pull down to refresh
30 sats \ 4 replies \ @Scoresby OP 23 Oct \ parent \ on: "Quasi-consensus": help me understand this conversation with Chris Guida bitcoin
I should have been more clear: when I said "this assumes a static system" I should have said "this assumes a static mempool system." The rules that validate blocks can be treated as static.
Mempool policies are effectively infinite in variety and pretty much can be ignored.
What I'm trying to argue is that any layer other than consensus ends up being irrelevant.
This is because the block validation (consensus) rules are designed to incentivize the mining of transactions that the "social layer" doesn't like. How the transactions reach miners is irrelevant, only important is the fact that there are effectively infinite ways of transmitting transactions to miners.
Ok, I think I understand what you're saying. It's actually a very PhD economist way of modeling the world ("the incentives must imply this outcome---how it gets to that outcome is irrelevant")
Now that I understand, I'd say I agree with you locally, but not globally. Locally in the sense that, right now, it seems like consensus rules are enough to ensure non-censorability.1
But I'd argue that this is only a function of current circumstances. Because miners' incentives exist both on and off chain, governments could, theoretically, make it so difficult or dangerous to mine disfavored transactions that miners refuse any side channels. In the current state of the world, that would be difficult to pull off, because we don't have a one-world government, and because internet privacy tools may be enough for individuals to circumvent national-level rules.
Still, the point I'm trying to make is that I don't think we should be lulled into thinking that a properly designed consensus layer is all that's required to keep bitcoin decentralized and non-censorable. I think that's perhaps too rosy of an outlook, and would tend to make bitcoiners complacent about social level threats.
Footnotes
-
You'd still have to respond to the question of "If filters don't work, why is there a cutoff at the OP_RETURN limit?" You answered earlier, "That indicates there's no demand above the limit." But even if there's little demand above the limit, you shouldn't see such a steep cutoff right at the limit. It should be a smoother dropoff of demand. Personally, though, I don't think this cutoff is relevant to the efficacy of filters, I think it speaks more to the impact of software defaults. ↩
reply
"If filters don't work"
I'd like to put the statement "filters don't work" even more strongly: "filters can't work."
If filters can be effective in the face of sustained demand for a certain kind of valid transaction, Bitcoin fails to deliver on its value proposition. There is absolutely nothing that stops a state actor from attacking Bitcoin via a giant network of state-controlled nodes. And if our defense relies on some social layer, the project fails.
It's not like this threat was never considered. The block validation rules of bitcoin are designed to resist exactly this kind of attack. I'd even go so far as to say that solving the problem of a social layer is what Satoshi was trying to solve with Bitcoin in the first place.
The only social layer I believe is necessary for Bitcoin is agreement on block validation rules (what I've been calling 'consensus rules'). This was Satoshi's big innovation: he created a way that if users agreed on a minimal set of rules, no other differences were relevant.
The problem Satoshi was trying to solve was how to create a peer to peer money, a money with a third party, that was stable -- not in the sense of price or anything like that, but in the sense of useful to anyone who wanted to use it, stable in the sense of not being subject to the turmoil of disagreement in a social layer.
If a social layer beyond agreeing to block validation rules is necessary to make Bitcoin work, we are immediately confronted with the problem of how to decide things in this social layer. Is it by vote? by CPU power? by who has the most coins? In answering this question, I believe you will find that the only known solution is the same one Satoshi came up with for block validation rules. But this is just to say the social layer can only be coming to consensus on what the block validation rules are. Nothing else matters.
And so, yes, I believe
a properly designed consensus layer is all that's required to keep bitcoin decentralized and non-censorable.
As to your distinction that block validation rules are enough locally but not globally
Because miners' incentives exist both on and off chain, governments could, theoretically, make it so difficult or dangerous to mine disfavored transactions that miners refuse any side channels.
Check out Rob Warren's recent X thread on how difficult it is to control hash power: #1263431. A state may do everything it can to put pressure on miners in its jurisdiction, but the permissionless nature of mining is such that we may hope it is in the self-interest of miners outside such a state's jurisdiction to put disfavored transactions in blocks.
Therefore, I maintain that consensus rules are enough to maintain censorship resistance and that it is not a "function of current circumstances."
reply
I think your point is that, as long as there's an on-chain incentive to mine the most profitable transactions, then the most profitable valid transactions will find their way into blocks.
I don't disagree with that, but my question is at what cost and at what degree of usability? I think censorability is a spectrum, not a 1 or a 0. So my argument would be that bitcoin's built-in incentive structure can keep censorship<1, but it's uncertain if it can maintain a high degree of usability and mainstream adoption.
So, I still maintain that social norms, ethos, and other soft crap like that still matters for the tradeoff matrix between censorability and usability on bitcoin
reply
The problem Satoshi was trying to solve was how to create a peer to peer money, a money without a third party, that was stable -- not in the sense of price or anything like that, but in the sense of useful to anyone who wanted to use it, stable in the sense of not being subject to the turmoil of disagreement in a social layer.
I never see the typos until after 10 minutes is up...
reply