Comment by Nursie | Hacker News Reader

Comment by Nursie | Hacker News Reader

Parent Full thread

I've enjoyed watching the Bitcoin experiment. It's been enlightening in a variety of ways, from technically to socially.

I've been fascinated and appalled in equal measure at the fanboy community, at the intolerance of criticism that sprang up very quickly, and how strong feelings ran (likely because of financial investment in the tech).

It's also been interesting watching it go from simple CPU mining, to multiple GPU rigs in dorm rooms, all the way through FPGA and then to massive installations of custom ASIC miners.

But I've always hoped it wouldn't go mainstream for two reasons - limited supply with weighting in favour of early adopters, and the massive electricity costs of the 'mining' and transaction validation process. Scalable, competitive proof of work systems for a widespread currency are an ecological disaster in the making, and deflationary currency with a handful of early users controlling a huge proportion of the total currency supply... these aren't "features".

I'll be very interested in what happens next, and for the reasons given I hope it's not just a BTC clone with better governance.

I 'll also notice what the whole experiment proved: That money can perverse moral incentives. The electricity/wasted energy issue you mention is just one of them. The amount of time spent on building and specializing the hardware is another. All this was in the hope of the abstract goal of creating money, not wealth. People weren't so zealous in building specialized hardware for, say, protein folding. Bitcoin advocates usually put forward "the good of humanity" argument, yet the product by design puts individual interests above all other interests.

In that vein, I'd like to point out Gridcoin[1], a crypto-currency that rewards you for throwing your computing power at a BOINC project (basically scientific data crunching) [2] instead of wasting it on pointlessly solving crypto problems. It's certainly not perfect, but an interessting idea nonetheless.

[1] http://gridcoin.us/ [2] http://boinc.berkeley.edu/

That's quite cool, I like that idea a lot more. I think I'd still rather a currency that didn't rely on burning power as fast as possible, but using that power for good is, well, good :)

The current system for USD is pretty close to ideal. Short term money is creating through lending, long term money through bonds.

It's a great mechanism and could be readily adapted to digital currencies. Have a group of banks that may issue loans in the currency and have a central authority that will issue bonds and pay interest on them in the currency. Set the lending rate to under the bond rate to "bootstrap" the initial currency flow, then manage it from there.

Management could be automated by computers to constantly find the optimal rates for loans and bonds.

Well, there needs to be some proof of work, something that is hard to do at scale but easy for individuals.

To that degree, Bitcoin has failed, because massive SHA256 crunching is now easy to do at scale and hard for individuals.

More varied challenges, like BOINC, or PoW that is bandwidth, connectivity, latency, or disk-space based, will provide a lot more value to the community and make it easier for individuals to contribute. I eagerly await the day that one of these alternatives becomes mainstream.

Alternatives are good but.... well, disk-space based would push up the price of hard drives. Bandwidth connectivity and latency could again be owned by big players...

I don't know. What I do know is that in a centralised crypto-currency system you don't need any of this as the issuing authority merely signs the transactions and we're away. I hope that some bright spark comes up with a way to make a decentralised one almost as simple...

I like the moral of this idea

> I 'll also notice what the whole experiment proved: That money can perverse moral incentives. The electricity/wasted energy issue you mention is just one of them. The amount of time spent on building and specializing the hardware is another.

The existence of high frequency trading already proved this. All that technology built and maintained, all that electricity wasted, merely so that one company's bots could play a zero-sum game nanoseconds faster than another company's bots, undermining them.

> That money can perverse moral incentives

This isn't something newly discovered, I hope.

Haha, I like the satire :-)

I was hoping gp was being sarcastic

> limited supply with weighting in favour of early adopters

How is that not true for the royal families of Europe or pretty much all the billionaires in China/India/Russia? Of course, it is not a desirable feature as we've discussed in criticism of pg's essay but what option do we have?

>> How is that not true for the royal families of Europe or pretty much all the billionaires in China/India/Russia?

While I agree there are problems when individuals can accrue control of such a disproportionate amount or influence and economic output... I don't think it's quite the same.

2/3 of all the Bitcoin that there will ever be already exist. The number of people involved in Bitcoin is comparatively small. Were it to become a mainstream world currency, on the scale of the dollar, some of these folks would likely become trillionaires. Those oligarchs in India, Russia and China are wealthy, but they haven't got such a huge proportion of all currency sewn up. And because of the nature of 'fiat' currency, we can always make more. With BTC this isn't the case.

Money itself is a legal phenomenon, the laws can be changed to amend how money functions.

From the bitcoin.org FAQ:

>Isn't Bitcoin mining a waste of energy?

>Spending energy to secure and operate a payment system is hardly a waste. Like any other payment service, the use of Bitcoin entails processing costs. Services necessary for the operation of currently widespread monetary systems, such as banks, credit cards, and armored vehicles, also use a lot of energy. Although unlike Bitcoin, their total energy consumption is not transparent and cannot be as easily measured. Bitcoin mining has been designed to become more optimized over time with specialized hardware consuming less energy, and the operating costs of mining should continue to be proportional to demand. When Bitcoin mining becomes too competitive and less profitable, some miners choose to stop their activities. Furthermore, all energy expended mining is eventually transformed into heat, and the most profitable miners will be those who have put this heat to good use. An optimally efficient mining network is one that isn't actually consuming any extra energy. While this is an ideal, the economics of mining are such that miners individually strive toward it.

> Spending energy to secure and operate a payment system is hardly a waste.

That's not the question. The question is if more energy is being spent than is necessary to achieve the goals of a well-secured, well-operated a payment system, and I think there are very strong reasons to believe the answer is yes. (Among many other things, Ripple and Stellar are existence proofs of being able to get something with comparable security and much lower energy usage.)

> Furthermore, all energy expended mining is eventually transformed into heat, and the most profitable miners will be those who have put this heat to good use.

what

> While this is an ideal, the economics of mining are such that miners individually strive toward it.

And it is equally valid to say that the economics of cloud computing are that cloud hosts should individually strive towards it, too, but is AWS so much as keeping a single pizza warm?

Is there any previous case in human history where "Negative externalities could be turned into profitable positive externalities with cleverness" has worked?

I disagree that the optimal network doesn't waste any energy - any hash attempt that fails is a waste of energy. As more people compete and the difficulty goes up, more and more energy is consumed.

I understand that this is the way BitCoin protects itself. I want to be clear that I am saying I think this is a bad way of doing things. As someone that understands a bit about crypto - most crypto schemes that are hard to break are not hard to break because the encrypting device has worked so hard at encrypting them. They're hard to break because they have been designed in a mathematically sound way.

I hope this endless grind is not something inherent to the idea of decentralised crypto-currency, but if it is I don't think the decentralisation aspect is worth it, and as we see in this article it can be illusory anyway and the amounts of energy involved are pretty huge.

>> Furthermore, all energy expended mining is eventually transformed into heat

This is just such a cop-out!

Well, there's always the conspiracy theory that the bitcoin network (and the blockchain) is a harness created by the NSA to crack encryption around the planet.

It uses vastly more energy than payment processors that have vastly more throughput. And the ideal miner that finds a valuable use for the waste heat is a fantasy, and it's an incredibly inefficient way of generating that heat.

I think at the end of the day there's an important axiomatic difference.

Are you willing to trust any central authority? Maybe a government, maybe a set of independent businesses you can chose among, maybe a specifically selected advisory committee, one really cool dude. Just anything.

If you are, then burning coal and compute cycles is just insane and wasteful.

If you aren't, then you might be willing to admit that from some angles it is an inelegant solution, but it is a solution.

Taking Bitcoin on its own terms, as a solution to a problem you may not personally care about, the wastefulness of the process isn't really a problem. If anything, I agree with what someone else said down thread -- it would be better if mining for Bitcoin wasn't profitable, and it was a lottery people played at a loss (wasting even more resources!). That would discourage the centralization of mining power. IMO the biggest problem Bitcoin faces is that it put voting power in the hands of miners instead of users, and then encouraged those miners to coalesce into a relatively small coalition of big-money players.

The energy waste of Bitcoin isn't a function of its decentralized nature, it's evidence that the decentralized nature is being subverted. A truly decentralized Bitcoin wouldn't have giant mining rigs that have to figure out what to do with all the waste heat, it'd be using spare CPU cycles that everyone has and the energy impact would be minimal. The energy waste comes from people using ASICs devoted to nothing but Bitcoin mining. And those same people, according to the article being discussed, are using their outsized impact to help steer the direction Bitcoin is headed in to the detriment of the majority of Bitcoin users. Bitcoin's ethos of "one CPU, one vote" was subverted as soon as people using dedicated Bitcoin mining hardware could have more votes than all the CPUs in the ecosystem put together.

In support of your points, I think it's worth remembering those "life in a bitcoin mine" videos that some journalists groups did

http://media.coindesk.com/2014/09/20140729120509_0210.jpg

The waste heat from these operations are not being reused. The energy most likely came from a coal fired power plant, or from a nuclear reactor.

Also we have no reason to believe that the trend will change away from advances in hardware efficiency just buying more of them - the ratchet mechanism doesn't reward stable capacity.

It's made me a little down on digital stamps, which was an idea I used to like.

Back before spam filters got good, one of the proposed solutions to the spam problem was to require everyone to compute an easy to verify, difficult to produce hash for their email message and bounce everything which didn't come with an attached hash. The idea was that a regular user could spend 1 or 10 seconds hashing each email, but a spammer couldn't spend the time hashing millions of spam messages.

Now I suspect that we would have just ended up with data centers full of highly efficient hashing hardware cranking out signatures for spam messages.

We didn't need bitcoin to know that HashCash and the like wouldn't work to prevent spam. The existence of botnets already means spammers won't be impacted. No need for expensive specialised hardware when other people are paying for the commodity hardware and electricity. And since a regular user on a not-so-powerful mobile device needs to be able to compute the hash in a reasonable time, that puts an upper bound on the processing that's guaranteed to not be an obstacle to a spammer with 3000 botnet slaves.

Yeah, the botnet / cell phone gap was already large enough.

I wonder if that could have been foreseen in the halcyon days of the late 90s, when the image of the l33t hax0r was someone sitting in the middle of a cobbled-together supercomputer in a garage with painted over windows, and not in the middle of an anonymous coffeeshop with a botnet.

> since a regular user on a not-so-powerful mobile device needs to be able to compute the hash in a reasonable time

I would assume this would be implemented at the SMTP level, so when I connect to gmail with my phone, it's their servers that need to generate the hash before they send it through SMTP.

gmail handles a lot of mail, which again puts an upper bound on how difficult the puzzle will be. I don't think this changes anything. And if you make it too hard, "free" (ad-supported) email will no longer be possible because the costs are simply too high for the financial return.

The root problem with Bitcoin isn't the program, it's the money supply equation. Deflationary currencies fail exactly because they are too volatile, which is the anti-thesis of designing a currency to be held as an asset. Imagine currency like the oil in an engine. Someone needs to create a coin that rewards mining by giving them futures contracts on the currency and still deflating it.

Reply on news.ycombinator.com