BlogNotes from
Notes from
the lab.
Plain explanations of the cryptography, contracts and agents we work on.
What a zero-knowledge proof actually provesA zero-knowledge proof convinces you a statement is true without showing why. What that statement is, and what it leaves out, decides whether the proof means anything.25 October 2026 · Rump Labs
- Groth16, PLONK or STARKs: choosing a proof systemProof size, verification cost, prover time, setup and assumptions: how the three main families compare, and how to pick one for a real system.
- Anon-Aadhaar, explainedHow a zero-knowledge proof over the Aadhaar secure QR code lets someone prove they hold a valid ID, and share only the attributes they choose.
- Agents that hold up on real data: evaluate before you shipA demo shows an agent working once. An evaluation set built from real cases, and a threshold agreed in advance, show whether it works on the cases you actually have.
- What a pre-audit contract teardown is (and isn't)A teardown is a hard, line-by-line read of a smart contract before mainnet. It is not a formal audit, and it is most useful when it doesn't pretend to be one.
- Secure computation and the quantum questionWhat secure two-party computation is, why fairness is hard, what changes when quantum computers enter the picture, and where IACR ePrint 2026/883 fits.
Next stepRead enough?
Read enough?
Talk to the authors.
Thirty minutes with an engineer, free. The people who write these are the people who build the systems.