Why ASIC's Should Not Be The Future Of Crypto Currencies

[Gabi]

"competition"? To me it looks like avalon and bfl use exactly the same asic  similar prices, similar performance and similar release date 

[Inaba]

Hah.  That's a joke, right?  You need to get your eyes examined, then.

Avalon is on an ancient 110nm technology while the BFL chip is the most advanced bitcoin mining ASIC in on the planet by several orders of magnitude, performance is 7x better than Avalon and the release date is 2 months prior to Avalon... yeah, exactly the same!

[Jutarul]

I've seen some POW ideas that I find intriguing for other reasons— like amiller's idea of using fast random access to the txout set as proof of ability to rapidly process Bitcoin transactions— but I've not seen anything that prevents people with custom hardware from getting a substantial advantage over the public— except by getting specialized hardware into the hands of the public.

I second that. We'd need more ideas of how to implement a fair POW system which cannot be easily corrupted. Satoshi ruled against Proof-of-IP, because it's even easier to attack.

A fundamentally fair system rewards the miner by their contribution to the network (in terms of transaction processing and related services). The difficulty increase does not add any beneficial factors to the bitcoin network, except resilience against a 51% attack. Thus the focus should lie on how to implement resilience against 51% attacks without the consumption of a lot of energy.

[scrybe]

I've seen some POW ideas that I find intriguing for other reasons— like amiller's idea of using fast random access to the txout set as proof of ability to rapidly process Bitcoin transactions— but I've not seen anything that prevents people with custom hardware from getting a substantial advantage over the public— except by getting specialized hardware into the hands of the public.

I second that. We'd need more ideas of how to implement a fair POW system which cannot be easily corrupted. Satoshi ruled against Proof-of-IP, because it's even easier to attack.

A fundamentally fair system rewards the miner by their contribution to the network (in terms of transaction processing and related services). The difficulty increase does not add any beneficial factors to the bitcoin network, except resilience against a 51% attack. Thus the focus should lie on how to implement resilience against 51% attacks without the consumption of a lot of energy.

Well we have the PPCoin experiment running, so we can see if proof of stake is a way forward, but they still use PoW for minting.

The other thing that PoW does that some folks forget, is it gives us an unhackable time source that does not rely on external time servers or anything else that could potentially be spoofed. You know locally when each block happened, so you have a reliable timestamp, and don't have to trust anyone. Moving to a pure PoS system would remove that timing mechanism, which is why PPC is a hybrid and was initially launched with check-pointing servers until the PoW activity became high enough.

You have to have a constraint that can reliably be applied to all members of the network. There are a lot of choices, but relatively few that are immune to manipulation. Bitcoin is based on pure CPU PoW, LiteCoin uses CPU and cache/memory IO, and PPC uses pure CPU as it's constraint for block generation to give us this timestamp. Other theoretical constraints include network latency or bandwidth (potentially manipulable by building well connected private clouds, but there could be some benefits to this like the Swiss are doing,) persistent storage, or even real-world physical token passing (and 2 guys pass a bag of tokens back and forth as fast as possible to do a 51% attack.)

When I start thinking about a potential future where we want to control power usage, I'm noticing that really what you are talking about is "increasing the cost of mining equipment so that power cost/performance parity is reached earlier" So instead of an ASIC that costs $1300 and uses 1W/GH, you have a computer that costs $2500 and uses 5W/(GH equiv) of power. Assuming that folks like bitcoin and keep mining, it would end up being a completely power neutral move that would have no advantage in security, minting, or timekeeping. (my opinion, but I think the conclusion is pretty obvious)

The good news is that you are stressing out over a short term issue that does not need to be solved because Satoshi designed the solution into it. Block halving will make this look like the 1849 gold rush (http://en.wikipedia.org/wiki/California_Gold_Rush) once the reward drops to a level that is not economic for all but the most efficient operations. These larger and more efficient operations will be motivated by profits, but will likely be more interested in maintaining the network security to ensure bitcoin profits in other endeavors rather than directly from mining itself. I would not be surprised at all if the BitCoin world of 2030 is almost 100% institutional (finanacial houses/banks) mining, with partnering agreements on planned hash levels to help reduce costs, and a big warehouse full of standby gear that can be powered on if a rouge does appear on the network.

So even if PoW is not ideal, it still seems to be the best solution, and I don't see any real advantage to changing it from a simple solution to a more complex one if it does not bring a substantial additional value.

[Jutarul]

The good news is that you are stressing out over a short term issue that does not need to be solved .... once the reward drops to a level that is not economic for all but the most efficient operations.

I don't feel stressed at all. Thank you. Let POW be with us for as long as we see fit.

I don't understand how the convergence towards efficient operations solves anything. It may be efficient in terms of profit margin, but can still be wasteful. After all, bitcoin users will pay for each consumed kWh in terms of transaction fees.

[scrybe]

The good news is that you are stressing out over a short term issue that does not need to be solved .... once the reward drops to a level that is not economic for all but the most efficient operations.

I don't feel stressed at all. Thank you. Let POW be with us for as long as we see fit.

I don't understand how the convergence towards efficient operations solves anything. It may be efficient in terms of profit margin, but can still be wasteful. After all, bitcoin users will pay for each consumed kWh in terms of transaction fees.

Sorry, stressed might be too strong a term, I like using overstatement.

I doubt very much that we will see aggregate fees increase to anywhere close to the level that mining revenue is at now, which means that overall the level of available money to miners will decrease over time.

Combine this with power costs and it should lead to a smaller network (some miners leave, others reduce investment) and more efficient hardware that consumes less electricity on aggregate. This is because after the minting operation is complete, 33% of the value of the mining community (to the bitcoin network) is gone, and now it's focus should be on making it's core functions more secure instead of subsidising a larger set of miners than needed. As part of this we will have to redefine a "small transaction" at some point and lower the fees on them proportionately, because having to pay a $5 transaction fee when buying a coffee because BTC makes it up to $1000 is not going to work.

I think large institutions will get into the game at some point and we will start seeing peering agreements that happen outside of the typical transaction fee mechanism. Once that happens all bets are off on fee structure as they are likely to try to eliminate them rather than increase them. That's because large institutions will be able to take a loss on their bitcoin farm (and standby equipment, and ASIC on-demand manufacturer contract) because they can make so much more from the services they offer WITH bitcoin, but they are going to want to minimize that loss.

[markm]

Buying a coffee with bitcoin might not make sense at all once bitcoin is a high value per coin, and transaction fees might be aiming more to have such trivially tiny purchases move to retail-purchase merged-mined blockchains instead of cluttering up the international balance of payments settlement network that is the original bitcoin chain.

Merged chains could also become significant sources of revenue for miners, as every Tom Dick Harry Airmile and Walmart starts looking into establishing its own brand of Canadian Tire Money.

-MarkM-

[scrybe]

Buying a coffee with bitcoin might not make sense at all once bitcoin is a high value per coin, and transaction fees might be aiming more to have such trivially tiny purchases move to retail-purchase merged-mined blockchains instead of cluttering up the international balance of payments settlement network that is the original bitcoin chain.

Merged chains could also become significant sources of revenue for miners, as every Tom Dick Harry Airmile and Walmart starts looking into establishing its own brand of Canadian Tire Money.

-MarkM-

Why the heck do we need 8 decimal places if the transaction fee is always .005? I thought a core tenant of bitcoin was it's ability to adjust to increased value by reducing fees, we have moved from .01 to .0005 already, why would that not continue? The real question is how big a network (in terms of power usage) we want to grow to, and adjust the transaction fees to match the typical costs so that those that are wasteful are losing money and those that are efficient are making a profit...

As for merged chains, they have their own rewards in their own currency, that revenue is of consideration for the pool and miner, but not for BitCoin itself. It's just piggybacking on the work available. If miners are making enough on these alt-chains then maybe bitcoin does not need to pay fees except for anti-spam reasons, it adds value to the miners by enabling the other protocol to profit in place of a direct reward.

(Of course, I have to ask WHY Walmart would be willing to let anyone mine their chain instead of just running a closed system.)

You sparked a new thought for me there, is/will there be an implementation that allows automated p2p exchanges using the mechanisms mentioned here: https://en.bitcoin.it/wiki/Contracts#Example_5:_Trading_across_chains? If so it could enable the future you mention where the transaction fee stays high, but we are going to have to have a completely automated conversion system in place to make it work as easily as just lowering fees. IIRC, the mandatory fees are anti-spam as much as anything else, if the blocks start filling up then a market for fees will start to be established as folks bid on space in the block for their transactions.

[FLHippy]

These larger and more efficient operations will be motivated by profits, but will likely be more interested in maintaining the network security to ensure bitcoin profits in other endeavors rather than directly from mining itself. I would not be surprised at all if the BitCoin world of 2030 is almost 100% institutional (finanacial houses/banks) mining, with partnering agreements on planned hash levels to help reduce costs, and a big warehouse full of standby gear that can be powered on if a rouge does appear on the network.

You had me up till the concept of miners hevily invested in hardware continuing to mine so that bitcoin continues.... I don't see that happening at all. Just like poeple will sell their GPUs when it's not profitable to mine, those who bought into ASIC expecting profits will shut down when there are none also.

I'm shutting down and selling my GPUs tuesday, maybe wednesday.

[scrybe]

These larger and more efficient operations will be motivated by profits, but will likely be more interested in maintaining the network security to ensure bitcoin profits in other endeavors rather than directly from mining itself. I would not be surprised at all if the BitCoin world of 2030 is almost 100% institutional (finanacial houses/banks) mining, with partnering agreements on planned hash levels to help reduce costs, and a big warehouse full of standby gear that can be powered on if a rouge does appear on the network.

You had me up till the concept of miners hevily invested in hardware continuing to mine so that bitcoin continues.... I don't see that happening at all. Just like poeple will sell their GPUs when it's not profitable to mine, those who bought into ASIC expecting profits will shut down when there are none also.

I'm shutting down and selling my GPUs tuesday, maybe wednesday.

ONLY in the case where there is another revenue stream that is directly supported by their continued unprofitable participation in bitcoin mining.

For example, thinking like a multinational bank, it might make millions or billions of dollars a year on fees related to bitcoin utilization, so they can factor that into the overall P%L sheet with both departments included. So while the mining department is like IT (sucking in money and turning it into heat (disclaimer, I work in IT)) at a loss of $4m a year, and you get $44m a year in fees, then you are $40M in the black, and you know that you have enough capacity (5% for a really big bank?) to prevent a double-spend attempt from any of your evil competitors, or ensure that your transactions are confirmed within a certain timeframe without excessive fees. Who cares if the mining is a losing proposition when you are making 10 times as much overall!

I'm not thinking that this makes sense for anyone who does not have something like "maintain X significant percentage of hashrate" as a key component to some other business plan. So independent miners are likely to have to run extremely lean, and it might not even ROI out if you have free power if the big boys start throwing their weight around. It's also not likely to stay that way permanently, the appetite for the loss leader status will run out and they will figure out a way to reduce hashing closer to parity, but it could impact us for several quarters if it happens.

I do think it will overall hit an equilibrium with a modest (5-15%) profit as time goes on, but in the start up phase we are going to see a lot of weird shit as we hit more diverse and larger markets. I'm hoping for at least a couple years before mining gets strange in some way, guessing maybe 5 at the outside.

[scrybe]

These larger and more efficient operations will be motivated by profits, but will likely be more interested in maintaining the network security to ensure bitcoin profits in other endeavors rather than directly from mining itself. I would not be surprised at all if the BitCoin world of 2030 is almost 100% institutional (finanacial houses/banks) mining, with partnering agreements on planned hash levels to help reduce costs, and a big warehouse full of standby gear that can be powered on if a rouge does appear on the network.

You had me up till the concept of miners hevily invested in hardware continuing to mine so that bitcoin continues.... I don't see that happening at all. Just like poeple will sell their GPUs when it's not profitable to mine, those who bought into ASIC expecting profits will shut down when there are none also.

I agree. Heck, look at the situation even now = unless it's profitable, 99.99% Bitcoiner won't do it, regardless of negative impacts.

took you off ignore because this was a good comment.

You point out "now" as the timeframe, I'm talking about 2-5 years out when we have much bigger fish swimming in the bitcoin pond.

You always have to take the whole business model into account. Google should be broke based on this logic, but they are not because the monetize the hell out of advertising and other advantages they can make out of their data and search capabilities. Amazon Kindle is another great example. All it will take is one large company that "needs to have x% no matter the cost" as part of a plan and we will see them going towards mining at a loss.

Of course now that I think about it, the loss might just be for Enterprises if they don't build efficient farms, so the solar powered mining rig might be a gift that just keeps on giving and giving.

[Syke]

Avalon is on an ancient 110nm technology while the BFL chip is the most advanced bitcoin mining ASIC in on the planet by several orders of magnitude, performance is 7x better than Avalon and the release date is 2 months prior to Avalon... yeah, exactly the same!

I suggest you read up on the definition of Orders of magnitude

[Inaba]

Avalon is on an ancient 110nm technology while the BFL chip is the most advanced bitcoin mining ASIC in on the planet by several orders of magnitude, performance is 7x better than Avalon and the release date is 2 months prior to Avalon... yeah, exactly the same!

I suggest you read up on the definition of Orders of magnitude

Or...  I suggest you look up the term vernacular before you try to educate me with grade school trolling.  I'll give you a freebie, too:  My use of "ancient" isn't technically accurate either.  You might also want to look up hyperbole since you are unfamiliar with both it and "vernacular".  They are useful words that you should have learned in grade school.

Both of these lessons are free of charge, the next one will cost you though.

PS - You should have a period after "magnitude" above. (Something else you should have learned in grade school.)

[Syke]

Avalon is on an ancient 110nm technology while the BFL chip is the most advanced bitcoin mining ASIC in on the planet by several orders of magnitude, performance is 7x better than Avalon and the release date is 2 months prior to Avalon... yeah, exactly the same!

I suggest you read up on the definition of Orders of magnitude

Or...  I suggest you look up the term vernacular before you try to educate me with grade school trolling.  I'll give you a freebie, too:  My use of "ancient" isn't technically accurate either.  You might also want to look up hyperbole since you are unfamiliar with both it and "vernacular".  They are useful words that you should have learned in grade school.

Both of these lessons are free of charge, the next one will cost you though.

PS - You should have a period after "magnitude" above. (Something else you should have learned in grade school.)

Ah, gotcha. You're just making things up, like that 7x performance increase and the 2 months prior shipping date. I guess the reality is Avalon is faster and will ship sooner. Thanks for clearing that up.

[cedivad]

Avalon is on an ancient 110nm technology while the BFL chip is the most advanced bitcoin mining ASIC in on the planet by several orders of magnitude, performance is 7x better than Avalon and the release date is 2 months prior to Avalon... yeah, exactly the same!

I suggest you read up on the definition of Orders of magnitude

Or...  I suggest you look up the term vernacular before you try to educate me with grade school trolling.  I'll give you a freebie, too:  My use of "ancient" isn't technically accurate either.  You might also want to look up hyperbole since you are unfamiliar with both it and "vernacular".  They are useful words that you should have learned in grade school.

Both of these lessons are free of charge, the next one will cost you though.

PS - You should have a period after "magnitude" above. (Something else you should have learned in grade school.)

Ah, gotcha. You're just making things up, like that 7x performance increase and the 2 months prior shipping date. I guess the reality is Avalon is faster and will ship sooner. Thanks for clearing that up.

Lol, this is even possible.

[Inaba]

Ah, gotcha. You're just making things up, like that 7x performance increase and the 2 months prior shipping date. I guess the reality is Avalon is faster and will ship sooner. Thanks for clearing that up.

No, AVALON is 400w at 60 GH/s.  Ours is 60w... now, what is 400/60?  Tell everyone, please.

BFL is scheduled to ship end of November, Avalon, sometime in January.  How many days are between December 1st and January 31st?  Tell everyone please?

Now, with your answer, how many days are in a month?  Divide that by the answer you got above.  Tell everyone, please?

Are you really so stupid that you need me to hand hold you through basic math?  I had given you more credit, but I see that credit was misplaced. 

[SolarSilver]

No, AVALON is 400w at 60 GH/s.  Ours is 60w... now, what is 400/60?  Tell everyone, please.

No, Avalon is worst case 400 Watts but once they have the real boards and not a prototype board, they will announce the real life power consumption. Stop twisting your competitors announcements (just like you did with the bASIC).

Are you still as sure about that 60 Watts / 60 GH/s just like the performance and power consumption of BFLs first product turned out? How well did that calculated guess work out?

Until you have a working product, you can only make guesses. Will you give people a refund if your product underperforms?

[Inaba]

How is that twisting the announcement?  That is the number they released, am I suppose to just pull numbers out of my ass?  Fine, the Avalon will use 900w and hash at 800 MH/s, hows that?  What did I twist for bASIC?  Tom still hasn't released numbers and he's suppose to ship in a couple weeks.

[cedivad]

How is that twisting the announcement?  That is the number they released, am I suppose to just pull numbers out of my ass?  Fine, the Avalon will use 900w and hash at 800 MH/s, hows that?  What did I twist for bASIC?  Tom still hasn't released numbers and he's suppose to ship in a couple weeks.

Where can I bet that he will ship at most 1 week later than you or better?

Can you STOP?

[Badonkadonk]

inaba, pls stop feeding the troll, or ill have to go make popcorn lol