Bitcoin – The Magic of Mining

http://www.economist.com/news/business/21638124-minting-digital-currency-has-become-big-ruthlessly-competitive-business-magic

by mgertner • 12 years ago

47 37 12 years ago

economist.com

thefreeman• 12 years ago

https://www.google.com/search?q=Bitcoin+%E2%80%93+The+Magic+...

drivingmenuts• 12 years ago

Thank you.

mrb3• 12 years ago

"last year it performed even worse than the Russian rouble and Ukrainian hryvnia"

A number of journalists have relayed this claim lately, but I find it disingenuous as it avoid talking about the real reason (volatility) and it hides the big picture (Bitcoin is up 3200% over the last two years [1]).

Otherwise, this is a good article about mining. But one comment I would like to make is that people, like this author, sometimes claim Bitcoin's mining consume a lot of electricity. But compared to what? It is fair to compare it to the many data centers owned and operated by credit card processors, banks, and all sorts of financial companies, but only to the portion of these data centers that fulfill a role roughly comparable to Bitcoin: processing transactions, verifying fraud, etc. Let's do some math: North American data center market power consumption is about 12 gigawatt [2]. I don't know what the world data center power consumption is, but let's assume ~3x that, or 30 gigawatt [3]. Let's assume that 1/3 of these data centers belong to financial companies, so 10 gigawatt are consumed by financial data centers. If half of these data centers are used to process transactions and for anti-fraud purposes, that's 5 gigawatt. By comparison, Bitcoin's mining operations worldwide are estimated to consume 150 megawatt [4]. So Bitcoin mining uses 3% of the electricity of what financial companies use for similar purposes.

[1] From $6 in January 2013 to $200 as of today: http://bitcoincharts.com/charts/bitstampUSD#rg730ztgSzm1g10z...

[2] http://www.datacenterdynamics.com/focus/archive/2014/01/15-g...

[3] 30 gigawatt is also approximately in line with the often quoted estimate that data centers use ~1% of the world's electricity (~1% of 2500 gigawatt).

[4] http://www.eetimes.com/document.asp?doc_id=1323522 - the global hashrate has increased by 50% since this article was written (~200 petahash/sec in August 2014, to ~300 petahash/sec today) but at the same time hardware efficiency (hash/sec per watt) has increased, so the ballpark number is correct. In fact the economist article also estimates it at ~135,000 average american homes, or 168 megawatt (1 average home = 1250 watt).

Edit: @the_mitsuhiko: By definition, "volatility" is a price variation over time, regardless if it is consistent or not.

the_mitsuhiko• 12 years ago

but I find it disingenuous as it avoid talking about the real reason (volatility) and it hides the big picture

I don't think a consistent drop over a year can be attributed to volatility.

bdcravens2• 12 years ago

In all these discussions of Bitcoin's price over N months, I don't see anyone consider this: 2 years is 44% of its tradable life (first trades on Mt. Gox were in July 2010). You might as well talk about Apple's performance since 2000, which is 44% of its tradable life. In which case, you see similar returns (almost 3000%)

So if Bitcoin goes down to $30, that's an amazing investment, right? After all, that's a 500% ROI!

cgjaro• 12 years ago

Nobody talks about "percentage of tradable life" because this is a pointless metric. You don't see banks publishing prospectus saying "this investment gained x% over y% of its tradable life". %/year is what matters to all investors.

So if Bitcoin goes down to $30, that's an amazing investment, right? After all, that's a 500% ROI!

Absolutely. Buying at $6 in january 2013, and selling even at $30 would be a 500% ROI and a great investment over 2 years. What is your point?

Animats• 12 years ago

That's a good article on Bitcoin mining.

Last week a big cloud mining operation, CEX.io, shut down. (http://www.coindesk.com/cex-io-halts-cloud-mining-service-du...) The price of Bitcoin has dropped to the point that their operation was unprofitable.

Because the total weekly reward to all miners is fixed, all miners are in direct competition. If mining effort doubles, return on mining investment halves. The big operations are starting to hit that limit. The little guys hit it a long time ago.

danbruc13• 12 years ago

Mining is what will ultimately make Bitcoin fail. You want to give everyone a single vote so that the majority can agree on which transactions are valid and which are not. But because you don't want a central authority that checks ID documents you come up with the idea of mining - make having a vote in the Bitcoin network an expensive thing by requiring to buy hardware and then spend more money on electricity to keep you hardware alive. It's not really fair because if you happen to have a lot of money you can buy a lot of votes.

But the really bad thing is that you just made Bitcoin transactions pretty expensive - I got six Dollars per transaction for electricity only about a month ago. Even worse, processing more transaction will or should not lower these costs. If Bitcoin would really take off processing transactions worth billions every day, you really don't want to protect them by hardware and electricity costs only worth a few millions because the growing transaction volume will also increase the amount of money someone would be willing to spend to mess with your system. At least up to some limit you really want the total mining costs proportional to the transaction volume.

And from there down it goes. Bitcoin transactions are not cheap even if the current low transaction fees suggest it. As the block reward goes down all the costs of running the Bitcoin network will end up as transaction fees. And transaction costs of several dollars are the death of micropayments and Bitcoin adoption in the underdeveloped world, two of the most lauded arguments for Bitcoin. I am not dismissive of the idea of cryptocurrencies but Bitcoin with its proof of work scheme won't be the queen of the prom.

johndevor• 12 years ago

As the block reward goes down all the costs of running the Bitcoin network will end up as transaction fees.

You haven't mentioned off-chain transactions at all. Don't you think Coinbase and others could enable those micropayments you're worried about?

danbruc13• 12 years ago

Off-chain transactions have nothing to do with Bitcoin - they are just changing the balances of their customers in their database. Essentially no difference to you average bank account besides that it only works between two users of the same service. Unless they make agreements between different services making them even more similar to your average bank.

You have to completely trust them because these off-chain transactions are neither visible to the public nor agreed on by the majority of the votes nor secured by any cryptographic or proof of work or whatever measures. All trades at Mt. Gox were off-chain transactions...

mrb3• 12 years ago

Your math is wrong, making your entire premise/argument invalid. A transaction costs $0.60 of electricity as of right now:

150000 kilowatt (power consumption of the global Bitcoin network [1]) ÷ (3600 (sec/hour) * 7 (transaction/sec)) = 6.0 kWh/transaction

The worldwide average price of 1 kWh is $0.10, so a transaction costs $0.60 in electricity to be processed by the network.

Also, the 7 transaction/second limit comes from the current 1MB block size limit which was chosen by agreement. It will be increased in the near future. If it is raised to, say, 10MB this would allow 70 transaction/sec, hence reducing the electricity cost of a transaction down to $0.06.

[1] http://www.eetimes.com/document.asp?doc_id=1323522 - the global hashrate has increased by 50% since this article was written (~200 petahash/sec in August 2014, up to ~300 petahash/sec today) but at the same time hardware efficiency (hash/sec per watt) has increased, so the ballpark number is correct

imaginenore4• 12 years ago

His logic is wrong too. $0.60 is not a transaction cost - the persons making the transaction are not paying that. It a cost to the miner.

sharpneli2• 12 years ago

Where do you get 150 000 kW? Even assuming all the miners are the best I could find (Achilles lab series) [1] that comes to 190000kW based on current hashrate [2].

Your transactions/second is also bit off. Right now based on blockchain.info the highest transaction day had 115787 transactions [3]. And that day the backlog of unconfirmed transactions got as high as almost 6000 unconfirmed transactions [4]. It doesn't matter at all if the network can "theoretically" support 7tps if it chokes at 1.3tps

That brings the actual kWh cost into 40kWh/transaction and up to the same ballback as what GP had.

Based on the arguments on forums right now it looks that increasing the block size is impossible. Because that would hard fork the blockchain.

When new coins are no longer being awarded to miners then people will have to start to pay the transaction fees themselves. Right now the mining is subsidized by what basically amounts to printing money and thus effectively is a small inflationary tax to everyone using bitcoin.

[1] https://en.bitcoin.it/wiki/Mining_hardware_comparison

[2] https://blockchain.info/charts/hash-rate

[3] https://blockchain.info/charts/n-transactions

[4] https://imgur.com/GAklIdZ The most delicious moment was not screencapped, but this is quite close. Even as this was happening the miners happily issued small blocks (500kB and less).

danbruc13• 12 years ago

I just looked it up in a comment [1] I wrote about a month ago - 250 PHash/s hash rate with 1 GHash/J mining efficiency, 100 $/MWh energy costs and 100k transactions per day. Sources for these numbers in the linked comment. But even your hypothetical $0.60 or $0.06 are still way to high for real micropayments if we, for example, think about paying maybe $0.10 for reading a news article as an alternative to subscriptions based or ad-financed online news.

[1] https://news.ycombinator.com/item?id=8782877

kanzure2• 12 years ago

You want to give everyone a single vote so that the majority can agree on which transactions are valid and which are not.

That's not how Bitcoin works, although I admit it is plausible that you were describing how you would like Bitcoin to work. There's no "vote" in Bitcoin. Instead, there's a set of rules for determining whether or not a transaction is valid. Also, the distribution of hash rate does not necessarily have any relationship to the number of unique participants.

danbruc13• 12 years ago

If you are building a system based on consensus you want to give everyone a single vote to vote for what is valid and what is not. But because Bitcoin doesn't want a central authority there is no easy way to ensure this by checking ID documents or something similar. Therefore everyone can create as many fake accounts as they want and let them vote for whatever they like. Bitcoin's solution to prevent everyone from casting as many votes as they like is mining, making it expensive to cast many votes. So the whole mining thing is nothing more than an elaborated fake account prevention mechanism without central authority.

brainy• 12 years ago

guys here in africa ze are starting to accpt bitcoin and i think this will have a positive effect on the coin http://techunzipped.com/bitcoin-conference-africa/

Bitcoin – The Magic of Mining

Related Stories