pull down to refresh
Indeed, they should be seen as custodial services.
Maybe I'm missing some of the debate here, because I always thought of them as custodial services. I don't follow the industry as closely, nor some of the technical claims, so I probably wasn't listening to certain inflated claims the L2 founders were making, trying to pretend that their L2s are self-custody (which they aren't and never were).
Even Lightning I'd say is not 100% self custody, but it limits the problem of trust to peer-to-peer relationships.
Lightning is like one person keeping a running tab with another person. And the network is formed out of these individual person-to-person tabs.
The other federated solutions are like one person keeping a running tab for many other people. They become the custodian and the trust requirements are higher.
Here are three liquid wallets. All describe themselves as self-custodial.
Having not followed the chatter much, on what basis did they call themselves self custodial? Is it because you can unilateral withdrawal?
I think so, but clearly you can't do unilateral withdrawal because the network is paused right now and because there isn't enough btc for all the l-btc. So they were wrong on that count.
I was looking into this a bit because of the SN Spark Wallet. (I don't follow the chatter, but I try to look into whatever service I use before I start using it).
The claim of unilateral withdrawal on Spark is fairly technical surrounding the creation of timelocked multisig transactions. But I came away with the thought that under certain network conditions, unilateral withdrawal is indeed feasible. But it depends on your software implementation to properly store and broadcast that transaction in time. I think you need to trust the L2 network's watchtowers too.
Based on my understanding of the Liquid hack, there's likely a disconnect between what the software implementation shows as your available balance and the underlying timelocked transactions that make up the unilateral withdrawal claim, i.e. the sum of the displayed wallet balances exceed the sum of the unilateral transaction claims.
All this to say, the trust assumptions are extremely technical. If you don't understand it, you should probably treat these L2s as custodial. Even if you do understand it, you'd understand that the self-custody claims only apply under certain technical assumptions and conditions.
but clearly you can't do unilateral withdrawal because the network is paused right now
That's where I'm a bit confused. As I was reading into Spark, my understanding is that you should have a signed multisig (unbroadcast) TXN that you can broadcast to the mainchain in order to initiate the unilateral withdrawal, in a way that doesn't depend on other Spark counterparties. Not sure if Liquid works similarly. But again, it depends on either you knowing how to broadcast this TXN, or your wallet software being able to.
That's my (limited) understanding, anyway.
I think you are accurate in your understanding of Spark. However, I'm pretty sure that Liquid never made such promises (despite wallets being built on liquid claiming to be self-custodial).
As far as I know, Liquid federation could always stop the chain or prevent your liquid transaction from being moved.
Interesting. I never really looked into Liquid. But again, it reinforces the point that the "self custody" claim is highly technical and what that really means is totally dependent on implementation details.
🤢🤮
Justin's point in his comment on this post makes sense:
So, I'm not arguing that we should abandon things like ecash and spark or even liquid, but I think I understand them as custodial services now -- which is more or less what you may be saying when you say they are fractional reserve banking.