Bitcoin Forum

There are tons of threads out there about BFL's ASIC, how it's good, how it's bad.  I doubt I've read them all, but I've read a good number, and I don't think I've yet seen anyone say this. 

Based on what I've seen, I was thinking we have a doomsday scenario on our hands for bitcoin if the following comes to pass:

1 - BFL delivers on their promise of ASIC throughput, at the cost they say.  Doesn't matter really if they are 3 months late, 6 months late, or even a year late.
2 - Network hashing power rapidly increases immensely, to the point that ASICs are 99% of the hashing power on the network.  Who's going to run a measly 5g h/s setup when there are multiple terra hash rigs out there? 
3 - A critical flaw is found in the encryption algorithm of bitcoin.  Either a flaw, or computing power makes the existing algorithm easily breakable.  Some people are saying this is more of an WHEN not an IF.

At that point, we have two choices, both of which spell the end of bitcoin:

a - Leave the algorithm alone, causing it to be hacked and unreliable, and bitcoin dies.
b - Change the algorithm and break ASICs, which are 99% of the hashing power of the network.  Difficulty at this point is astronomically high for the remaining 1% of the die hard FPGAs and GPU users.  Yes, difficulty adjusts periodically every so many blocks.  But what if it takes weeks to solve a block with the remaining 1%?  That's essentially the end of bitcoin.

Yes, no, maybe so?

M

If you are going to change the algorithm, you can also reduce the difficulty at the same time. Nothing forces the devs to keep the same difficulty if devs are making a change as drastic as changing the algorithm.

It will definitely squeeze out the GPU and CPU miners, and put the majority of the bashing power in the hands of people who can afford the specialty mining hardware.

It is unusual for a hash function to go from "secure" to "utterly broken" in one exploit. Usually, what happens is that it gradually weakens.

So let's say there's an exploit which makes SHA-256 only 1% as hard to brute-force as it used to be. At that point, it's a long way from being broken - there's 256 bits there, after all, and finding a collision is still a 2^121 hard problem (instead of 2^128). But... FPGAs and GPUs can take advantage of that exploit, and ASICs cannot. Suddenly, BFL's devices have ceased to be worth the amount people are paying for them, and anyone with a good graphics card can mine again. And you don't need to change the algorithm, because 2^121 is still plenty hard to serve as proof of work.

Now, if 2SHA256 was broken in half in a single leap of cryptographic genius, then I think we have a problem. Not because the network becomes impossible to secure; the change in algorithm can be accompanied by a reset of difficulty, so the sudden contraction of hashing power isn't a big deal. But from the time the exploit is found, to the time the hash function changes, everything is uncertain  and double-spend attacks can run roughshod over the world. Of course, I think (and hope!) this is an unlikely scenario, since, as I've said before, usually crypto functions aren't broken all at once.

My impression is that this is the point of the BFL "Jalapeno" board - to have ASIC mining hardware within reach of consumer budgets.

Not that $150 is cheap, you understand! Just within reach.

how about change the algorithm and break GPU and FPGA farms?

Being able to change the difficulty is the piece I missed.

That's good, no doomsday.  And here I was working out the math and saw it coming awful close to December 21, 2012. :)

M

The difference there, from my understanding, is those can be programattically changed because they are open and driven by software.  ASIC, however, is stamped in the hardware, and can not be changed. 

M

First of all, the reason Bitcoin has been adopted so rapidly is because it used hardware most people already had access to, so just about anyone could mine themselves some coins, and those who could afford to invest money in something so volatile could simply buy more GPUs with the confidence that they would at least have some decent resale value if bitcoin flopped.

Once you require specialized hardware to mine bitcoins, you alienate a very large swath of your mining user base. Which isn't the end of the world, but it makes acquiring bitcoins that much more expensive for the curious public.

Then once you consider that hashing power is proportional to how much money you invest, and that there are some very big investors out there looking to buy ASIC devices, the bitcoins you can mine with a $150 device will be a very tiny fraction of the allotted block reward.

Remember, since the rate of coin generation is constant, buying more hashing power only increases your share of the reward, not the total amount of the reward. So if everyone buys more hardware, then you're right back where you started, just with less money in your pocket.


EDIT: As a great man once said, "you never want to be first", and in the case of decentralized crypto-currency, it's Bitcoin.
What people need is a CPU only currency. It's not the best solution, but in terms of evenly redistributing a currency, it may be the best option, and I'd love to hear a better one.

I don't know about that.  I'd love to have been one of the early adopters and made (and saved) thousands of coins when they were easy to make and worth a few cents a piece.


When the last bitcoin is mined, might it not come to that?  Why throw terra hashes of power at processing blocks of transactions when megahashes will do just fine?  Right now it's greed that's driving the generation of coins.  That's driving difficulty up as people are putting more and more money into crazy schemes from companies with horrible reputations (no names) to buy dirt cheap ASICs that might turn out to be total scams. 

When there are no more coins to generate, then what?

M

New algorithms will not affect the FPGA or GPU.  If the algorithm changes only ASIC will be deemed useless.

The quote was referring to newly invented solutions, not the users. They used tivo, altavista, myspace, and xerox as examples.

I wish there was more competition for ASiC products.  It's been a pyramid, with CPU as the base, then $200-$400 AMD video cards.  A handful (like 5?) companies then make FPGAs, now BFL is the apex predator at the very top of the pyramid.  So, as the apex predator, they can make you wait as long as they want, pay as much as they want, not respond to emails/ phone calls, etc. 

I only think ASIC will spell doomsday if this monopoly lasts after BFL rolls out the SC devices.  People in all other areas of the pyramid will simply get frustrated and leave Bitcoin, IMO.  You'll have only early adopters with plenty of disposable income citing over transaction fees.

Full disclosure: I own five singles and have pre ordered one SC single.

development of bitcoin specialized hardware, ASICs  = is a good sign that bitcoin moves in a right direction and is actually growing and expanding. a year ago it would have been considered a pipe dream what someone invests in producing specialized bitcoin chips and making them more or less affordable so soon, although obviously it is inevitable and only further proves that bitcoin technology is solid.
bitcoin ASICs mean that this digital currency experiment, this project can and will be taken seriously, it's growing out from kiddie, geeky hobby thing into a next level of mass adoption

if you want to cut off ASICs (which will not happen btw else the project doomed to fail) - you might as well dismiss all human and industrial progress so far and go back living in caves

How does cutting off asics doom bitcoin as a project to fail? Everyone who is mining has a vote in whether or not to allow asics to join the network. Speaking from my perspective as a gpu miner with cheap power, I don't want asics at the moment, especially only coming from one company. If everyone doesn't accept them then they won't happen.

cutting off asics means bitcoin will never be able to grow out of "geeky currency". additionally there is no reason to change sha256 algo as it's backbone of bitcoin network security until something drastic happens with sha256 or network security.

another point,
GPU - highly inefficient in terms of power consumption
FPGA -transitional technology, it has never meant to be final solution
ASICs - is the only way to move forward, and even with ASICs, chip's effitiency will always be improving and there will be times when older generation chips may become obsolete and inefficient in future. but in my view only way to progress is building highly specialized and efficient tools - that's why i say Bitcoin project is moving in the right direction. and we are lucky that at least one company stepped up so far in actually developing ASICs

currently bitcoin mining is in expanding and transitional mode,  everyone investing into mining should realize that margins will only get tighter and at some point your new rigs will become obsolete.

I agree with you.  I just wish BFL had a better reputation.  No way I'm risking that much BTC for something that may or may not deliver.

M

The CPU miners are already squeezed out by the GPU miners, which puts the majority of the hashing power in the hands of gamers and people who can afford mining hardware which they wouldn't otherwise buy.

The Jalapeno is far cheaper than most viable GPUs and serves as a way for the average Bitcoiner to join in mining and to secure the network.

So if bitcoin is currently expanding, and soon everyone's investments will no longer yield profits worth investing in (due too being obsolete), then won't bitcoin inevitably decline? Currently, there are folks investing large amounts of capital because profits can easily be made if you category plan your mining operation. But if profits cannot be made, very few people will invest and there will be large exodus of miners leaving the scene en masse.

Edit: Or, are you saying that by the time any technology for mining becomes obsolete (and this is inevitable), then there will be new profitable technology available? Because in that case, forget my first argument.

I'm not sure why you feel like making the entry cost to mining lower is going to drive less people to enter mining.

If current GPU and FPGA rigs/farms are shut off, difficulty will decrease and the profitability of mining will increase.

Low entry cost does not matter. In fact, that only means the network hash rate will grow even faster. My point is that people invest their money to make MORE imoney. If there is no money to be made from investing in bitcoin mining hardware (no matter how cheap), then people simply won't invest.

And that is true about the gpu and fpga miners... at least until everyone grabs a low power ASIC alternative. At that point the network hash rate will be adjusted to the increase in production except now there are more independent miners with proportionally less mining power because the entry cost is so low and every greedy bastard on the face of the internet wants a piece of the bitcoin pie. Since everyone has such a small proportion now, profitability will drop and no new investors join.

I think that ASIC is the bes tthing for bitcoin right now. For one thing, it defends again botnets.

But a more important advantage to me is that ASIC companies are now invested in BitCoin. We need more SIC companies, enough to make a voice. This voice is important because a major threat to bitcoin is the government passing legislation against it because of the blackmarket uses. The majority of volume in bitcoin is in USD, and if the american government passes a law against mining or trading in bitcoin, it could seriously cause issues for the network. Basically, it would need to be shifted to the darknet and exchanges would need to be drop cash in the mail. The inconvenience would cause the price to crumble. However, with devoted hardware companies, they can lobby to keep it legal. ATI will never go to bat for bitcoin, but BFL will because their business depends on it.

One other scenario to consider is viable quantum computing coming into being. If a 256qbit computer were made, then theoretically, it could churn out a solution every second regardless of the difficulty. In this case, most algorithms that run on GPUs and CPUs won't work.

There's no sub-exponential quantum algorithm for reversing SHA-256. Sure, Grover's algorithm reduces the work to the square root of what it was, but it's still exponential, and quantum computers are unlikely to "catch up" to Turing machines in that respect for a long time, even if a 256qubit computer that could execute Grover's algorithm on 2SHA256 was made.

In short, quantum computers will at best be another competing technology, and one with an uphill battle to achieve parity with ASICs.

I stopped reading here.

We have nothing to worry about guys..

Bitcoin miners need people to hold value in the currency to net a profit. The transaction fees and 50 BTC rewards mean nothing if no one is there to trade them and believe they hold a value.

It's not about whether or not the miners will be able to make more money than their peers, it's whether or not mining will always be profitable or break-even in the long term. Which it should, for a VERY long time.

Your naiveté is amusing. You think that $150 jalapeno is going to generate any significant portion of BTC? If BFL is able to meet the demand and specifications there will be dozens of ASIC based Mini Rigs and hundreds of ASIC based singles. I agree that it is a big if, but, if they do then all of us with medium-large GPU farms are going to convert and the guys with a dinky $150 jalapeno will be left fighting over fractions of a BTC/day.

Better hope BTC spikes in price. In the end this sort of advancement is just going to consolidate a great portion of hashing power into early adopter's hands. I personally am playing the waiting game but I am eagerly watching for the first sign that it is time to dump my $500/month powerbill.

I think the cutoff for "early adopters" will be the first block split. To think, half of ALL bitcoins to ever be mined will have been mined by the end of this year. And I doubt BFL will deliver any ASIC devices before new years.

ASIC will mark the commercialization of Bitcoin, as it is clearly growing, and investors are growing more confident and more bold. It will also be the end of of the "casual miner", but now that half the bitcoins will have been mined, they've had their time, so to speak, and can still buy the coins if they want to take part.

If anyone want's to mine casually, it will be time to move to another crypto-curreny.

Bitcoin will survive past December 2012 if the majority of it's users are interested in it as a medium of exchange, not source of profit. If majority are miners, wanting to cover their costs ASAP, selling btc, then it's doomed to collapse. Find ways to promote Bitcoin in your area, and if everyone does that, we should be safe.

Here in lies the problem! 95% of miners don't mine because they genuinely want bitcoin to succeed and the collapse of fiat currency to follow... They mine because anyone with capital and the knowledge to run simple programs and manage hardware can more than double their investment by sitting on their ass!

When mining no longer provides reasonable ROI and profitably, we'll see a huge decrease in mining interest.

I won't care about btc profitability once I cover my costs. I want btc to survive and be a medium of exchange, not sourse for profit. I hope get the farm running on renewable energy, so when the profit runs out, the network will still be secure and fast at zero cost :))

The community honestly needs more people like you. Hell, even I'm just in it for the profit. I would love to do to help the network, but I just can't be throwing cash for the betterment of a crypto-currency.  :-\

I'm really worried.

I don't know if my router will be able to cope with running 240GH/s worth of SC singles.

I'd be more worried about being able to recoup the cost in a reasonable amount of time, if ever.

M