I ran the TBV testnet flow last night expecting the vault itself to be the trust-sensitive part. It's not. Locking BTC and seeing it show up as Aave v4 collateral was clean and fully self-custodial, no complaints there.
What got me was liquidation. Bitcoin doesn't actually verify any of it. The vault's spending rules are checked off-chain by a garbled circuit, using a cut-and-choose setup — meaning if someone posts a bad proof, a specific party has to notice, run the verifier, and post the correction before a timeout runs out. No one watching in that window, the bad proof just goes through.
It's reminded me of a building with fire alarms but no sprinklers — the alarm only works if someone's actually there to hear it and act. TBV's "trustless" custody is the fireproof walls. Liquidation correctness is still the alarm system, and it depends on a whitelisted set of liquidators staying online and watching the price oracle, not an open permissionless set.
So there are two separate trust claims stacked in one product, and only one of them is actually trustless. Custody, yes, fully. Liquidation integrity, no — it's trust-minimized, resting on a smaller group doing their job on time.
Not a dealbreaker on testnet. But it does mean "trustless Bitcoin vaults" is doing more marketing work than protocol work for that one specific piece.
Curious whether Babylon opens the challenger/liquidator role up as the product matures, or keeps it whitelisted by design since Bitcoin settlement speed just doesn't leave room for an open race to catch fraud in time.
@BabylonLabs_io $BABY #baby $IDOL $UAI #Babylon
What got me was liquidation. Bitcoin doesn't actually verify any of it. The vault's spending rules are checked off-chain by a garbled circuit, using a cut-and-choose setup — meaning if someone posts a bad proof, a specific party has to notice, run the verifier, and post the correction before a timeout runs out. No one watching in that window, the bad proof just goes through.
It's reminded me of a building with fire alarms but no sprinklers — the alarm only works if someone's actually there to hear it and act. TBV's "trustless" custody is the fireproof walls. Liquidation correctness is still the alarm system, and it depends on a whitelisted set of liquidators staying online and watching the price oracle, not an open permissionless set.
So there are two separate trust claims stacked in one product, and only one of them is actually trustless. Custody, yes, fully. Liquidation integrity, no — it's trust-minimized, resting on a smaller group doing their job on time.
Not a dealbreaker on testnet. But it does mean "trustless Bitcoin vaults" is doing more marketing work than protocol work for that one specific piece.
Curious whether Babylon opens the challenger/liquidator role up as the product matures, or keeps it whitelisted by design since Bitcoin settlement speed just doesn't leave room for an open race to catch fraud in time.
@BabylonLabs_io $BABY #baby $IDOL $UAI #Babylon