ASIC mining will force some protocol change rendering 1st gen.ASIC miners useles

[happygeorge]

I found this bet and it made me wonder what everyone here thinks

http://betsofbitco.in/item?id=1432

ASIC mining will force some protocol change rendering 1st gen.ASIC miners useless before Dec.31,2013

All the ASIC hype and huge hashing power will change how the whole Bitcoin protocol will behave and due to some weakness, vulnerability or some unforeseen reason, there will be a small/large change in the protocol that may or may not cause a fork in the blockchain, but regardless of the reason(s);

Before the end of 2013, AVALON's Batch #1,#2,#3 and other such "1st gen." ASIC miners will become useless.

"1st gen." includes the promised BFL ASIC miners (if they ever come) if not, then this only refers to AVALON's first 3 batches.

Later batches that may include modifications do not count.

- an additional reason for killing these ASICs may be a "revolt" by the GPU, FPGA mining crowd to DO SOMETHING!


I think for the purpose of this poll, the exact ending date in the bet is not really interesting.

[1714466224]

1714466224

[1714466224]

1714466224

[1714466224]

1714466224

[erk]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

[happygeorge]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

[jspielberg]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

Only with one player with access to the tech.   Asic miner was in that position briefly, but it has passed.

[erk]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

No because ultimately it will be only ASICs mining, so it will be ASIC vs ASIC for control.

Ignoring ASIC as the dominant technology, is like insisting that we all keep using incandescent lamps instead of LED or CF. When electricity costs $1/kWh if you aren't running ASIC you aint gonna be mining.

[happygeorge]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

Only with one player with access to the tech.   Asic miner was in that position briefly, but it has passed.

Yeah, but even Yifu said that he was amazed that with relatively little "investment" they were able to create the technology (ASICs) to 51% the network (if they wanted to)

Now, I think that still holds, the BIG GUNS can get the ASIC tech, and since the protocol is "locked in" a huge production of ASICs can easily do damage I think, and Yifu said that he thought that others entering the ASIC manufacturing process would need about 8 months minimum, I think he is underestimating the power of professional, and directed effort by one or ALL the BIG PLAYERS who may collude to break the current form of the Bitcoin protocol...forcing a change in the protocol (fork), causing chaos in the community, and maybe a collapse (which they would like very much)

[jspielberg]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

Only with one player with access to the tech.   Asic miner was in that position briefly, but it has passed.

Yeah, but even Yifu said that he was amazed that with relatively little "investment" they were able to create the technology (ASICs) to 51% the network (if they wanted to)

Now, I think that still holds, the BIG GUNS can get the ASIC tech, and since the protocol is "locked in" a huge production of ASICs can easily do damage I think, and Yifu said that he thought that others entering the ASIC manufacturing process would need about 8 months minimum, I think he is underestimating the power of professional, and directed effort by one or ALL the BIG PLAYERS who may collude to break the current form of the Bitcoin protocol...forcing a change in the protocol (fork), causing chaos in the community, and maybe a collapse (which they would like very much)

If a big player wanted to destroy the network they could do it easier and cheaper by just doing stuff to trading side where there are no rules as opposed to the protocol which is based on rules.  The game plan of destroying share value is well established in the OTC market.  The easiest way to destroy confidence in a stock/asset is to prove that market is uneven.   A big player could do that by buying and selling large blocks in the current thinly traded market without having to get their hands dirty with all the engineering issues.

[happygeorge]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

Only with one player with access to the tech.   Asic miner was in that position briefly, but it has passed.

Yeah, but even Yifu said that he was amazed that with relatively little "investment" they were able to create the technology (ASICs) to 51% the network (if they wanted to)

Now, I think that still holds, the BIG GUNS can get the ASIC tech, and since the protocol is "locked in" a huge production of ASICs can easily do damage I think, and Yifu said that he thought that others entering the ASIC manufacturing process would need about 8 months minimum, I think he is underestimating the power of professional, and directed effort by one or ALL the BIG PLAYERS who may collude to break the current form of the Bitcoin protocol...forcing a change in the protocol (fork), causing chaos in the community, and maybe a collapse (which they would like very much)

If a big player wanted to destroy the network they could do it easier and cheaper by just doing stuff to trading side where there are no rules as opposed to the protocol which is based on rules.  The game plan of destroying share value is well established in the OTC market.  The easiest way to destroy confidence in a stock/asset is to prove that market is uneven.   A big player could do that by buying and selling large blocks in the current thinly traded market without having to get their hands dirty with all the engineering issues.

I am not a financial whiz by any stretch of the imagination... so please help me:  If they try to play this "thin" market, wouldn't they have to be VERY CAREFUL not to make it "thick" by their actions, and since it is so wild-west-like, they may just end up loosing, the same way I have, playing the market...

If the market goes up, up, up... there is not much they can do, I believe, they will just become part of all the "action"... and loose/win with all the rest of us. No?

Please explain how the market-makers would destabilize things.

[DPoS]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

No because ultimately it will be only ASICs mining, so it will be ASIC vs ASIC for control.

Ignoring ASIC as the dominant technology, is like insisting that we all keep using incandescent lamps instead of LED or CF. When electricity costs $1/kWh if you aren't running ASIC you aint gonna be mining.

not a good analogy..  you do not have any art in your home

[Carlton Banks]

Avalon played their hand carefully and well;

1. Produce and ship a reference unit, but make the "reference" designation in hindsight.

2. Sell remaining chips to all-comers indiscriminately, see whether and how the various types of modules sink or swim.

Continuing to produce the reference unit would make the specific design of that unit into a target for attack. That's why I say they've played their hand carefully. They were dealing with a landscape with zero commercial competitors when that decision got made, and now they've only got 1.5 competitors (2.5 ASIC manufacturers does not make for a competitive environment).


Maybe those that admire ngzhang + team Avalon's complete designs will get a better chance to have their own when Avalon team are satisfied that the market is diverse enough. It's been a frustrating few months for those (lucky) people waiting on units, but I believe they've adopted the most responsible strategy they could. It was needed.  

[happygeorge]

Avalon played their hand carefully and well;

1. Produce and ship a reference unit, but make the "reference" designation in hindsight.

2. Sell remaining chips to all-comers indiscriminately, see whether and how the various types of modules sink or swim.

Continuing to produce the reference unit would make the specific design of that unit into a target for attack. That's why I say they've played their hand carefully. They were dealing with a landscape with zero commercial competitors when that decision got made, and now they've only got 1.5 competitors (2.5 ASIC manufacturers does not make for a competitive environment).


Maybe those that admire ngzhang + team Avalon's complete designs will get a better chance to have their own when Avalon team are satisfied that the market is diverse enough. It's been a frustrating few months for those (lucky) people waiting on units, but I believe they've adopted the most responsible strategy they could. It was needed.  

Isn't the "chip" the heart of the ASIC mining rigs?  If the protocol changes the various modules would ALL DIE... no?

so there really is no competition really, and as soon as someone else gets a bunch of AVALON or other ASICs the network will become centralized, or it will become a "war betweeen AVALON chips and the new design... and then...

phew, my brain started to SMOKE, I am not meant to think about these things... thanks for your explanation anyway, I will let it simmer for a bit.

[tom_o]

No it could not be changed to outlaw ASICs, unless they change the proof of work from double sha256.

Also the fork you propose happening in the first post would result in massive amounts of hashpower (ASICs) vastly out powering GPUs. Meaning the blockchain that would be seen by the network as 'valid' would be the one the ASICs are creating.

Remember also that most designs have a programmable microcontroller embedded, so as long as the protocol includes a double sha256 hash they will still work with a little reprogramming (Avalon uses a FPGA and klondike a PIC for example)

[PaperClip]

Gee, Brain, what do you want to do tonight?

1) let's say corporation X invest millions of USD to build mining rig with >50% hashing power of network
2) then it compromises BTC by performing 51% attack
3) this becomes noticed by public -> everyone panics and sells BTC -> BTC/USD ratio falls
4) now corporation X has equipment worth millions of USD, and millions of BTC in their wallet. BTC is compromised, nobody will buy it. Mining equipment also becomes not needed by public. Corporation X will have everything, but it will be worth nothing.

Great plan, huh? Nobody capable of this attack will be stupid enough to actually do this.


Also, imagine that protocol change would be performed. At some point of time two protocols would exist - old one (supported by ASICs) and new one (supported by CPU/GPU with updated clients). This will lead to two block chains in network. Currently ASICs generate ~90% of all new blocks, old hardware like CPU/GPU - 10%. ASICs supported protocol will be winning over everything else, and old protocol will continue to exist without any problem.
[edit:] this is the same that tom_o has already posted, while I was typing this one

[Vandroiy]

Normally, the nice thing about Betsofbitcoin is that people with a strong opinion on it can test their opinion by placing some money.

Hoever, this

Bet deadline: Dec. 11, 2013 end of day Eastern Time
Event date: Dec. 31, 2013 end of day Eastern Time

is ridiculous. BOBs weighting function will not compensate against late bets. So betting now makes no sense and the time-frame in the end will be very short.

Also, this is a trick bet. Note the wording: "but regardless of the reasons" invalidates the protocol fork as a condition on the bet! If the 1st gen miners are no longer profitable, they also become useless in this context.

Anyway, this bet setup reeks of trouble. I'd advise to stay away. Might be smart for an admin to remove it because of ambiguous terms.

[happygeorge]

No it could not be changed to outlaw ASICs, unless they change the proof of work from double sha256.

Also the fork you propose happening in the first post would result in massive amounts of hashpower (ASICs) vastly out powering GPUs. Meaning the blockchain that would be seen by the network as 'valid' would be the one the ASICs are creating.

Remember also that most designs have a programmable microcontroller embedded, so as long as the protocol includes a double sha256 hash they will still work with a little reprogramming (Avalon uses a FPGA and klondike a PIC for example)

I hope you are right, because the way you put it, I think I UNDERSTAND IT!



Thanks for the great explanation!

[happygeorge]

Currently ASICs generate ~90% of all new blocks, old hardware like CPU/GPU - 10%.

Is this known for a fact? How can we check this statistic?

The 90%:10% ratio, that is/

[tom_o]

Currently ASICs generate ~90% of all new blocks, old hardware like CPU/GPU - 10%.

Is this known for a fact? How can we check this statistic?

The 90%:10% ratio, that is/

Well ASICMINER have 23% of the network in their own operations so that's a number to start from. http://www.asicminercharts.com/live/

[minternj]

Currently ASICs generate ~90% of all new blocks, old hardware like CPU/GPU - 10%.

Is this known for a fact? How can we check this statistic?

The 90%:10% ratio, that is/

Current hash rate - hash rate before avalons shipped. I think 90 is way off... but i am too lazy to look it up.

[PaperClip]

Currently ASICs generate ~90% of all new blocks, old hardware like CPU/GPU - 10%.

Is this known for a fact? How can we check this statistic?

The 90%:10% ratio, that is/

Current hash rate - hash rate before avalons shipped. I think 90 is way off... but i am too lazy to look it up.

I was using the same idea.
This is estimation of mining devices in network:
http://s20.postimg.org/45jotrtvd/miners.jpg
(https://bitcointalk.org/index.php?topic=227792.msg2461310#msg2461310)
Overall CPU/GPU/FPGA level was almost the same for 5 months before Avalon, and its unlikely to be growing after that. I suppose actually it even started to drop. 10% maybe is too low, but anyway its not more that 15%

[Double-Spent]

It's a silly proposition, because when most coins have been mined the network still needs to be maintained for transactions, and that maintenance needs to be as energy efficient as possible as there won't be much profit in running it, nothing beats an ASIC in energy efficiency for mining crypto.

Yeah, but I think that is only half of the equation, the other half is "control" and with ASICs the big players can very easily manufacture a gizzilion chips and 51% the network, no?

Only with one player with access to the tech.   Asic miner was in that position briefly, but it has passed.

Yeah, but even Yifu said that he was amazed that with relatively little "investment" they were able to create the technology (ASICs) to 51% the network (if they wanted to)

I'm amazed by him being amazed. It's so freackin' obvious that the investment required to pwn the network is minimum with ASIC techology, I don't understand how he can be "amazed" by that fact.