Bitcoin ASIC Miner from BFL, finally | Hacker News Reader

Bitcoin ASIC Miner from BFL, finally

codinginmysleep.com

83 pointsenmaku99 comments

The more people get excited about faster ways to hoard bitcoins and eventually make a nice profit (presumably in USD in the end) the dramatically more cynical I become about its future. When the bitcoin discussion was mainly around its use as a currency for things like silkroad, VPNs etc I actually thought it seemed to have a promising future. I really hope I'm wrong, but I have a feeling in 2023 the bitcoin discussion will be relegated to nostalgic jokes about how crazy we used to be.

Indeed, it is pretty crazy. I know someone with an Avalon ASIC who claims Satoshi is a genius. I think he's biased by the fact he has a small device that plugs into the wall and makes him more money than his day job.

Is there any doubt that Satoshi is a genius? Look at what he's accomplished.

> The more people get excited about faster ways to hoard bitcoins

People seem to be oblivious to the fact that a currency must be available in order to be useful. There is a reason that Gresham's Law has achieved "Law" status.

Oddly, the right thing to do for currency acceptance would be to shutdown all of the ASICs.

I think people are more oblivious to the fact that a "deflationary currency" is inherently flawed because it by definition results in hoarding, which is not what you want people to do with your currency. Why would you bother loaning or investing it if you can make more money just sitting on it?

If USD was limited to currency in circulation then it would not be nearly as inflationary as it is.

Once an established lending system based on bitcoins is established it will have the same problems as our current currencies. It wasn't printing money that caused the real estate bubble, it was the way to banking system has the power to expand credit and the systemic risk that expansion of credit causes.

Once there are fractional reserve bitcoin banks it will no longer be a deflationary currency.

Will people want to borrow in defalationary biased currency though? Better to borrow in Bernanke Bucks where there is an unstoppable torrent being emitted monthly.

IMHO the Real estate bubble was from the Fed bailing the prior tech bubble by suppressing interest rates - which encourages borrowing. When loans are paid back, the currency created by taking the loan is extinguished but they have been juicing the system by continually suppressing rates via QE/bond purchases and not letting the real market decide the risk (ie: price of bonds). Now that interest rates have hit the floor - the next bubble to pop is the USD itself.

If the USD pops then BTC may be less deflationary. The bubble did not pop because mortgages were paid back.

Loans are generally repaid in currency, paying back a loan creates even more currency as the additional currency created by the interest becomes an asset against which loans can be issued.

The idea the USD is going to collapse is farcical. It would be a global catastrophe of a scale not yet seen, but also not threatened.

Meanwhile, BTC gains and loses 50 - 80% of its value in a single day. I wonder which is more stable.

Agree BTC is way too unstable but being the the best nag in the glue factory doesn't mean much.

USD collapse would certainly be a far reaching global event but that doesn't mean it will not happen. I'm sure all prior entities with reserve status also thought their position unasailable.

http://www.zerohedge.com/sites/default/files/images/user5/im...

Global support for it is waning. China is doing direct currency swap deals and is no longer supporting USD by stacking treasuries at the rate it was. Russia is dishoarding treasuries too. The current trade deficit is propped up by the Fed buying USG bonds and suppressing interest rates. USD is not long for this world in its current form.

Plenty of currencies exist without being de facto global currencies and survive just fine. The question is what makes the USD likely to collapse as a currency altogether - rather then simply behave more like the normal internal currency of a sovereign nation?

People don't go around predicting the decline and fall of the Australian dollar for example.

The USD is the global reserve - therefore it is needed in vastly greater quantities to lubricate global trade than would otherwise be needed internally. To buy oil, France needs to export to the US in order to obtain USD to purchase off oil off Saudi Arabia. The US gets to print off (and spend) a lot of currency that would have otherwise caused inflation domestically but instead is mopped up by the need for global trade lubrication. Central Banks also absorb USD for reserves - for now. The USGovt is now addicted to its budget deficit that has been enabled by this reserve status but it gives the USA an exhorbitant privelige against other nations - this is starting to unwind - CB's no longer trust the debt of other sovereigns. There is a glut of USD in existence - vulnerable to a change in other nations willingness to hold them. The military is powered by oil and goods and services bought on USD. The USD enables the military, not the other way around.

> Plenty of currencies exist without being de facto global currencies and survive just fine. The question is what makes the USD likely to collapse as a currency altogether

Bingo! You got it! The USD is not going anywhere or collapsing. The global "reserve asset" named dollar however seems to be slowly on its way out.

Fair game right?

Problem is, there's a wave of trillions global dollars that will now seek to "go home". Inflation hasn't happened yet when fresh currency notes are "exported" (exchanged for real goods from the world) immediately. But those USD, whether in bonds or treasuries or bank deposits or hard cash, will seek to come back to their native soil when global savings or trades no longer desire them as they used to. Wait for it! The money has already been "printed", then "kept in custody" all over the globe for some amount of time (circulating at times, finding a good temporary residence for some other times), but still waiting to go back to their permanent home, to circulating with ever-increasing velocity..

> The idea the USD is going to collapse is farcical. It would be a global catastrophe of a scale not yet seen, but also not threatened.

In the short term you are probably right. In the long term you are almost certainly wrong.

Hardly. What long term destabilizing factors would make people abandon the USD - the currency of the current global military hegemon, if nothing else?

rampant inflation.

You mean that 2-4% range with the Fed is legally mandated to keep it in, and has been wildly successful in doing so for the past 40 years?

How are you fitting the 1970s into that statement?

By observing that the USD did not "collapse" despite it?

Plenty of BitCoin people seem adamant the USD is definitely going to collapse, but can't explain why it hasn't already, or where the instability causing it too would come from, or why smaller, less wide-spread currencies don't do the same barring exceptional circumstances (i.e. Zimbabwe).

I more meant the 2-4% range part. If you had said 30 years, fine. But inflation was (well) over 4% for much of the 10 years prior to 1983.

I didnt say there was inflation, its just one of those things that can make people drop the US if it happened.

US debt cannot increase indefinitely without consequence, and in a time of desperate need (such as the 2008 bailouts) it has to print lots of money and hence cause inflation. If the US dollar suffered 10% inflation 2 years in a row, it could originate some drastic changes, as dollar holders might want to get rid of the coin. Figure china selling out its entire dollar reservers in one year, US goes to the tank.

Im not being alarmist, its just a possibility. Comparing Bitcoin to USD makes no sense at this stage in many ways, starting with volume.

Did I hear "global military hegemon"? Try paying your soldiers far from home in a dying currency, go for it! ;D

(Not making a case for "collapse gonna happen", just saying "if, then your military's worth sinks with your currency"...)

> It would be a global catastrophe of a scale not yet seen, but also not threatened

It would have been in the 70s -- which is why USD support has remained despite the Nixon shock for a few more decades. By now, most central banks and governments are increasingly, dare I say "prepared" or slowly preparing for the possibility without it impacting their economies or trade too badly. Nobody wants to see the USD fail or die --- but day-to-day economic / monetary policies on US soil should become less and less of a rest-of-the-world economic worry, concern or impact compared to previously.

Now, the last remaining problem is savings of millions tied up in USD promises worldwide. Still a big problem, and rest assured the USD will continue to get "support" to the proportion that these savings haven't "migrated" over to "something else" yet.

Anyway, one day all these notes the US has been "exporting" globally for many years will come home one way or the other. Hope the new cash supplies won't, ahem, "overwhelm" a struggling domestic economy at that time, shall we?

"One day". It will be a gradual process - it has to be - because anyone stupid enough to try and cash-out mass USD in bulk will not get there money's worth back in it.

But more importantly, is there any real sign that that time is soon? Global financial crisis...and the world's investors are literally paying the US to take their money with the current returns on US treasuries.

But it's hardly the USD "collapsing" - it's a gradual process over time, which crucially doesn't effect the buying power of individuals over meaningful timespans to them (decades, not days - unlike BTC).

And, probably more importantly, whatever replacement reserve currency comes to the fore (and what would it be, exactly?), it won't be something like BTC.

I see your point - the principal is repaid but the interest stays in circulation.

Agree mortgages didn't pop due to being paid back.

Deflation is not desireable in itself, but it is not a crippling disadvantage. Eg Japan has had a very long deflation spiral and it didn't affect the usability of their currency. And the west has been living in a world of negative real interest rates since the 2008 crisis.

As long as we don't have a big bitcoin credit market and we don't want to regulate the bitcoin "printing press" to adjust for market cycles, it won't even be a disadvantage. = As long as it's only a secondary currency to most users.

Why does the method of mining mean that a currency should not be accepted?

Gold serves as a currency yet the methods of mining have changed drastically in the last 1,000 years.

> Gold serves as a currency

Not for the last 80 years it hasn't..

Actually 39 years. The USD was a token for gold until Nixon shock.

Yeah but that I define as "monetary metal for Central Bank settlements" gold, not "day-to-day normal-people transactional currency". Was it still "money" / monetary after 33? Certainly, until '71 (42 years ago, not 39). But "currency" (circulating to facilitate/lubricate individual micro rather than macro exchange&trade)? Hardly ;) Also not "legal tender" -- as in, for courts enforcing contract liabilities or tax payments.

TL;DR: only outside the US, only to Central Banks.

Interesting point. Wouldn't an individual Frenchman see a USD as a token for metal until the shock? And until the shock, national currencies circulated at a fixed exchange with the USD - thereby his own currency was effectively gold backed?

At a national level, yes the Central Banks settled in gold, but that was after the micro transactions bubbled up to be settled from the commercial banks and then eventually cleared at the Central Bank level (Talking on the edges of my understanding here - just trying to see if the reality for an individual level was the currency was gold backed until 71).

> And until the shock, national currencies circulated at a fixed exchange with the USD - thereby his own currency was effectively gold backed?

Yes, "effectively gold-backed" until the first guy came along to openly demand an actual physical (rather than "potential") exchange of USD reserves for real ounces, in size, from the US. Bam, Nixon shock. Essentially "caused" by one "individual Frenchman" (de Gaulle) testing whether the system can in fact, in size and in physical actuality consistently deliver what it promised to the world.

Gold fell out of widespread use as a currency after the mining became industrialized. Nobody wants to play in a rigged game.

Tell me more about how gold fell out of use as currency because mining became industrialized.

http://en.wikipedia.org/wiki/Bretton_Woods_system http://en.wikipedia.org/wiki/Executive_Order_6102

I'm not a gold bug and I see many problems with using gold as a currency, namely that the amount of currency in circulation should depend on how much stuff we can dig out of the ground. (In the case of BTC something about how many times a SHA-256 hash can be calculated.)

The idea that gold fell out of use as currency because of industrialization is either a completely new discovery in economics, or completely wrong. I'm pretty sure it's the latter.

I don't think the Bretton woods system counts. That's between countries, I'm talking about everyday preference for gold by individuals. I suppose I should have said specie instead.

The same number of coins will be mined, no matter how many ASICs are online.

Mining = securing network from double transactions. This is about security of transactions, not about saving. Please stay on the topic.

This is working out exactly as planned - the security of the system against double-spending is predicated on honest nodes controlling the majority of the hashing power in the bitcoin network, so the fact that people are working very hard to add hashing power to honest nodes is desirable for the system as a whole.

I doubted that these ASIC rigs would ship any time soon. Wouldn't it be more profitable for the vendors to simply to delay/cancel orders and run the rigs 24/7 themselves?

Totally agree.

Why would anyone wants to sell the goose who lays golden eggs while gold cost more than the goose?

They'll pile up cash from "preorders", mine bitcoins and when time is right - will dump outdated devices to customers.

are there any alternatives to this? and some thing that's already selling them and not pre-orders?

I see it as a futures contract on the value of the BTC. Selling a mining ASIC locks in profit for the seller, betting that this will exceed the time-value generated from the rig itself. Buying a rig locks in a loss, betting that it will pay for itself. We can't be certain whether the BTC will be at US$1,000 or US$1 a year from now. You could make a case for either side, that's why the contract exists.

Which, by the way, doesn't mean they are betting against BTC. They just elected to take sure profit up front, over less certain profit of potentially greater value long-term.

I don't think so. They probably needed the preorder money to fully develop the device and to be able to manufacture it.

The profitability of ASICs is going to decrease at a rapid rate once they start shipping to consumers. Additionally, in a few months the market is going to be flooded with cheap ASICs.

From BFL's perspective if they don't release first they're dead in the water. They don't know where their competitors are compared to them but they recognize it's a winner take all situation. The natural thing to do would be to take pre-orders to fund the manufacturing and start realizing profits now. Also, if your competitors are offering pre-orders you should be too.

Did they collect the money in advance? If not, it seems to me that the shipping time is almost irrelevant.

Yes, you pay at time of order, and wait a few months for the unit.

few months - year...

This waxing and waning of mining devices is certainly going to become the norm.

* Buy up $10K worth of devices

* Make your money while flooding the market with fast and easy BTC

* Sell off all your hardware after the profit margins have cratered

* Wait for new hardware and rinse and repeat

Personally I will be extremely glad when the first wave hits as the hoarders will start feeling the pinch of inflation.

bitcoin does not work that way. Inflation decreases at a set rate.

Difficulty adjusts week-to-week. Being the only guy with an ASIC is like a license to print money. Being the 10,000th person with an ASIC will be like trying to mine with a netbook today.

Supply will always be constant however how it's divided changes -- people who previously enjoyed lack of competition are now having their feet held close to the fire.

The evolution was CPU -> GPU -> FPGA -> ASIC. Now that we're at the ASIC stage growth is going to slow considerably. There's no longer order of magnitude gains to be made.

>at the ASIC stage growth is going to slow considerably

Yes, buy these are 110nm v1 designes, it looks like they will be able to squeeze something like 10x per watt by moving to better process and iterating on design. And probably additional 10x in cost by ordering more than current quantities that are in the ballpark of the smallest possible order.

Right now entire network runs at speed equal to all hardware in first two Avalon batches (300 and 600 units). When batch 2 and 3 and BFLs hit in the future, everything else is going to be wiped out. But then it will be wiped out again by the next iteration.

This seems like it would be begging for a class-action lawsuit.

That said, I've half-joked that probability of receiving my miner is probably about 0.98^(USDBTC).

How can one litigate against an entity beyond the jurisdiction of the US court system, e.g. say the rig manufacturer was the crook and not in the United States? What are the options there?

The options are to sue them in their foreign country or sue them in the United States. If you sue them in the US, you face jurisdictional issues (do they have sufficient contacts in the US to make a lawsuit here fair?) as well as judgment recognition issues (you would probably have to attach the damage award to a contract they have with someone in the US to collect if the foreign country court system doesn't recognize the US judgment).

Thanks for your answer. I have two questions if you don't mind:

- Are reciprocal judgements possible in a lawsuit such as this hypothetical example?

- What is to stop the defendent from liquidating and reappearing as a new legal entity?

Reciprocal judgments: I'm not perfectly clear what you mean by reciprocal judgment. I think that you'll find this helpful on the matter, though: http://travel.state.gov/law/judicial/judicial_691.html

Liquidation: Corporate governance issues are a matter of the country where the defendant is organized. So if the law there allows them to do it, they can. I'm less familiar with the process of chasing the money. Trying to remember back to law school here: In the US, the debt doesn't just disappear. It would either attach to the old entity (which would have to be paid a fair price for the acquisition of its assets) or the new entity (which acquired its assets and liabilities per the acquisition agreement). So if it has no assets to begin with, it's possible. If it has real assets, it's more difficult to do. Not impossible, just more difficult.

Part of the value of Bitcoin (if any) is that it's decentralized. It should be more profitable to sell the miners out than to hoard them because if they're distributed the cryptographic safety of Bitcoin is greater.

"It should be more profitable to sell the miners out than to hoard them because if they're distributed the cryptographic safety of Bitcoin is greater"

Can you explain this in more detail? I don't understand what decentralization has to do with miners' profit.

Decentralization boosts miners' profits by making bitcoin sounder, and therefore making its price higher.

Miner's profits have to do with the stability of the network and the demand for bitcoins. At high bitcoin prices, their return is better than when the prices are low, other things (thing, mining difficulty) being equal. It's much like gold mining: it has become much more profitable of late because the difference between extraction price ($250 to $400 per ounce in most places) and sale price (roughly $1,500 per ounce) is how much the miner takes home.

Decentralization increases this price because it makes bitcoin sound. If there's no one door to knock down, or one person to subpoena, the cyberpunk currency can robustly survive losses in its network and still have enough computing power trained on it to prevent attacks.

Their competitor Avalon sells the units for "the first 30 days profit of the device". So whatever the current price of Bitcoins and mining difficulty affects the price.

I'm sure the same thing was said during the gold rush days and it was the store owners that sold prospecting equipment that made some very stable money.

Those store owners couldn't use a thousand pickaxes simultaneously.

They could hire people to operate the pickaxes, like how BFL would have to pay for electricity to power their mining hardware.

> I doubted that these ASIC rigs would ship any time soon. Wouldn't it be more profitable for the vendors to simply to delay/cancel orders and run the rigs 24/7 themselves?

Only if they're "long BTC". What makes you think they are? They're just providing shovels & tools to "miners in a gold rush". Historically, such providers have always done phenominally well.

At least they get the rigs at production cost to run them---selling them for good ol' USD may just be a hedge or sideshow, only time can tell.

The difference with shovels is you need to be there and shovel yourself. Here if you have tool, you just plug it in and it does all the work for you.

There is energy expended during the process though. I imagine if this continues 'banks' of the future will be data centers. Also might be a way to make money on idle servers at data centers.

If anything this is an extremely interesting time in currency experimentation. Side effect of this is massive energy cost, another awesome side effect is hardware/supercomputer miniaturization and evolution. Great for really hard problems like verification (as in mining here) and maybe brain research/processing or other really hard data crunching problems.

I'm amazed that there hasn't been a scandal with some Googler or Microsofty or Appler or academic finding a way to use some fraction of their unused cycles for mining.

http://en.wikipedia.org/wiki/Bitcoin#Botnet_mining

http://thenextweb.com/au/2011/06/23/abc-employee-caught-mini...

Doesn't mention anything about the computing capability of those servers though.

From what I understand these devices started shipping right after the BitCoin peaked... Maybe the vendors kept "testing" them for just a bit longer than they expected?

what makes you think they aren't doing just that? they could start shipping them out and make a profit on the machines just as the difficulty rate picks up and there is more profit to be made in selling miners rather than mining.

I work this out to be about 180 Mhash/J, which is 80-100x more energy efficient than any GPU mining.

Also, for anyone wondering: At current difficulty and rate of exchange this will cost you <$0.10 a day to run and generate $40 of Bitcoin.

$40 per day?

Yes?

I estimate it'd produce around 0.25 BTC/day, mining in a pool. That's about $30/day USD at today's price.

As ASICs come online, the difficulty is going to go up, which will reduce the daily generation rate for any given miner. I've been trying to decide whether or not to order a device now based on my profitability prediction for months from now when I might actually get it. I'm guessing by the time I get one I might be able to produce at least 0.05 BTC/day. Assuming that's $6/day, the $274 device will take about six weeks to pay for itself. Not too bad, but I don't know if keeping tabs on it afterwards is worth it for $6/day. It'd only be worth it if I think BTC are going to be worth a lot more in the future... getting back to hoarding coins as an investment rather than using them as currency.

>getting back to hoarding coins as an investment rather than using them as currency

But then, may be it's better just to hoard BTC directly right now by reallocating thousands of dollars required for device? Can't decide for myself.

Yes, sorry.

Also 1/5th the energy efficiency that BFL originally promised when taking orders, and their most vocal spokesperson claimed that there was no way that they'd be significantly more power-hungry than predicted, that anyone who suggested otherwise was trolling, and that the only reason to draw attention to the fiasco when their FPGA mining hardware which was announced with impressive efficiency and blew out the power budget so badly they had to redesign the boards was to shill or troll. Right up until they blew their power budget and had to redesign the ASIC boards.

Additionally, he tried to convince people to buy BFL rather than one of the competing boards by arguing it was so much more power efficient miners using them would be forced to shut down. It's barely more power efficient than ASICs which shipped months ago, and those companies are already looking at newer models.

Oh, for what it's worth that's also less than 10x the power efficiency of the best FPGA boards...

selling shovels and pick-axes to miners was always much more lucrative than being a prospector - plus ca change

Well, there were surely a few prospectors who made out like bandits- but on the whole, yes.

Isn't this more selling the pre-used shovels and pick-axes after you've already been at the mine for months using the investors' equipment?

The heat is of course a byproduct of the higher wattage, but I wonder what on earth happened to make them miss the wattage target by 6-7x?

They probably missed the clock-rate by a factor of 2x, or decided to use half the number of chips and overclock them to save money. It may also have been their backup plan because people are willing to pay 8x as much for power, but they're less willing to get 1/2 the hashrate.

If you run a CPU at the same voltage, but at half the speed, it uses 1/2 the power because CMOS circuits only really use power as the transistors change state. However you're also able to run at about 1/2 the voltage without getting errors. Because V=IR, 1/2 the voltage at the same resistance is 1/2 the amperage, and since P=IV, that's 1/4 the power for normal circuits. Of course CMOS circuits already run at 1/2 the power. In the real world there's also leakage and such, but in a perfect world doubling the clockrate takes 8 times the power.

Transmeta's Crusoe was the first processor to really take advantage of this, by varying the voltage and running slower to save battery life. Before this time (2000), laptops would use a duty cycle to save power. Soon all the major chip manufacturers started copying them.

I ordered the 60gh/s SC Single the night preorderes opened like a year ago. I think there are maybe ~7 orders before me. I hope I get it soon. I paid 220 btc@5.45 usd for it, let's see if its worth it.

I ordered a Jalepeno almost immediately. I chose to pay by wire, and the email i received included "Your order position won't be penalized for any time it takes to send you this payment information. Please send me an email with a copy of the wire confirmation form. Once payment has been received, I'll respond with final confirmation of your order."

I still haven't paid and am going to now. I wonder if they will honor it or not.

"time it takes to send you this payment information. "

I don't see where it says you won't be penalized for the time it takes for you to make a payment. With that said - I'm curious, did people pay in advance of shipping?

oh no! i read that wrong everytime. In my said it said "it takes you to send this payment".

but yes, they have been taking pre-orders for like 9 months and people have been paying.

> 220 btc

Holy moly: $27K.

If he paid with Bitcoin, it was probably handled through BitPay, which might have in turn converted it to USD before paying it out to the company. I'm not sure the company could have foreseen the spike in BTC and held off the USD for that long.

I wonder who reaped the most out of this.

Arrived just after the peak too.

appreciate the thorough burn-in test

I'm sure with enough digging around I could find this out for myself, but could someone familiar with things comment on how the performance of these boxes compares to a standard ATI card miner?

High-end ATI cards will do < 1GH/s, with most midrange cards below 500MH/s. This does 5GH/s, while using an essentially an order of magnitude less power than a PC-based rig will use.

https://en.bitcoin.it/wiki/Mining_hardware_comparison

The more important specification is the MHash / Joule (also found in the link below). The video says this thing does 5 GHash/s @ 30W. That's 100x the efficiency of the best videocards (1000x if you count the power consumption of the rest of the PC).

More discussion on this article at Bitcoin Talk: https://bitcointalk.org/index.php?topic=181849.0;all

That is the current expected bitcoin per gigahash earning rate?

I thought they had been ripped off by their suppliers and they just never got wise to it.

Reply on news.ycombinator.com