Closed enricobottazzi closed 9 months ago
TLDR: the blinds don't affect the grand sum.
For normal advice, not just the blinds are random but the columns themselves are filled with random cell values after the circuit values. To the contrary, the unblinded advice columns are not filled with any random values:
https://github.com/summa-dev/halo2/blob/v2-PoC/halo2_proofs/src/plonk/prover.rs#L366-L378
The blinding factors also do not contribute to the KZG commitment.
https://github.com/summa-dev/halo2/blob/v2-PoC/halo2_proofs/src/poly/kzg/commitment.rs#L305-L316
Thanks for clarifying that. Closing the issue
The goal is to understand how the blinding factors affect the grandsum calculation and if a malicious prover can exploit its feature. For example by overflowing the grandsum.