Antelligence ZK
An experiment in proving facts about Antelligence simulations without revealing their inputs.
- Status
- Cryptographic prototype
- Year
- 2026
- Role
- Solo, on the SP1 template
- Area
- Research
In plain English
People want to trust a simulation's results, yet some of its inputs may need to stay private. Zero-knowledge proofs (math that proves a claim without showing the data) offer a way to have both. This is a small Rust experiment applying that idea to Antelligence.
How it works
- Start from the SP1 template, a starter project for a zkVM (a virtual computer that produces proofs).
- Feed in the simulation's parameters as private inputs.
- Check that each input sits in its allowed range and that the kill-rate arithmetic adds up.
- Publish the result in a format a smart contract (a program on a blockchain) can read.
What I did
I adapted the SP1 template on my own, adding the Antelligence-specific checks in Rust. The surrounding project structure comes from the template.
What I learned
Checking inputs is the first step, and proving that a whole simulation ran correctly is the bigger goal. Seeing that gap clearly showed me where the hard part of verifiable science lives.