0 sats \ 2 replies \ @elvismercury OP 3 Dec 2023 \ parent \ on: What's really scarce in btc? bitcoin
The proof of work that accrues to a project is a function of its commitment to a particular instantiation of an idea. (I'm using PoW in its most literal instantiation here -- hashpower. If you mean "proof of work, in aggregate, across all system activities" then I'd say that's isomorphic to the idea I'm proposing.)
Proof of work is being directed to something. What is that something, and why? That's the distinction here. Here's a thought experiment:
Imagine that a supervillain devoted to the enslavement of mankind gets control of 75% of btc hashpower, and proceeds to censor transactions, double-spend strategically, decry carnivory, etc. What happens?
If the community -- defined in the hazy way that's necessary when talking about something as nebulous as the people who have some kind of stake in btc -- believes, collectively, that this state of affairs violates the ideas that it's committed to, they'll fork away from that chain, despite its preponderance of PoW, and direct their various efforts toward the new enterprise. The fact that there's exahash sitting on the sidelines is not the issue.
In the same way that hashpower follows price, and not the reverse, hashpower is a function of commitment, which itself is mediated by price.
Imagine that a supervillain devoted to the enslavement of mankind gets control of 75% of btc hashpower, and proceeds to censor transactions, double-spend strategically, decry carnivory, etc. What happens?
That's a really good thought experiment, but the point that's missing is that the coordination problem.
That's exactly what the Byzantine Generals (and Satoshi's update "the king's wifi") are about - if you presuppose the ability to globally coordinate and trust non-defection, then proof of work as a solution is not needed.
And indeed, your analogy is not at all contrived - it's the true state of the world. Governments and other large powerful entities could coordinate to attack bitcoin at the level of hashpower, but they have not, and I contest, they will not succeed in doing so, exactly because of that inability to coordinate at a sufficiently overwhelming scale (in geography, in time, in category etc.).
We've seen an example already - China hosted the majority of the world's hashpower and then suddenly outright banned mining - but they couldn't get other large powers to coordinate, so Bitcoin was (more or less!) fine.
If the community -- defined in the hazy way that's necessary when talking about something as nebulous as the people who have some kind of stake in btc -- believes, collectively, that this state of affairs violates the ideas that it's committed to, they'll fork away from that chain, despite its preponderance of PoW, and direct their various efforts toward the new enterprise. The fact that there's exahash sitting on the sidelines is not the issue.
In the same way that hashpower follows price, and not the reverse, hashpower is a function of commitment, which itself is mediated by price.
Yes I always get this line of counterargument when I say these things, and of course, it must be partly true - one of the main two factors here (let's say "network effect" and "proof of work") doesn't work without the other.
The reason I see network effect as a proximate cause, not an ultimate cause, is a matter of simple logic - network effect could somehow arrive in any other system of money, why doesn't it (hence me saying "unstable equilibrium")? You're not getting at the ultimate cause of bitcoin's success if you focus on network effect.
reply
Yes I always get this line of counterargument when I say these things, and of course, it must be partly true - one of the main two factors here (let's say "network effect" and "proof of work") doesn't work without the other.
All those feedbacks and recurrences add up to quite a muddle, I think. You got peanut butter in my chocolate, etc.
I keep intending to play around with a differential equations model so that at least it could be talked about concretely and coherently, but have not got around to it. It's probably the best (only?) way to push harder into the distinctions you and I are proposing. I know there are other models, e.g., valuation models, that purport to do this, but I last ran into them years ago and can't remember what they are to see if there's inspiration there.
Anyway, I appreciate your comments. I'll keep noodling on them.
reply