Rusty Russell: The Corrosion of Ethics in Cryptocurrencies
medium.com
medium.com
I look forward to the day where cryptocurrencies are more commonly understood (and thus feared) and the technology isn't chronically lagging behind the hype.
When you bought, someone else was saying the same thing about you :)
This logic is flawless until one day the market collapses and we find out who the "greatest fools" are.
Thus the hierarchy of idiocy. I believe the decision, as someone who has studied CS and economics and became deeply interested in the white paper and the game-theoretic and macroeconomic consequences of cryptocurrencies, to buy post-MtGox (which I did) vs. everyone and their grandmother feeling FOMO (and quite literally expressing this as their motivation) as they seek to buy during last year's explosion in popularity, suggests quite different risk assessments, neither of which I can morally defend in retrospect.
clearly, cryptocurrencies offer significant value over fiat currencies and their advantages will only continue to grow as development maturity surrounding the actual technology becomes more solidified
0: http://isg-one.com/research/research-detail-page/cryptocurre...
I bought at $3 and I started selling at $200. I sold a ton at $1000. Then a little at $5k and the rest at around $15k.
The entire time I held them I thought "the government will either ban or regulate this and they'll probably ban it" and you know what? It doesn't stop you from making money anyway.
Now that's a fun story right? Look at smart 3pt14159. What you don't see is the other 7 things I did with 0 payoff whatsoever. I bought a ton of emoji .ws domains. Why? I dunno, I stupidly thought they'd be the only TLD that would render as emojis. I hindsight, of course not, and once the others came out I reduced the number of domains I controlled. Nobody sees the lost weeks and $$ there, unless I happen to bring it up in a comment like this.
But the core idea behind both of those actions was this: Look for the low probability, high payoff investments. Unless you have $100m+ it's usually worth your time and most people don't have an intuitive grasp of inelastic supply curves and their impact on price.
At 2,000$ each they can't have 1000x returns. So the person buying at 2000$ is more likely to be a sucker than someone buying at 3$, but less likely than someone who buy's at 15,000$.
95% in ultra-safe government bonds and bank CDs.
5% in ultra-high-risk moonshots.
Cut the losers quickly and let the winners run.
It is not. It is idiotic, however, to buy something that you don't know it's probabilities of payoff.
The fact that so many people could upvote this is a troubling signal of the misunderstanding and lack of awareness that exists even among educated, intelligent professionals.
Categorically speaking, they are the same. The difference is one of degree: financial markets consist of many two-party competitions with material loss, and sports consist of multi-party or competition with non-material loss.
Yes it is. The frequency of events occurring in a specific measure of time is both mathematically and financially meaningful. In particular, the periodicity (if any) of such events is very important.
> Also, mining makes crypto coins negative sum, to be positive some they would need to generate more value than the cost of mining which is unlikely.
At this point you’re introducing new measurements, and you need to rigorously define the “game” to determine if it’s zero sum or not. Proving that a set of objects and their actions on each other is zero sum is very difficult because it’s very rare. As a trivially obvious example, for any integer n in the infinite set of all integers Z, there exists only one -n such that the sum of the pair is exactly 0.
Most real world systems do not satisfy that property, and you’re sort of hand-waving to make it fit. When you model a bid and an ask, it satisfies a zero sum because trading is specifically designed to do that. But the aggregate activity of the market is not zero sum, because it results in net positive growth in aggregate financial value over time.
> Also, mining makes crypto coins negative sum, to be positive some they would need to generate more value than the cost of mining which is unlikely.
If you’re defining your game as the interactions comprising cryptocurrency mining and cryptocurrency trading, then mining generates net positive value in conjunction with trading. If it did not, it wouldn’t be sustainable due to gradual decay over time. Again, frequency and periodicity of profit and price movement here are meaningful. We don’t have to talk about what’s likely, because it’s demonstrated for us economically.
The only way forward for you to assert that cryptocurrency trading or mining is negative sum or zero sum is by introducing new, reasonable interactions to the game and quantifying their relationships with the existing interactions.
More abstractly miners must be paid to do work. Billions of hours worth of compute time has direct material and energy costs.
No, because the aggregate value of the market is increasing. And what you’re describing is a liquidity cost, not a value cost - those coins still exist and their value has not changed, your ability to capture their utility has. The conventional markets are very similar in this regard with respect to derivatives and trading fees.
> More abstractly miners must be paid to do work.
Yes, and at the same level or abstraction: if the activity was not positive, it would (by definition) not be sustainable, given that we are measuring interactions between miners (and their aggregate costs) and the market (and its aggregate costs). Which brings me to...
> Billions of hours worth of compute time has direct material and energy costs.
Yes, and the same principle - the compute time and energy costs have quantifiable costs, and those costs are under the potential profit of mining.
Note that I’m not saying growth is sustainable in perpetuity (that’s thermodynamically impossible). I’m saying that the game, with the interactions and constituents you’ve described so far, is positive sum, and therefore sustainable with growth for the foreseeable future.
Worse you are proposing exponentially increasing investments forever.
I’m sorry, I don’t really know how to express the point more clearly at this juncture, so I guess this will have to do:
1) in finite, discrete measurements of time, the activities of a game can be positive sum;
2) corollary to point 1, the aggregate market valuation trends upwards over time.
You keep focusing on infinite or underspecified measures of time for your arguments, which is frankly incoherent for the discussion at hand. This indicates to me that you don’t believe anything can be zero or even positive sum, because according to your logic - as literally stated - the value of gold is negative sum because resources are finite and time is not, so gold will eventually be stolen.
Not only is the probability of a thing being stolen outside the interactions of the game we’re considering, it’s not relevant to valuation unless you rigorously quantify its cost (such as through insurance) for a discrete measure of time. In essence, you’re conceptually refuting central tenets of game theory by invoking Murphy’s law.
Consider planting crops. The 'output' of the system over time is lower due to entropy, but the value of chemical energy in corn is vastly higher than the value of nuclear energy in the Sun. note it's the transformation that's providing value.
The problem is coins and gold are not being transformed AND there are costs that are inherently part of the system. Poker with N players is ~zero sum, poker with N players and a house stake is negative sum among the players. Clearly, coins follow the second example with a house (equipment) taking profits off the table.
Now, if you want to say commerce is positive sum and coins can provide value as a medium of exchange that's an argument around their relative utility. Which is currently generally* less than money as you go money > coin > stuff vs money > stuff. However, coins as an investment is a different argument. If I hold a coin for 200 years then it's not part of any transaction.
* Coins as a means of circumventing currency controls is yet another argument.
- to avoid obsessing over something shallow I ended up focusing 2x times more on learning and coding, so vanity made me more active in a very significant way
- if it turns into a passive income and I can leverage a bit more, I plan to give regular food money to homeless people. I find the idea of grabbing a few bucks from the cryptowealthy to the poorest fun and worthy. (robinhoodcoin)
- also, when I do PE when reading charts so it's never totally wasted
Picture 500 years of interest vs a small bit of gold. Granted this assumes they buy the gold as an investment but if it's not sold for 500 years that seems reasonable.
Except I am saying it goes to 0 so they don't actually get any gold. So, it's option A get nothing or option B possibly get something of value. I don't see how option A is better.
While the single dimension of buy and sell price for any asset can be negative sum if it depreciates after your purchase, whether or not the complete value of the asset is net positive or net negative is a function of 1) the quantifiable advantages, like (but not necessarily) dividends or risk reduction, and 2) the amount of time spent holding it. This second point is also why the frequency of price changes (and net change in any discrete measure of time) is a very important part of the analysis.
Say I buy 1,000,000$ worth of gold and agree the vault owner can sell gold to cover the costs for the next 100 years. Alternatively, I can spend 1,000,000$ but get less than 1,000,000$ worth of gold to pay for the vault. In either case in 100 years someone needs to keep paying for the vault. Any finite amount of gold is not enough to pay for it's storage costs forever.
PS: You can spend less on security but you risk having nothing of it's stolen. Given enough time and it will be.
You can quantify those costs for the same measures of time that I keep talking about in this thread. As I keep saying: the cost of storing the asset is economically, demonstrably lower than the value of the asset for any meaningful, discrete timespan.
> PS: You can spend less on security but you risk having nothing of it's stolen. Given enough time and it will be.
By this logic, literally everything is negative sum according to thermodynamics. You’re not using a reasonable, accepted definition for this concept in the context of game theory.
We know the heat death of the universe is going to happen, so does that mean the aggregate activity of the market is negative sum? This isn’t just incorrect, it’s incoherent.
Gold sitting in a vault stays gold sitting in a vault the outputs are identical in form and lower in amount.
PS: A bank safety deposit box costs ~60$ a year store 1,000$ worth of gold for 20 years and you lost money unless it's worth more than twice as much inflation adjusted. Store 1,000,000$ worth of gold costs more than 60$ / month, but not 1000x more so it's not as pointless.
Speculation is more zero-sum, but can send useful price signals when tied to an asset that relates to production.
However, speculation that is completely disconnected from production is just gambling... zero-sum. Actually negative-sum, if we include wasted energy and time.
Yes.
> Speculation is more zero-sum, but can send useful price signals when tied to an asset that relates to production.
Yes, with the qualifier that the individual trades are what’s zero sum, with the aggregate activity and emergent signals causing the positive sum in real investment.
> However, speculation that is completely disconnected from production is just gambling... zero-sum.
No. Uninformed speculation is less useful for price discovery than informed speculation, but it does not nullify informed speculation. Furthermore, efficient price discovery is facilitated by liquidity, and uninformed speculation is a net improvement in liquidity.
Furthermore, the incentive may be much higher than is actually required, and creates a lot of collateral waste in the process.
Every bit of time, energy, and resources spent in this speculative endeavor crowds out other activities. And a large amount of this effort is just spent on a cybersecurity arms race over these hypothetically valuable tokens.
I realize you were trying to make a point about speculation in general, such as with futarchy. Any kind of speculation over things related to real production, including bets about CEO performance, would have some obvious informational value. But bets about sports do not.
I certainly believe there is useful hedge functions in forward pricing expectations. Speculation on coins just doesn't feel like it has that dimension. Its gambling but instead of taking on "the house" its taking on somebody else.
"Gamble responsibly" is one of those "half pregnant" statements which raises a wry smile but the underlying message is "you can't"
Considering hydro power supposedly consumed by mining would be generated no matter what (the costs are largely fixed), almost all the power used by mining is ultimately produced by coal plants, which are the main source for adjustable electrical generation in China.
In the USA, mines are mostly located in 3 counties in central Washington State, along the Columbia river. Power rates there are much cheaper (around 2.5¢/kW), which is the main incentive to put your mine there.
Mining on coal is stupid because it's expensive and miners want to make a profit: https://en.wikipedia.org/wiki/Cost_of_electricity_by_source
You've misunderstood the argument: "being close to" implies that the electricity would otherwise go to waste.
But because transmission losses are low, distances to markets aren't as large as one might believe, and the Chinese weren't stupid enough to build hydroelectric dams in places where the energy would uselessly go to waste, that argument is wrong.
A true peer-to-peer digital currency enabling you to make payments globally without a third party being able to censor has very real benefits.
For example:
* Donations to wikileaks or similar organizations
* Payments in gray areas, for example porn
* Transfers to and from inaccessible areas, for example war zones or dictatorships
I hear much of the electricity is hydro power in China.
Lastly, Bitcoin isn't finished yet. I was dialing up to the internet on a 56kbps modem just 20 years ago. I can't imagine where crypto currencies will be in 20 years.
How is that any different?
The answer is we consume for our understanding (or assessment) of value. Bitcoin is not burning electricity. People are burning electricity to generate bitcoin.
I was never a true believer, but even being skeptical things are worse in the cryptocoin world than I even thought they would get before the hype died down and it turned into whatever kind of niche hobby thing it's gonna be in the long term.
Just take a look at the developers behind Bitcoin Core. It's difficult to ignore the obvious conflicts of interests of doing everything to block on-chain scaling while their company's business model is off-chain solutions.