You forget the cost of the ATM, operating cost, site lease, armored car pickups, maintenance, vandalism, theft, counterfeits, etc.

I don't see ATM as a viable business model unless an area has a very high concentration of bitcoin users.

Can we have honestly from one builder ONCE?  You guys were doing so good up to this point.

That is pure and utter crap and you know it to.

The host in this case is a OpenWRT board pulling what 10W? Ok so 620W vs 610.  Claiming 400W on your site is disingenous.  Sure you have DC to DC losses so does anyone but that is part of the power load.  AMD doesn't quote the power consumption of the raw chip when selling an entire graphics cards.  They quote the card power usage which includes GPU, RAM, VRMS, fans, and any associated control circuitry.   

Your system doesn't use 400W.  Don't advertise that it uses 400W.  It uses ~600W.  Claiming anything less and then saying "oh it needs a host" when you know the host board is pulling a tiny fraction of that is just plain dishonest.  Period.   620W vs 400W is a 55% increase.  That is significant.  Rise above your competitors and start acting like a real company.  You know a real one which wouldn't pull a stunt like this because they would be facing a class action lawsuit for blatant misreprenstation of a material fact.

Maybe for batch #1 which can take advantage of a lack of competition, but batch #2, batch #3, upgrade modules?  Power matters in the long run.  

Purchase price: $1500
620W for 1 year @ $0.10 per kWh = $543
620W for 2 year @ $0.10 per kWh = $1357

620W for 1 year @ $0.10 per kWh = $1086
620W for 2 year @ $0.25 per kWh = $2715

So depending on the economical lifespan of the miner and the power costs we are talking from 26% to 65% of "total cost of ownership" being electrical costs.  A 55% increase over quoted power consumption is significant.

Why is ASIC going to make mining market harder.  To have any meaningful amount of hashing power currently require a multi GPU rig or honestly multiple multi-GPU rigs.  You are talking a $2K to $10K investment.  To have any meaningful amount of hashing power post ASIC you are talking one or probably multiple fully upgraded Avalons so you are talking $3K+.   I don't see much difference.

The amount of hashpower generated by casual miners who earn fractional bitcoin pennies a day with a single NVidia gaming card are a rounding error for network security.  Maybe at one time that kind of casual decentralization existed but it hasn't been that way for a long time.  Just because someone can mine with just about any single low end GPU and contribute 1/100,000 of 1% of global hashing power doesn't mean it provides anything more than feel good security.

Now if the cheapest Avalon unit was 500 GH/s and costed $10K you might have a point but I don't see this massive centralization of hashing power.

We have some clients that do this.  Same large valued purchase once a week, every week, for week after week after week.  Personally if I was looking to acquire say $250,000 in BTC that is how I would do it.  $5K a week for a year.   Sure trying to time the market can be thrilling (buying $250K when BTC drops under $3) but it can also be painful (buying $250K worth of BTC at the peak).   Looking back there is no year long period where buying fixed dollar amounts would result in a net loss.  

Another thing to consider miners who mine and hold are essentially dollar cost averaging even if they don't realize it.  They are trading a relatively fixed sum of dollars (in electrical costs) each week for a variable sum of BTC.

Well you combined a bunch of things as if they are a single entity.

C# is an open international standard.
http://www.ecma-international.org/publications/standards/Ecma-334.htm

Mono is an open implementation of .NET.
http://www.mono-project.com/Main_Page

MS SQL Server is closed source however to answer the question "why"?  Rapid application development.  Nice tools for rapidly constructing entity framework when using Visual Studio and MS SQL Server.  Of course it is somewhat of a strawman argument as it isn't required.  .Net will work fine with just about any datstore SQL or otherwise.

Hashrate is equally irrelevant.  What matters is the cost to the attacker.  Using GPU while attacker uses ASICs only ensures that the cost is much lower for the attacker.  Based on Avalon (assume 100% markup) the cost to attack the network using ASICs is ~$12,000 per TH/s.   Would you prefer it cost the attacker $300,000 (current 25TH/s defended w/ GPUs) or cost the attacker $30,000,000 (2.5 PH/s defending w/ ASICs).  

The attacker is ALWAYS going to use the most efficient method.  There is no gentleman's agreement to only attack a "GPU chain" with GPUs and not cheat with superior weaponry.

There was never a chance that a government or other entity large enough to attack Bitcoin was going to buy hundreds of thousands of GPUs from AMD and build warehouses of GPU farms.  That was never going to happen.  If they wanted to attack Bitcoin they would use the absolute most cost effective way to do so.  

If defenders are using a less efficient method then they are simply limiting how much computing power (and thus how expensive) it will be for the attacker.


The first one option is pathetically easy to attack by collusion and the second one is pathetically easy to attack by brute force.  Neither is truly secure.  How about 2,500 TH/s (or 25,000 TH/s or 100,000 TH/s) divided between a couple hundred miners?

Worst case scenario I would imagine transplanting them to a 4U 19" rackmount case wouldn't be that hard.  Maybe in the future they can improve the module pricing and allow one to buy the controller separately (or even better open source the controller).   Just grab the case & PSU you want and buy up to 6 modules from Avalon.

Gabi didn't litterally mean everyone.   He means that ASICs will replace GPU.

"Bad Guys" could ALWAYS have used ASICs.  Good Guys using them didn't suddenly unlock that option.  I mean your logic is like .... we know guns exist but if we use them we could be outgunned.  So instead we will use only rocks and thus our enemies will need to use rocks and we can beat them in a rock war.  No the reality is one guy with an automatic weapon would defeat your entire army of soldiers with rocks.

Moving to ASICs takes that "cheat card" away.  Sure an attacker could try to outspend the network (they always could) BUT now they can't cheat and build a small number of ASICs to defeat a huge number of GPUs (or an even huger number of CPUs had Bitcoin never moved to CPUs).

Maybe some numbers would help.  Lets say bad guys can use ASICs to build a 51% cluster at a cost of $12,000 per TH/s.  Which is more secure a network consisting of 25TH/s or one consisting or 25,000 TH/s?


TL/DR:
Bad guys will ALWAYS use the most cost effective attack vector.  You want the good guys using the most efficient hardware possible (someday hopefully even MORE efficient than Avalon) to ensure the bad guys have to pay the most possible.

There is no double edged sword.  ANYTHING presents a risk is hashing power is concentrated.  There is nothing unique about ASICs.  Litecoin for example is very vulnerable to Botnets.  The limitations of LiteCoin make is highly likely that hashing power will never be decentralized enough to overcome that risk.  LTC "bad" scrypt params are what made GPU mining even marginally competitive and for those bemoaning no "pure CPU coin" had LTC been made more memory hard (and thus non competitive on GPUs) it would have been even MORE vulnerable to Botnets.

Nobody built Bitcoin to be a "GPU coin" or "FPGA coin" or "ASIC coin".  It was simply built to be efficient and secure on available hardware and overtime that openness lead to more and more efficient hardware.  Today Bitcoin has the potential to secure the network with 17,000% higher efficiency than the first CPU miners.  There is no possible scenario where that is a bad thing.

Then you make it incompatible rather than trying to make one way compatibility with the other fork and thus ensure your efforts are continually overwritten.  I think you misunderstand the point.  The point isn't that forks are impossible it is that they are FORKS and they will not rejoin in the future. 

Which has nothing to do with the topic.

An Avalon ASIC is roughly 17x more efficient than the average FPGA (170 MH/J vs 10 MH/J).
An Avalon ASIC is roughly 85x more efficient than the average GPU (170 MH/J vs 2 MH/J).
An Avalon ASIC is roughly 680x more efficient than the average CPU (170 MH/J vs 0.25 MH/J).

Nobody said faster and more efficient ASICs aren't possible but there is no technology which would allow a 680x increased in efficiency over what the Avalon is capable of.  Going to state of the art, fully custom, optimized 28nm ASIC might allow a 24x increase in performance however after that miners would be limited to Moore's law.  There are no more "shortcuts".

When you consider that at one time "good miners" were operating at 0.25 MH/W efficiency and bad guys could "cheat" (using 28nm full custom ASICs) acheive a roughly 16,000x "shortcut" the gap has been significantly closed.

So why didn't these happen on GPU (which are pretty damn easy to reprogram).  GPU algorithm mining efficiency has all but stalled. 18 months ago people were finding 10% here, and 3% there and that all essentially flatlined almost a year ago.

I doubt there is much algorithm efficiency left.  Now it is possible SHA-256 will be partially compromised which allows for the creation of "optimized" hashers which take that cryptographic flaw into account but that is hard to predict when or if it will happen.

The only major jumps left are catching up to current state of the art technology (28nm process).  After that any future improvements would be limited to Moore's law.  

The Avalon is built using 110nm so there are a jumps in performance using the same chip manufactured on a smaller process (80nm, 65m, 40nm, 28nm).  Each one of these jumps would roughly double the efficiency (MH/J and MH/$).  Also the design of the Avalon is likely not the "end all" in SHA-256 Microprocessor design.  Some additional efficiency can be tweaked out of an improved chip design and improved construction, etc.  How much?  No idea lets guess 50% improvement is possible.

Still even if you put all that together you likely are talking a full custom, perfectly optimized, 28nm miner would be ~(110/28)^2*150% "only" 24x high efficiency as a theoretical upper bound.

Yes.  I got to run so I don't have time to check that math but lets just look at this logically An Avalon is only 57x more energy efficient then a GPU.  IF x GWh requires 21 million Avalons then it would also require 57x 21 million = ~ 1 billion GPU.  I am pretty sure there aren't a billion GPU being used in mining today.  Right?

Of course we will.  GPS vs ASIC it doesn't really matter.  Hashing power (and thus energy consumption) will continue to rise until margins are essentially break even for miners. 

You got some unit conversion issue here.  Pretty obvious from the first state.  If we take the first line as accurate 170 MH/J vs 3 MH/J  (57x) then for same amount of hashing power GPU require ~57x more energy not 58,000 more energy.  

If GPU take 29 GWh (assuming that is correct) the ASICs (a 1/57th the energy) would take ~0.5 GWh (not 0.5 MWh).  

Everything else being equal (total hardware costs, exchange rate, global energy rates, risk, etc) given enough time for the "market" to absorb and react to this change we would expect difficulty to rise by a factor of 57x.  So while the energy per hash will be less it will require more hashes and thus the energy per BLOCK will be roughly the same.  Same energy being spent but now 57x as many hashes are protecting the network.  We saw essentially the same thing in the move from CPU to GPU.

Also impossible baring some flaw in ECDSA and/or RIPEMD-160 SHA-256.  If such a flaw does develop it likely will be over years or decades giving active owners plenty of time to move funds to new higher security addresses leaving only the coins which can't be moved.

I honestly don't think he needs better PR.  Batch 2 will sell out completely.  Every single unit.  Any time/money/effort spent on PR is a waste as people will still find a reason to doubt it, or call it a scam.

A good product sells itself.  If Avalon is a good product (and everything so far indicates it will) he should just focus on getting units shipped, ensure good quality control, and resolving problems/failures.  THAT will bring in the coins.

Generally speaking the companies that spend the most on PR are those who either have an inferior product or are in a highly competitive environment where they can always produce more than they can hope to sell.

Blocks have the prior block's hash as part of their header.  One there is a incompatible fork the chains will never rejoin without orphaning the new fork's blocks and what miner is going to miner a new fork's blocks knowing there is a high probability that will result in an orphan when he/she can just mine the old fork's blocks without risk.

Example:
Block A (original)
Block B ( increased precision - "sub satoshi values")

Both chains are the same until block 1,000,001.
In Block 1,000,0001B a miner produces a block with a "sub satoshi".  This Block will be seen as invalid by network A.

Now Network A produces a block 1,000,0001A and eventually 1,000,0002A.    Network B can't use 1,000,002A without 1,000,0001A (which will orphan the miner which produces the "first sub satoshi block".  No miner would risk including sub satoshi's when that increases the risk its block will be oprhaned.   So the fork will never be used.  The only way to prevent that economic cost is to make the chain incompatible.  If the "B" chain introduces incompatible changes to the blockheader then "A" chain could never oprhan the "B" chain.

That would make 1,000,001A and 1,000,001B mutually exclusive and incompatible with the other network forever.