Perplexed about bitcoins
March 23, 2013 8:49 AM Subscribe
Can anyone explain to me, as if I were a child, how bitcoins work?
The 30,000 foot view: it's a peer-to-peer network, except instead of sharing files, it shares data about a collection of wallets and how many "coins" each wallet contains. The owner of a wallet (who has the corresponding private key) can ask the network to transfer coins to another wallet. Every transaction is announced publicly to the entire network and recorded in an ever-growing block chain, so that all the nodes can agree on how many coins are in each wallet. But wallet IDs are randomly generated, and anyone can create as many new ones as they want, so there's some degree of anonymity as to who the coins actually belong to.
As a visual aid, here's a live view of the most recent blocks in the chain.
The critical feature is that there's no central record-keeper of who has how many coins, and yet the entire network can securely agree on the current state. The details are technical, but at its heart, it relies on a particular cryptographic problem that takes a lot of CPU time to solve, but is very easy to verify once a solution has been found. Miners compete to solve the current instance of the problem, and whichever one does it first gets to decide on a list of transactions to be included in the log. As a reward, the network allows the winning miner to grant itself a small number of bitcoins out of thin air. The point of all of this is to create a "proof of work" mechanism, whereby in order to take control of the network, a malicious user would have to control more CPU power than the entire rest of the network combined.
posted by teraflop at 9:45 AM on March 23, 2013 [7 favorites]
As a visual aid, here's a live view of the most recent blocks in the chain.
The critical feature is that there's no central record-keeper of who has how many coins, and yet the entire network can securely agree on the current state. The details are technical, but at its heart, it relies on a particular cryptographic problem that takes a lot of CPU time to solve, but is very easy to verify once a solution has been found. Miners compete to solve the current instance of the problem, and whichever one does it first gets to decide on a list of transactions to be included in the log. As a reward, the network allows the winning miner to grant itself a small number of bitcoins out of thin air. The point of all of this is to create a "proof of work" mechanism, whereby in order to take control of the network, a malicious user would have to control more CPU power than the entire rest of the network combined.
posted by teraflop at 9:45 AM on March 23, 2013 [7 favorites]
A penny is just a token of value determined by market forces and law. Authentic due to the difficulty of making and again law. Hard to steal as it's in your pocket.
The bitcoin is an attempt to create a software token with those basic attributes using encryption. It's hard to counterfeit, not so hard to steal, market forces have not been too nice to it, and so far the law has somewhat ignored it.
posted by sammyo at 9:49 AM on March 23, 2013 [2 favorites]
The bitcoin is an attempt to create a software token with those basic attributes using encryption. It's hard to counterfeit, not so hard to steal, market forces have not been too nice to it, and so far the law has somewhat ignored it.
posted by sammyo at 9:49 AM on March 23, 2013 [2 favorites]
If you can clarify what questions you have, it might help. The Steve Gibson discussion is good for explaining the technology, but, really, Gibson and Laporte are not economists, and are not enlightening from that point of view.
Economist Tyler Cowen has some notes on Bitcoin. If you search the site, you will come across further discussion. Here's Paul Krugman talking about it.
From a practical point of view, it depends on what you think "money" or "currency" is. Bitcoin provides a relatively involatile unit of account -- you can't counterfeit it, and you can't inflate it (so the issuer does not gain from seigniorage). On the other hand, it's convertibility, barring some publicized situations, is poor, so, it probably fails as a medium of exchange. This may improve in the future, but people will need to believe in it.
It should be noted that distributed monetary systems (i.e., without a central bank controlling the supply) are not new. Take a look at the idea of free banking.
posted by chengjih at 9:50 AM on March 23, 2013 [4 favorites]
Economist Tyler Cowen has some notes on Bitcoin. If you search the site, you will come across further discussion. Here's Paul Krugman talking about it.
From a practical point of view, it depends on what you think "money" or "currency" is. Bitcoin provides a relatively involatile unit of account -- you can't counterfeit it, and you can't inflate it (so the issuer does not gain from seigniorage). On the other hand, it's convertibility, barring some publicized situations, is poor, so, it probably fails as a medium of exchange. This may improve in the future, but people will need to believe in it.
It should be noted that distributed monetary systems (i.e., without a central bank controlling the supply) are not new. Take a look at the idea of free banking.
posted by chengjih at 9:50 AM on March 23, 2013 [4 favorites]
It helps to remember that money and stuff are not fungible. You always have to find someone to buy your stuff, or take your money.
But if you think of it as trading stuff for stuff, then it gets easier. Bitcoin is just a "stuff" that hopes to be considered as easily tradeable as money.
The alleged benefits of bitcoin are that it was designed to not be a fiat currency. There are only so many valid "bitcoins" out there, and the idea is that as a Bitcoin economy grows, the currency deflates and what was once worth 1 bitcoin will someday be worth .01 bitcoin. Nobody can create new bitcoins, they can only "find" ones that haven't yet been found. It's basically electronic gold. There is only so much gold on earth, and as we exploit the easiest to mine supplies, it gets harder and harder to find more, so the relative value of the gold that is out there goes up.
(Or, what Krugman says. The money supply needs to be able to expand and contract with the demands of the economy. One that doesn't is a hindrance.)
posted by gjc at 11:35 AM on March 23, 2013
But if you think of it as trading stuff for stuff, then it gets easier. Bitcoin is just a "stuff" that hopes to be considered as easily tradeable as money.
The alleged benefits of bitcoin are that it was designed to not be a fiat currency. There are only so many valid "bitcoins" out there, and the idea is that as a Bitcoin economy grows, the currency deflates and what was once worth 1 bitcoin will someday be worth .01 bitcoin. Nobody can create new bitcoins, they can only "find" ones that haven't yet been found. It's basically electronic gold. There is only so much gold on earth, and as we exploit the easiest to mine supplies, it gets harder and harder to find more, so the relative value of the gold that is out there goes up.
(Or, what Krugman says. The money supply needs to be able to expand and contract with the demands of the economy. One that doesn't is a hindrance.)
posted by gjc at 11:35 AM on March 23, 2013
I think there are a few interacting issues here.
1. As with other applications of cryptography, the technology underlying Bitcoin has way outstripped the education most people receive, either in school or in daily life. This makes it difficult to explain stuff "as if [one] were a child." (Meanwhile, even people who are supposed to be experts in implementing crypto algorithms and protocols screw up shockingly often.) But if you're interested in getting into the nuts and bolts, a good place to start would be looking at the following concepts:
The important common thread is that there are mathematical functions that are cheap for a legitimate user to compute a small number of values of, but expensive for an attacker either to reverse given values of ("preimage attack") or to compute a large number of values of ("exhaustive search").
This feature is intrinsic to much of modern crypto, and depends on a conjunction of (a) the architecture of essentially all modern computers with (b) some very deep conjectures in number theory and computer science. There's also a lot of speculation (with varying degrees of well-foundedness) on the near-term practical effect of quantum computers, which are not subject to the constraints of (a) and would essentially enable us to sidestep some of the key implications of (b). IMO it's evident that in the long term quantum computing will change everything, but the time horizon is not very clear right now.
The notion of making something decently easy for a legitimate user but excessively costly for an attacker is key to Bitcoin, and the system is carefully designed to keep that balance in tune as computing technology improves.
2. There are both economic and legal aspects of digital cash systems, but the line between them is different than with "real money." For instance: physical, state-backed currency like a US dollar has the interesting characteristic that the features providing authentication (i.e., anti-counterfeiting measures) are the same features designed to discourage double spending (if you photocopy that $50 and try to spend it, the Secret Service will come after you). In digital cash systems there's the potential that you have to handle the authentication and double spending problems separately, since the problem of forging a "valid" token de novo is different from the problem of double-spending a legitimately obtained token. Bitcoin uses digital signatures verifiable by any party and reliable timestamps to achieve both of these aims, but in a way that doesn't map neatly to physical currency. In general, digital cash systems (as sammyo said above) don't get a lot of support from laws, and a lot of the considerations we might think of as legal have to get shifted to the economic domain. In other words, since attackers can't be effectively prevented from doing bad things or effectively punished once they have, our only hope is to make it so expensive they won't bother to try. This brings us back to proof-of-work problems, and paragraph 1 above.
3. There are a ton of design choices needed to make a functioning digital cash system. (Do you want centralized or not? Anonymous or not? Traceable or not? ...And many more. Again, there are tradeoffs, and these criteria all interact in gross and subtle ways.) Bitcoin is one such system, but many other choices are possible. That's probably good to keep in mind as you read.
I realize that nowhere in here have I actually attempted to describe what Bitcoin is or does, in so many words. But hopefully this is enough context to help you make additional sense of the other answers here, and of what you learn from external resources.
posted by AkzidenzGrotesk at 3:26 PM on March 23, 2013
1. As with other applications of cryptography, the technology underlying Bitcoin has way outstripped the education most people receive, either in school or in daily life. This makes it difficult to explain stuff "as if [one] were a child." (Meanwhile, even people who are supposed to be experts in implementing crypto algorithms and protocols screw up shockingly often.) But if you're interested in getting into the nuts and bolts, a good place to start would be looking at the following concepts:
- Proof-of-work problems
- Hash functions and birthday attacks
- Trapdoor functions
- Merkle puzzles
The important common thread is that there are mathematical functions that are cheap for a legitimate user to compute a small number of values of, but expensive for an attacker either to reverse given values of ("preimage attack") or to compute a large number of values of ("exhaustive search").
This feature is intrinsic to much of modern crypto, and depends on a conjunction of (a) the architecture of essentially all modern computers with (b) some very deep conjectures in number theory and computer science. There's also a lot of speculation (with varying degrees of well-foundedness) on the near-term practical effect of quantum computers, which are not subject to the constraints of (a) and would essentially enable us to sidestep some of the key implications of (b). IMO it's evident that in the long term quantum computing will change everything, but the time horizon is not very clear right now.
The notion of making something decently easy for a legitimate user but excessively costly for an attacker is key to Bitcoin, and the system is carefully designed to keep that balance in tune as computing technology improves.
2. There are both economic and legal aspects of digital cash systems, but the line between them is different than with "real money." For instance: physical, state-backed currency like a US dollar has the interesting characteristic that the features providing authentication (i.e., anti-counterfeiting measures) are the same features designed to discourage double spending (if you photocopy that $50 and try to spend it, the Secret Service will come after you). In digital cash systems there's the potential that you have to handle the authentication and double spending problems separately, since the problem of forging a "valid" token de novo is different from the problem of double-spending a legitimately obtained token. Bitcoin uses digital signatures verifiable by any party and reliable timestamps to achieve both of these aims, but in a way that doesn't map neatly to physical currency. In general, digital cash systems (as sammyo said above) don't get a lot of support from laws, and a lot of the considerations we might think of as legal have to get shifted to the economic domain. In other words, since attackers can't be effectively prevented from doing bad things or effectively punished once they have, our only hope is to make it so expensive they won't bother to try. This brings us back to proof-of-work problems, and paragraph 1 above.
3. There are a ton of design choices needed to make a functioning digital cash system. (Do you want centralized or not? Anonymous or not? Traceable or not? ...And many more. Again, there are tradeoffs, and these criteria all interact in gross and subtle ways.) Bitcoin is one such system, but many other choices are possible. That's probably good to keep in mind as you read.
I realize that nowhere in here have I actually attempted to describe what Bitcoin is or does, in so many words. But hopefully this is enough context to help you make additional sense of the other answers here, and of what you learn from external resources.
posted by AkzidenzGrotesk at 3:26 PM on March 23, 2013
What is Bitcoin? The future digital currency explained.
And a bit more.
Personally I would like this idea to work as anything that takes over from the awful and obnoxious paypal is a winner.
At the moment the best thing to spend them on seems to be drugs so it has that advantage as well, as it cuts out the middle man and has a sort of built in quality control.
posted by adamvasco at 3:45 PM on March 23, 2013 [1 favorite]
And a bit more.
Personally I would like this idea to work as anything that takes over from the awful and obnoxious paypal is a winner.
At the moment the best thing to spend them on seems to be drugs so it has that advantage as well, as it cuts out the middle man and has a sort of built in quality control.
posted by adamvasco at 3:45 PM on March 23, 2013 [1 favorite]
I'm returning to this post, having found it so interesting. I haven't checked if anyone else has posted this link, which I found on Ed Yong's ever excellent: I’ve got your missing links right here (13 April 2013)
Economist Explains How Does Bitcoin Work
posted by Dub at 6:20 PM on April 14, 2013 [2 favorites]
Economist Explains How Does Bitcoin Work
posted by Dub at 6:20 PM on April 14, 2013 [2 favorites]
This thread is closed to new comments.
Transcript & Audio
YouTube video
posted by kidbritish at 9:18 AM on March 23, 2013 [1 favorite]