Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
$100k of FPGAs, Mining Bitcoins at 110 Ghash/sec (bitfury.org)
73 points by mrb on May 22, 2012 | hide | past | favorite | 70 comments


I think the incentives of the bitcoin system to hoard money and the massive bias towards early adopters will be the downfall of bitcoin. All the advantages of anonymity and irreversibility are voided if there is every reason to keep hold of your money and let it increase in value. Several people on HN claim to have hundreds of thousands of bitcoins, making them worth more than 0.1% of the entire economy. Whilst the real world has its (possibly unfair) share of billionaires, the majority of these have come through substantial business innovation. Bitcoin "billionaires" come through luck and early adoption, and in my view it makes it more unfair than our current system.

I think that bitcoin will continue to have some success in areas that wish for privacy, but will not reach the heights that some people predict.


No. The "hoarding" argument has been beaten to death. Evidence shows people are not hoarding the coins: every day, 40 thousand coins change hands on the single largest exchange: http://bitcoincharts.com/markets/mtgoxUSD.html This is six times the number of coins created daily (7 thousands).

In other words, people are not hoarding them, but are trading them very, very frequently.

And this is just measuring MtGox's volume. Other trades (merchant sales, other exchanges, etc) are likely surpassing MtGox's volume in aggregate...


> Evidence shows people are not hoarding the coins: every day, ... This is six times the number of coins created daily (7 thousands).

That's not very convincing. A better comparison evidence would compare the daily exchange to the total inventory.

However, the real question is more granular. If we look at #exchanges/age for each bitcoin, what do we find?

I'd expect to find that some bitcoins have higher turn-over than others even if there's no hoarding, but what fraction have never been traded, or have been traded less than 3 times?


I apologise for re-hashing old arguments, I haven't been following them, and I'd be interested to read rebuttals if you know where they are.

I'm not sure that evidence of trading is evidence of non-hoarding. It's only 0.5% of the total number of bitcoins out there that are traded daily. Is that a lot? I'm not sure, but I don't think we can compare it to the number of coins created daily and draw conclusions from that.

Some people are hoarding them. A lot of them. Or at least they claim to be.


Are there any statistics what those might be used for? Is there a list of sites/services that use Bitcoin with published numbers?


How much of this volume are traders trying to flip bitcoins? I am willing to bet it could be as high as 95%.

As someone who has mined, bought bitcoins, and tried using them at bitcoin stores, there is really nothing special I can buy with them, and can and have dropped 20-50% in a matter of hours.


How many coins were created daily 3 years ago?


The same 7200, it halves every 4 years.

Add: so 7200 coins/day in 2009, 3600 coins/day starting 2013, 1800 coins/day starting 2017 etc.


Early adopters used powerful computers; then GPUs; then GPU farms; now people are using FPGA clusters. They've seen considerable investment in hardware depreciate, and now they're stuck with a massively powerful setup that isn't much use to them. (Perhaps well-funded scientific groups should start buying all these hobbyist GPU farms while they're being sold cheap and put them to use folding proteins or modelling whatever?)

Will anyone go to the expense of an ASIC?

Would a well funded group with access to FPGA technology early on in the chain, when others were using GPUs, have been enough to crash the currency?

And does anyone believe that a government would not be able to destroy bitcoin, trivially, by spending a few days mining?


>Perhaps well-funded scientific groups should start buying all these hobbyist GPU farms while they're being sold cheap and put them to use folding proteins or modelling whatever?

This seems likely. I don't imagine that FPGA clusters depreciate much over periods of a few years. I bet you could make most of the investment back by selling them off to some research firm once the return from the mining dropped. There must be plenty of hi-tech research based startups that need to do serious number crunching on a budget.


The most the government could do is generate all the new coins, this would drive up the price of coins until it became economical compete.

Given the almost limitless divisibility of bitcoins I don't think it's much of a threat. I'd be far more worried about laws being passed that made bitcoins punishable by jailtime.


Actually, at least as of a couple of months ago, if you are generating all new coins, you can prevent all transactions from being validated (by refusing to include them in your blocks) or undermine community faith in bitcoin integrity by double spending. Both capabilities would be pretty devastating.


You need more than 50% of the total bitcoin network computing power to do that kind of attack.

While goverments have the money to do it, do they have the capabilities and incentives?


Of course many governments have the capability.

The main reason a government would want to do this is that control over the coin of the realm is a major source of state power. If a government ever concluded that bitcoin posed a credible threat to that power, that would be a strong incentive to undermine it.


Capability =/= Ability


What?

You think NSA or GCHQ[1] don't have the knowledge to implement an efficient bitcoin machine? You think they don't have the money? You think they don't already have significant computing power?

Or are you suggesting they'd be hampered by legal oversight issues?

[1] GCHQ probably have the most powerful computing facility in Europe. That puts them amongst the most powerful computing centres world wide.


Sorry, not sure what your point is. Could you say a bit more about it?


Just because the government has the resources to throw at a problem it does not mean that they will do so effectively or efficiently.


I don't disagree with you, but there are a significant number of hereditary billionaires (in the real world), which inherit money through luck and early adoption (of business or banking)


> Bitcoin "billionaires" come through luck and early adoption, and in my view it makes it more unfair than our current system.

I don't see how the hoarding/early adopter thing would destroy bitcoin. Bitcoin isn't fair, but lack of fairness doesn't necessarily make the project unsuccessful.


If people don't think bitcoin is fair they won't adopt it, and may even fight against it.


>> fair

I was going to post exactly that. People are funny. If they perceive something is biased against them, they are much less likely to adopt its' use or even encourage its' existence. In fact, I bet there is a fair bit of quiet satisfaction around here every time another "$xxxx bitcoin stolen" story gets posted simply because most of us feel as tho the ship has sailed and we missed it. FWIW, I don't really care one way or another - if this currency exists in ten years, it'll just be another opportunity I missed early on. I'm at peace with that, but I am also not going to buy a bitcoin mining rig. :P


I don't know. Money isn't fair. Some people have it, some not.

And usually people don't purchase things because they are "fair". People purchase luxury cars, big houses, drugs etc. for purely selfish reasons, to enjoy life. If we think that currency is just another selection in the market, people will use the currency that will benefit them most. Fairness has nothing to do with it.


The point is that they have no choice but to use fiat currency.

That is not the case with bitcoin so far.


That was my opinion in the beginning, but I'm not convinced anymore. First of all, the bitcoin billionaires have a strong incentive to spend, because if they don't, the currency will crash and they will lose their money.

Second, if you were to create "inflatacoin", with a small but real inflation rate, who would adopt it? It would probably be better economically speaking, but people prefer putting their money in things that stay stable or grow over time.


If people are spending their money in fear of a crash then hoarding is not bitcoin's main problem.

I don't know the solution to bitcoin's deflation, but IMO it's still a problem.


There is a huge psychological bias for this argument. I believe it is 'sunk cost bias' or something like that.

It goes like this. Say you are a miner. You mine. At the moment you get your bitcoins, you've realized value in the sum of ( $Bgenerated - $Bcost (your hardware, etc)).

From that point on 'hoarding' is no different than going to an exchange and buying $Bgenerated in bitcoins with dollars.

So your argument for 'hoarders' having a tremendous unfair advantage is seriously flawed. They have no greater advantage than any speculator going and making a bet in bitcoin long term value.

The real value generated is the one they capture at mine time, and there is a super efficient market for that with very slim profits.

http://tpbitcalc.appspot.com/


" the majority of these have come through substantial business innovation." [citation needed]


Discussion thread on bitcointalk.org: https://bitcointalk.org/index.php?topic=82941.0

I don't know what the skeptics think, but to me, the fact that at least 2 wealthy individuals or entities have invested $100k in 360-FPGA farms should send a signal that at least some consider Bitcoin sufficiently reliable, trusted, and stable to warrant this kind of investment. They put their money where their mouth is.

I will soon write a blog post documenting the numerous other FPGA board vendors that started designing custom boards targetting the Bitcoin mining market.

Disclosure: I am myself a long-time miner.


Isn't it a lot like historical "gold rushes"?

These guys are selling shovels to the miners.

"Step right up! Get your ~10k/month bitcoin mining gear right here! Only $90,000 and _guaranteed_ to make you rich!"


> "Step right up! Get your ~10k/month bitcoin mining gear right here! Only $90,000 and _guaranteed_ to make you rich!"

I don't see any false advertising here. I think the people who invest to these things know the risks. After all, bitcoin is pretty open project. You can easily calculate how much you will be making money with this with current difficulty/exchange rate. Of course both of them will probably fluctuate heavily in the coming years.


Sure - I wasn't intending to imply any "false advertising" (and apologies if I did). I just saw a real parallel with local (Australian) school taught history of the gold rushes in the late 1800s - a few people got rich mining gold, but a lot _more_ people got rich selling shovels and camping gear to all the people who _didn't_ get rich looking for gold.

Another notable modern example is Apple and the App Store - sure Rovio "struck it rich" with Angry Birds, which Apple "only" skimmed 30%off the top, but how much _more_ money do you suppose Apple has made as a result of Rovio's success spurring on many many imitators? (all of whom pay Apple $100/yr + 30% of whatever success they manage)


Tell that to all the people who bought expensive video cards only to see the price of bitcoin fall until their recoup period's measured in years.


The rate of BTC generation falls off in a very predictable way. The only gamble is USD/BTC exchange rate.


USD/BTC is not the only gamble. BTC Generation system-wide is predictable, but you can't predict the size of the mining industry. So your 1GH/day contribution will be worth a smaller and smaller cut as more people like you are mining. These mining rigs are in competition with one another.


I think it's telling that the manufacturer doesn't accept payment in BitCoins.


The more logical conclusion is that they simply see money to be made.

I don't believe they are sending ANY signal that Bitcoin is reliable, trusted or stable. In fact I personally see this as a clear sign that Bitcoin is fundamentally unfair and easily gamed.


Wait, what? it's a sign that it's fundamentally unfair and easily gamed that the market is responding to the incentives implicit in the design? Miners don't just suck away all the value in the system, they are the value in the system; no mining, no blockchain, no blockchain, no transactions.

More mining, more secure blockchain, etc.


How is it easily gamed? Production of bitcoins is related to processing power, if you can purchase the processing power you can increase your capacity to mine.


> I personally see this as a clear sign that Bitcoin is fundamentally unfair and easily gamed.

Curious that you used the word "unfair". It is quite likely that the fairness of bitcoin, is not related to it's future viability.

Disparity is alive and well in the real world, and seems to be doing just fine.


They are taking a huge risk with this investment. With the mining reward set to automatically halve in 2013, and the currency still in its infancy, its a bold move. And despite dropping 6 figures on the hardware it barely a percent of the total hash rate.


They are only risking the amount that the hardware depreciates while they own it, which is apparently quite a lot less than the purchase price for this hardware.


Who said the creation of wealth should be or is fair?

(not that I think it shouldn't be)


Mining is the problem - these people at Newmont/BHP/Barrick are trying to destroy the economy by digging gold out of the ground.


I'm not sure what 110 Ghash/sec means. Can someone give me an idea of: at current Bitcoin <-> currency conversion rates, what return per hour should you expect from this?


Roughly this farm makes $10k per month.

110e9 (hash/sec) * 3600 (sec/hour) * 730 (hour/month) / (2^32 * 1733208 (current difficulty factor)) * 50 (coins/block) * 5.1 (current exchange rate: usd/coin) = $9903 per month

Minus about $730/month in electricity for the 10kW it consumes, assuming an average worldwide rate of $0.10/kWh.


I think I'm right to say this will drop off as the potential number of Bitcoin dwindles (although I don't know the mechanics)... I wonder how long until this thing pays for itself, assuming Bitcoin value stays roughly constant


So, 9k per month.


Just to give you an idea of how does it compare to different solutions: https://en.bitcoin.it/wiki/Mining_Hardware_Comparison


Anyone know enough about Amazon and Bitcoin to know how many bitcoins you'd get for a $10,000 spend on Amazon EC2 GPU clusters?



... unless you're paying for them using stolen credit cards.


or just use the stolen cards to buy goods from amazon?


Handling physical goods is an easy way to get caught. Bitcoins are a very effective way for criminals to cover their tracks.


They would still have to buy the BTC using the credit card, although I agree it is not as easy to trace as physical goods, the eventual BTC for physical transaction could be traced all the way back to the credit card payment for BTC.

So could be a case of security by obscurity here?


They wouldn't be buying the bitcoins, they would be mining the bitcoins -- and newly mined bitcoins don't have any history showing who paid for the GPUs which found them.


I don't see how paying for the GPU instances using a credit card is any more anonymous and less traceable than paying for physical goods?

Sure mine the BTC and it won't be possible to trace back to the machine.


The margin on EC2 hardware is very high compared to a do-it-yourself solution. The GPU machines are also NVidia based which is nowhere near as efficient as the ATI ones for this particular application.


Using the excess heat for household hot water and space heating is a great idea.


Not really "excess". You put 10kw of electricity in, you get 10kw of heat out.


Excess heat in the sense that it is unwanted heat.


Perhaps you should read up on the laws of thermodynamics. You do not get "10kw" of heat out.


He is more than 99.9% accurate, so perhaps you need to go read the laws of thermodynamics. Heat is energy transferred from one system to another by thermal interaction.


He is mostly correct, since the only work being done in a computer system is the motion of the fans and hard disks.


A great way to sell showels in a gold rush.


I started a small bitcoin consultancy as well. I thought bitcoin was going to be the next gold rush and replace currency all together, so I started building a super efficient OS. I have since moved on, but I still believe in bitcoin.

If anyone is interested in the project, you can check it out here: http://projects.tonigemayel.com/bitcoin-os/


I wonder how long until we see worms/botnets that use the infected PCs to mine bitcoins.


They already do: "I operate a ~10k botnet using a ZeuS software I modified myself, including IRC, DDoS and bitcoin mining (13GH/s - 20GH/s atm)." - reddit IAmA http://www.reddit.com/r/IAmA/comments/sq7cy/iama_a_malware_c... http://news.ycombinator.com/item?id=3960034



already happening. not for a short time.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: