I used to read "trustless verification" on Babylon's TBV page as if it meant the whole chain of trust had been removed, not just one link in it.
Then I actually dug into what the proof covers.
Three pieces make it work: on-chain contracts, light-client proofs running on ZK-SNARKs, and independent indexers, all stitched together so Bitcoin's consensus and UTXO state can be checked from somewhere else entirely, without anyone having to trust one custodian to vouch for it.
Hmm.
Because the SNARK proves something narrower than the pitch suggests. It confirms a claimed Bitcoin state transition matches the rules baked into the circuit, the same circuit that encodes Bitcoin's own consensus logic, math checking math. It assumes two things without proving either. That the header and transaction data going into the circuit was accurate to begin with, and that the circuit itself was built correctly in the first place.
Closed the page and traced both assumptions back to where they actually live.
The first one is the indexer's job, a separate, swappable piece sitting next to the proof system. The second sits with whoever built and audited the circuit, a one-time trust question rather than an ongoing one.
I'm not saying this makes the design weak. Every ZK light-client scheme built on external chain data carries seams like these somewhere. Babylon just names them instead of folding them into the pitch.
If the SNARK verifies the state transition perfectly but either assumption was wrong, did the trustless part actually hold, or just the part that was never trustless in the first place?
#baby
$BABY
@BabylonLabs_io
$BABY
#baby
Then I actually dug into what the proof covers.
Three pieces make it work: on-chain contracts, light-client proofs running on ZK-SNARKs, and independent indexers, all stitched together so Bitcoin's consensus and UTXO state can be checked from somewhere else entirely, without anyone having to trust one custodian to vouch for it.
Hmm.
Because the SNARK proves something narrower than the pitch suggests. It confirms a claimed Bitcoin state transition matches the rules baked into the circuit, the same circuit that encodes Bitcoin's own consensus logic, math checking math. It assumes two things without proving either. That the header and transaction data going into the circuit was accurate to begin with, and that the circuit itself was built correctly in the first place.
Closed the page and traced both assumptions back to where they actually live.
The first one is the indexer's job, a separate, swappable piece sitting next to the proof system. The second sits with whoever built and audited the circuit, a one-time trust question rather than an ongoing one.
I'm not saying this makes the design weak. Every ZK light-client scheme built on external chain data carries seams like these somewhere. Babylon just names them instead of folding them into the pitch.
If the SNARK verifies the state transition perfectly but either assumption was wrong, did the trustless part actually hold, or just the part that was never trustless in the first place?
#baby
$BABY
@BabylonLabs_io
$BABY
#baby