Eugen,
* Eugen Leitl (eu...@leitl.org) [110618 04:42]: > On Sat, Jun 18, 2011 at 12:21:42AM +0530, Sirtaj Singh Kang wrote: > > This is a very thorough and compelling takedown of bitcoin. > > Currencies are consensual belief systems. No they are not. Example: In the USA, I *must* pay taxes in USD, not bitcoins, regardless of my level of belief or consent in either. > Bitcoin will remain useful as long as there are > no effective attacks against the cryptosystem and/or > the infrastructure, and people continue > to believe in it. The idea of Bitcoin is grand; however, it seems deeply flawed on many levels. First, here's my unified framework for understanding the relationship between fiat currency based systems and n-way barter asynchronous systems: [1] Multiflow on timed, colored, prioritized Petri nets with variable & uncertain topology approximates barter. Currency is a bartered "color". [2] Colors in a barter Petri net have time & place dependent {estimated|actual|perceived} marginal: o "intrinsic local" value o "gamed network" value [3] Modeling currency as a multiflow-bartered color, Petri net place color values depend on place intrinsics, topology forecasts, and supply/demand game theory. [4] "Faith" in a bartered item is reliance on estimates of its {present|future} "local intrinsic" and "gamed network" {estimated|actual|perceived} marginal values. [5] All items in barter exchange are created "by fiat" when they are created by someone else. For these reasons, it doesn't bother me at all that Bitcoin is a "fiat" currency per-se, but it *does* bother me that it is not a proxy for something with intrinsic local value to someone somewhere. Essentially, we are asking people to barter things the make or do that have intrinsic local value for something that has none at all -- without the force of law twisting their arm to do so (as the US government does when I pay my taxes in USD). Further, Bitcoin is encouraging the production of useless items (ie: key block solutions) merely because of the mathematically easy with which they can be verified, and the apparent difficulty of faking them. To do that robustly in the context of a computational arms race, Bitcoin makes these blocks harder to solve as things go along, thereby effectively *centralizing* the process of mining coins. It all comes down to blocks/$ efficiency, so if you can't compete with Exxon access to cheap energy, or the Fed for cheap USD to buy energy/computers, then you can't economically mine coins in the long run. This argument leads to a very sobering conclusion: centralized mining is implicit in the trust model Bitcoin uses. Ouch. Why would I want a purely faith-based "color" in this barter Petri net anyway, if I could just as well get one with some intrinsic value? I'd prefer to trade in proxies for something with easily convertible worth outside the system during both the bootstrapping phase and afterwards. What Bitcoin mining looks like to me is a computer-science-only solution to a verification problem that has some extremely robust solutions outside of the world of computer science: multiple independent audits. While it's true that someone might pay off N auditors (thereby creating fraudulent receipts for barter), lawsuits and insurance have already solved this problem fairly well in the real world of commodities markets, and in a way that does not count on some hashing cipher never being broken. Decentralized webs of trust could make a system like this extremely resilient to fraud. Note that if the "colors" in a barter Petri net are proxies for commodities, then you really *do* have a robust way of auto-scaling currency supply to market size, and a decentralized system of currency production (ie: audited receipts for commodities) that encourages the production of things with intrinsic worth according to their actual market value. Bitcoin has none of this. Bitcoin's choice of trust model for mining seems to make it not very different from the USD fiat currency we have today: controlled by energy/finance oligarchs, inherently worthless, and divorced at point of injection into the system from meaningful market signals. > Bitcoin is unlikely to be it, but it will have successors and > competitors. Yes, I agree. -Jon