Topic: ASIC = The end of decentralized mining - page 3. (Read 22780 times)

Turn left by turning right 3 times successively?
Metaphor fails  

ASIC are like Buying a Car that can only turn Right.(SHA-256 ASIC)

What happens if someone finds a pot hole in the road? (SHA-256 Weakness)
Scrap all the Cars?

---

My concerns are switching to Hardware that can Only do SHA-256 or SHA-512.
Also having one provider of that Hardware.
It reduces the ability of the network to adapt to a new encryption algorithm.

Its only a matter of time with Moore's law, SHA-256 is already 10 years old.

True, unless there are several ASIC builders there, but decentralize is not very important, since the coin generation speed is the same, from time point of view, the distribution of BTC has always been decentralized

Lol, well this is a possible scenario, but so unlikely as to not be worth worrying about IMO. But carry on - this is after all the subforum for speculation about mining.

No they weren't. Most miners actually made profits, a lot of them still do. As do BFL and I assume other FPGA vendors.


I dont disagree.
Although if BFL had plans to mine themselves, you would expect them not to say so while they are taking (pre)orders . The comparison with AMD is flawed in that respect; there is no point for AMD to mine bitcoins, its peanuts to them and they make more selling the hardware, if not for bitcoin mining then to gamers.

But ASICs have no other purpose, and more over, electricity cost is pretty much a non issue, so if/when sales dry up, BFL might decide to go for it themselves. Not saying they will, but there is nothing to prevent them from doing so either, other than a financial incentive from miners overpaying for their equipment.

But how does the end of decentralized mining follow from this? AMD was the sole beneficiary of bitcoin mining during the GPU era, yet we had decentralized mining because AMD had no interest in mining themselves, or in selling GPUs to a select few. We have no good reason to think that BFL is interested in mining themselves, or in selling ASICs to a select few. Complete hegemony by a single manufacturer is perfectly consistent with decentralized mining. This point has been made before of course. 

The problem is not about being rich. Its about almost infinite pricing flexibility allowing the first vendor to market to price their chips somewhere near GPU/FPGAs prices per GH upon launch, while being able to drop prices per GH by one or two orders of magnitude before reaching variable costs.

Not only is there this possibility, but because market price of a bitcoin ASIC is directly linked with past sales (more sales -> higher difficulty -> lower ROI for miners) ,  the ASIC vendor is effectively competing with its past customers, forcing them to drop prices even absent any competing asic vendor. ASICs represent an enormous leap in performance/$ of 100 or 1000x where it is today; its this leap that allows the vendor to get a return on his very large investment. But this is a one time deal, another asic might be better, but not 100x better. So once difficulty catches up, and market prices for asics will have dropped to something much closer to variable cost, because the market is inherently limited, it will make no financial sense to invest another x million dollar for a slightly better ASIC, let alone an equally good one.

In a nutshell, if BFL chooses to play their cards right, they can end up being the sole beneficiary of bitcoin mining for the next few years at least. Its future competitors as well as its customers could easily be forced in to losses.

ASIC would only be centralized if bfl mined themselves and didnt release the hardware into the wild.

So basically the debate is "Scince only rich can start producing low cost ASICs' hardware that must mean that they are centralizing the hashing power favor to them and in turn somehow are going to influence how Bitcoin works?"

or am I just not understanding how ASIC is the end of decentralized mining?

there is a high upfront cost to begin production, but once that's done, repeated production is dirt cheap on a chip by chip scale. hard to start up, but once going, easy to keep running.

So I'm not up on all the type of mining technologies but I'm guessing ASICs' will cost a lot?

I wonder who you werent talking about when you said "well...mostly".

LOL 

Wait, GPUs can do more than calculate SHA256 sums?

When cars came out, horses were re-purposed for other activities, because they could still do things cars couldn't. Ride on the beach, ride up a mountain, etc. Nowadays, most people who ride horses do so for fun, not transportation.

Point: People didn't have to shoot the horses just to drive the cars.

People don't have to shoot their GPUs, just because (at some point) they will no longer be profitable for mining. GPUs will always have other purposes (anyone here ever play a game?!).

I still say, does it matter?

ASIC is happening.  No matter what you say in this thread, ASICs will still happen.

I would say we should ask a different question that accurately describes what some of the ASIC detractor's more serious claims.

I would propose the question would be: "How many cars would be on the road and in service if to have Cars (read: ASIC) on the road meant we had to shoot 98% of the horses"  

I say it in this extreme manner because the effect of the ASIC technology is effectively doing that.   Sure, just like with the horses analogy, we still have horses around that are used for transportation.    But all in all, cars or some variant of them is used for ground transportation that doesn't involve walking.


I consider this the credible and serious question being asked of this transition.   This is a point that I believe we are not fully evaluating. I am not dismissive to this question.  


Thoughts?
Dalkore

http://bit.ly/MTChaj

As some members have pointed out, GPUs are all ASIC... What people are really talking about when they say ASIC is "SHA256 Specialized ASIC"

Sorry for my ignorance, but what is ASIC and why is it significant?  In idiot-layman terms, please.  Thanks!

I agree completely SgtSpike!

What about all the mug warmers jjshabadoo  

All my mugs will be warm for next to nothing, and my friends mugs too because they are so inexpensive,  and I can keep my GPUs on at a profit because they are always making more than if I heat by other means?

It's hard to get these arguments, GPUs are hot noisy sensitive monsters that draw down the world's juice and you're trying to say ASICs  are bad idea? It seems simpler to just scale up you're idea of what the difficulty should be. If I spend the same on ASICs as I did GPUs I'll get a ton more hashes but so will every one else. This can only make Bitcoin bigger better and faster. I'll probably still be using less electricity and getting more btcs a day with an easier, more secure, continual and standardized upgrade path in line with Moore.

BFL has to sell its units to recoup its investment and right there all conspiracies die even without further competition from others that are bound to show.

Isn't there a thread that predicted the same thing when GPUs started up?

The Ethernet jack could even just be for initial configuration via a networked computer.  After which the device can operate on the network wirelessly (maybe even on battery power for a short while depending on efficiency, and have an indicator light when it needs to be plugged back into a power source).