The Bitcoin Saga


It's economical if you don't invest in more electricity or a better GPU. In other words, it's economical if you run the miner in background on a computer that is running all the time anyway, like a home server. You won't get rich, but you wouldn't get rich running SETI@HOME or a screensaver either.

So large scale is the only way to go, so does that mean there has been a glut of very rich people setting up huge 'mining farms'?

"Expensive" does not mean "limited to rich people". It means "limited to people who find investors". After all, opening a small restaurant can cost even more but it's certainly not limited to very rich people.

I'm just not feeling it, i can do many more things on my PC to make a better return

How many of those things can you do with an unattended PC?


No, unlike you I have invested in neither, but thanks for the disclosure.

But of course, their fees are not reasonable at all. Sending 100 KES incurs a 5% fee. Sending 101 incurs a fee over 25%. Any economist can tell you the problem with that, and unfortunately for you their model is not likely to be successful in other countries.

Yes, how dare sending very small amounts have it's actual costs properly reflected, and not passed off onto other users! And ohnes there are some edge cases, which is entirely solvable. Moreover, it's not? Really? I see...

Kenya (2007), Tanzania (2008), Afghanistan (2008, notable success in paying police salaries to fight corruption), South Africa (2010), parts of India (2011, 2013), Romania (2014), Other EU Countries Soon (TM)

There's also, of course, existing (multiple) mobile payment systems in the Nordics and Baltics...

The only flaw is trusting the wrong person with your credentials.

...I know someone who lost bitcoins because his printer had misaligned heads.


Ha! Did Bitcoin ever accomplish THIS!?

Dogecoin car will be in NASCAR:


Lemme guess: they're waiting for Dogecoins to appreciate enough so they can also afford tires and an engine, right? :)

Actually, it is cool that the Dogecoin Foundation is doing fundraisers rather than, say, just trying to get rich selling `coins to gullible speculators. Not that NASCAR really needs the help, but their other causes look worthy.




Very wow.


so left turn


such nascar


I don't know how new this news is, but Newegg now apparently supports Bitcoin payments.


In other news, the bitcoins seized from Silk Road were auctioned off and bought by Tim Draper, who claims to be responsible for the success of Hotmail and Skype. Well, no, he didn't work for them, but he made them successful because he invented viral marketing. Well, was the "ideator" of viral marketing, whatever that means... He's currently working on a plan to split California into six smaller states: two that would be the richest states in the country and four that would be among the poorest.


A good home.


Ok, so I read about bitcoin mining, and it keeps saying that you are "rewarded" with bitcoins (or I assume, fractions of a bitcoin) for having your computer process bitcoin transactions, which they call "mining". The other way it's put is that "mining" is "doing specialized math problems".

Since this is a crypto currency, can I assume this transaction processing is running encryption/decryption algorithims? Bascially, the transactions take enormous computing resources to calculate/run the algorithms in order to transfer the digital currency. Is this correct?



Here's the short answer, as I understand it. Any errors contained herein are my own.

A crypto currency like Bitcoin has a potential spoofing problem. If I have a bitcoin and I simultaneous tell two people I'm giving them the same bitcoin, i.e. send transfers within milliseconds of each other, they could both accept the same bitcoin and give me something in return. There's no central authority, so they both have to take my word for it that I have the bitcoin and that I own it.

Regular currencies use banking systems to avoid this, and they can reverse transactions if I try and give two people $1 simultaneously when I only have $1 in my account.

Bitcoin handles this problem by introducing a deliberate delay. I broadcast the news that I'm sending a bitcoin to X. Lots of people get that, and start a brute-force decryption of a chain of transactions, one of which is mine, and broadcast when they have a solution. This takes time, typically about 8 minutes right now. X listens for these broadcasts. When X gets enough to feel safe, X accepts the transaction.

Under this system, if I try and send the same bitcoin to both X and Y, I have to tell the world about both. Either X or Y will hear about at least one of these double transactions before they close the deal.

The system depends on a lot of people being willing to do time-consuming calculations. Hence there's a built-in system where if you do enough of these calculations, you get some free bitcoins. These are totally new, and like printed paper money, slightly devalue all existing bitcoins. The intent is that the rate of coin production goes steadily down to avoid inflation.


The vast majority of mining is solving arbitrary problems at this point, only a small percentage is actually running the network. The difficulty of the problems is simply increased to whatever is required to keep the creation of new coins at a constant rate, so when twice as many people are mining, the problems are twice as hard. But this is just arbitrary math. There portion of profits from processing transactions will increase over time as the mining rates decrease.


That seems like such a waste of computing power, in Quaro's scenario. You'd think they would at least take some complex calculation that needs to be worked out, like some quantum physics calculation or protein folding model or something. Lot of energy being used for nothing.


It's pretty much an arms race, where you have to keep buying more and more hashing power just to keep your share of the mining the same, or you'll rapidly fall behind and earn nothing (well, more nothing).

The problem with finding a better, more useful problem to solve is that it has to be something difficult to solve, but easy to verify. Cryptographic hashes work really well for that, but something like protein folding is just as difficult to verify as it was to solve in the first place. Long verification times limit the network's ability to keep up, and if for no other reason than greed, people want their nodes to be spending their time solving, not verifying. Some of the alternative coins are experimenting with different challenges, and Curecoin is trying the protein folding thing, but there doesn't seem to be a lot of activity around it.


It's hard to think of how bitcoin's computational needs fit into any other paradigm, though. Basically bitcoin and its ilk are

f ( transaction data + X ) = A
Where f is an easy function to compute, but f[SUP]-1[/SUP] is really hard to compute
Solve for X (via brute force) to verify the transaction data

How would you fit the payload into scientific computing? And that's without considering the criteria that the function must be easy to verify (f), but difficult to solve (f[SUP]-1[/SUP]). IIRC protein folding is difficult to compute (brute force), and the same brute force steps are calculated by a verifier as well.


It is depressing when you think about it. A lot of people don't really understand it either, that increasing the overall efficiency of mining (via asics for example) doesn't actually improve anything, it's all zero sum. All that matters is the dollar cost of getting 50% of the network power. If everyone moves to asics that are 1000x more efficient the situation is unchanged.

There have been some efforts to make coins that do work with the calculation. Some things like Namecoin to do a DNS like service. Others like CureCoin or GridCoin to do scientific computing. But as mentioned these are pretty hard to pull off in a distributed way.


Some cryptocurrencies try to avoid wasting computer power by using a "proof-of-stake" system instead of a "proof-of-work" system.

From my limited understanding, this means that transactions are validated by coinholders, not CPU cycles. To oversimplify, the first transaction would be validated by "votes" from those who hold coins #1 - 10, the next by those holding coins #11 - 20, etc. The more coins you hold, and the more likely it becomes that you will be involved in validating the next transaction.

The CPU load is minimal, and spoofers are deterred by the difficulty in owning 51% of all coins rather than the difficulty in owning 51% of all computing power. Likewise, the rewards of validation do not accrue to those who own the most computing power, they accrue to those who own the most coins. This is intentionally meant to mimic earning interest in a bank.


That's only transaction processing -- mining is still just arbitrarily difficulty hashing.