have to do the entire computation locally?

Working on building zkTLS infra/products. A ruthless empiricist and meme connoisseur.

/opacity
/eulerlagrange

A naive implementation of letting the circuit make assumptions yields a 5x gain in speed.

It takes the same time to prove and verify.

https://github.com/Hmac512/noir-merkle-root-POC-lookup

…then use infra with highly optimized circuits to validate assumptions.

😅it’s actually simple stupid hahaha

BDFL of Quilibrium, Eng @ Farcaster, ex-Coinbase, always opinionated, never hydrated

/quilibrium

Because the brain bending arithmetic behind confidential compute feels like a paradoxical abstraction for most people

Technologist, Natural Philosopher, Wizard

It’s a combined PoS and PoW right? Actually have never looked to much into it.

When you say offload, is it like a traditional prover network?

Aleo has a clever proof of proof* consensus scheme. The side effect is an excess supply of prover compute that is auctioned to end users.

This offloads the heavy lifting from the users' mobile devices to specialized FGPAs in the "cloud."

*I think they have a different name for it

If you can do it locally, do it locally.

Every middleman needs to be paid, even if you can guarantee 100% privacy on remote computation (IP? Metadata?), I would still prefer a local first approach.