Satoshi Nakamoto: "Bitcoin can scale larger than the Visa Network"

[franky1]

I am truly unsettled; I have switched from Core 0.11 to Classic 0.11 to Core 0.12 to Classic 0.12 and it is totally possible/likely I will switch again.  Perhaps I will try using more advanced mathematics, above and beyond high school maths.

#0: Reducing unproductive contention in the Bitcoin community is a personal goal.

#1: Right now my top interest is in the vulnerability to transactions with lots of inputs that lead to long compute times for verifications.  One wonders why we don't see bad actors using this against Bitcoin relentlessly.  Maybe there are less malicious folks attacking than I hear about?  Maybe the bad guys don't know how much this can hurt?

#2: Secondarily I would like to see a concerted effort be made to 1) educate users about setting fees well to facilitate quick confirmations, i.e. to avoid long/painful/anxiety-causing commit times and 2) enhancing wallets to make it automatic and default to have high enough fees.

#3: Thirdly; I do highly prefer adoption much more than increasing fees which means to me capacity.  Increasing fees will have to come eventually but in the meantime I would happily sacrifice fees until adoption is really widespread.

#4: Fourthly; Is there something to the Bitcoin Unlimited stuff?

#5: Lastly; all of those other features/functions like transaction malleability, etc.  Perhaps these should be higher on my list but I haven't dug into them yet; sorry.

With those in mind;

block size limit, e.g. 2MB; helps with #3, hurts #1, we should resist doing it just because it seems obvious, obvious is not a reliable attribute

SegWit; helps with #3, hurts #1, introduces other complexities that might have subtle consequences

limit inputs: all by itself this is great for #1, seems simple enough, workaround is trivial, i.e. create multiple small transactions

#0 this can be solved by core adding in the code, to not cause the contention they cry about. then with all the different code bases (btcd, bitcoinj, classic, core, etc) having the code, it is then just a waiting game to see if miners decide to upgrade too knowing all the users are ready to accept such changes.

#1 this can be solved by knowing REALISTICALLY the times it actually takes to sort through a transaction with 50 inputs, 100 inputs, 1000 inputs. and then see how many times people have made a genuine(not attack) transaction of such. to set an arbitrary rule to ignore transactions with such amount of inputs, just like they ignore transactions with no fee.

#2 1) also educate people how to code/make transactions with least bloat/random use of inputs. to help them decrease their own cost of sending a tx, aswell as helping reduce chances of verification time problems.
#2 2) its a little too early to push the transaction fee up.. id say it should be a slow process over years-decades, not a rush job to try making 3000+ transactions every 10 minutes have high fee's meaning the fee naturally increases because there are only 2000 transactions(average) allowed in per block.
imagine it. 3000 tx pay 4c.. 2000 get allowed in block1 and 1000 in block2.. then the next 3000 tx only 1000 is in block 2 and 2000 are in block 3. meaning that if the 3rd set of transactions have any hope of getting in block 4 or 5 they are likely to have to pay a premium to fight off competition..
so within 5 blocks the price begins to rise.

the fee is not an essential part of mining income. it is just a small bonus. it does not need to, nor should offset the reward for many years. because the reward is part of the mechanism that helps the speculation of the deflationary price(price rise). by flooding miners with lots of coins will make them spend them just as fast which not only affects the valuation of bitcoin. but also pushes miners to add more ASICS, and raising the difficulty. making bitcoin more centralized due to the smaller pools losing the competition. also pushing customers/users away from bitcoin because a fee does not actually guarantee the very next block(no guarantee of confirmation in 10 minutes)
pushing the fee has a knock on affect on many aspects. and there is no logic to push it too fast.

#3 agreed. allowing twice as much buffer room(2mb) to grow, or even 4x as much buffer room(2mb+segwit) allows for NATURAL SLOW growth without pushing or irritations and without demanding to dev-team for just an extra spoon of buffer every couple months knowing it takes 2 years to get the spoon.

#4 i personally am not in BU camp, xt camp, classic camp or core camp. i just want more buffer space without beeding to be spoonfed by developers. BU is the premiss that there needs no hard limit. and miners can set their own preferential soft limits. but best to ask someone(hopefully they reply unbiasedly) who has researched it in more detail

#5 i agree it doesnt need to be a classic OR core. it needs both. in combination

edit
your final sentence answered the #1 for yourself and more elegantly then i did

[franky1]

Premise:
- 64MB blocks are allowed
- 64MB blocks take over 10 minutes to verify
- Someone wants to destroy the network ( 100 million dollar investment )
- Other miners want to let him destroy the network

Conclusion:
BITCOIN IS DOOOM!


there's nothing stopping the other miners from orphenning the attackers blocks... if his block give them much inconvenience for whatever reason they will orphen them...

EDIT. my now removed statement was on the bases of a 1mb attack. now lets make it 64mb..
if it takes 10minutes to validate a malicious 1mb. then it takes 640 minutes to validate a 64mb.. and so malicious miner is 638 minutes behind the competitors before it even begins to hash a block, because neutral miners will ignore such stupid transactions just like they ignore 0 fee transactions and make a standard block using more rational size transactions

[sgbett]

The elephant in the room just will not go away though. The limit was a temporary fix.

You do realise that a block as big as 64 MB will take a lot longer than 10 minutes to verify on anything but the most costly hardware?

(so if Satoshi had "got it right" with the original block size then Bitcoin wouldn't be able to even produce blocks as quickly as 10 minutes and no-one but corporations would be able to afford to even run full nodes - but hey who cares about inconvenient truths such as those)

Premise:
- 64MB blocks are allowed
- 64MB blocks take over 10 minutes to verify

Question:
What financial incentive exists for a miner to produce such a block?

That's not even really the appropriate question. The real question you should be asking is "What would prevent someone who wants to destroy the network from producing such a block?"

But you don't ask that because ... (you tell me).

I do not ask the question because, an attacker intending to destroy the network would not produce such a block, because such a block does not destroy the network.

This transaction monopolization can only happen IF all the other mining pools choose to mine on top of that monopolizing mining pool's blocks. Yet, doing so will result in lower profitability for the other (the majority) of pools since they (having smaller validation capacity) must always mine 1-txn blocks and are therefore unable to reap transaction fees. This is an unstable situation - if a single mining pool chooses to ignore the large block and is able to find a small competing block while other pools are still validating a large block, it is in the other pools best interest to switch to this new sibling6. By switching, the other pools reduce the risk that they are mining on top of an invalid block, and can mine blocks with transactions. But if mining pools know that the majority will switch to a discovered sibling, it is rational for all pools except for the producer of the large block to search for a sibling rather than produce a 1-txn block.

Knowing that an expensive to validate block is mitigated as an attack vector, the question remains as to why somebody would produce such a block?

The implication being that if you have the money/hardware to produce blocks, your most profitable course of action is to just mine blocks honestly.

[adamstgBit]

miners are free to determine if a block is valid or not by whatever standards they choose. they are by no means forced to accept any block. the majority of miners need to accept the block for this block to be valid. therefore any argument that starts with " if a miner is a bad actor " is not valid.  you need to start with " if majority of miners are bad actors ". At best the bad guy miner could cause some disruption during the time it take all the other miners to agree to whatever new rules they need to implement to start ignoring the attackers blocks. these new rules do not need a HF.

is there anything false with this statement?

[franky1]

miners are free to determine if a block is valid or not by whatever standards they choose. they are by no means forced to accept any block. the majority of miners need to accept the block for this block to be valid. therefore any argument that starts with " if a miner is a bad actor " is not valid.  you need to start with " if majority of miners are bad actors ". At best the bad guy miner could cause some disruption during the time it take all the other miners to agree to whatever new rules they need to implement to start ignoring the attackers blocks. these new rules do not need a HF.

is there anything false with this statement?

nope
i agree with your statement, infact it adds another level ontop of my statement. because my statement was under the premiss that 64mb blocks were acceptable as standard and that there was little to no code to void a competitors solved block due to malicious bloat. and only code for miners to just ignore or accept malicious transactions within their own attempts of making a block.

similar to the 0fee ignore game. some miners ignore transactions without fee's but if a competitor had a block solution and those had tx's without fee's the ignorant miner would still accept its competitors solution.

but you are quite right the miners could also reject a block if it does not like the content of their competitors block,

[btcusury]

The elephant in the room just will not go away though. The limit was a temporary fix.

You do realise that a block as big as 64 MB will take a lot longer than 10 minutes to verify on anything but the most costly hardware?

(so if Satoshi had "got it right" with the original block size then Bitcoin wouldn't be able to even produce blocks as quickly as 10 minutes and no-one but corporations would be able to afford to even run full nodes - but hey who cares about inconvenient truths such as those)

Premise:
- 64MB blocks are allowed
- 64MB blocks take over 10 minutes to verify

Question:
What financial incentive exists for a miner to produce such a block?

That's not even really the appropriate question. The real question you should be asking is "What would prevent someone who wants to destroy the network from producing such a block?"

But you don't ask that because ... (you tell me).

I do not ask the question because, an attacker intending to destroy the network would not produce such a block, because such a block does not destroy the network.

This transaction monopolization can only happen IF all the other mining pools choose to mine on top of that monopolizing mining pool's blocks. Yet, doing so will result in lower profitability for the other (the majority) of pools since they (having smaller validation capacity) must always mine 1-txn blocks and are therefore unable to reap transaction fees. This is an unstable situation - if a single mining pool chooses to ignore the large block and is able to find a small competing block while other pools are still validating a large block, it is in the other pools best interest to switch to this new sibling6. By switching, the other pools reduce the risk that they are mining on top of an invalid block, and can mine blocks with transactions. But if mining pools know that the majority will switch to a discovered sibling, it is rational for all pools except for the producer of the large block to search for a sibling rather than produce a 1-txn block.

Knowing that an expensive to validate block is mitigated as an attack vector, the question remains as to why somebody would produce such a block?

The implication being that if you have the money/hardware to produce blocks, your most profitable course of action is to just mine blocks honestly.

You are missing my point because you are assuming that the motivations of all participants is that of monetary profit. You can't just ignore the observation that decentralized cryptocurrency represents the only real threat to the central banksters in a very long time. The point is the same for 64MB as it is for 2MB, only more obvious in that the problems 64MB would create are much greater.

[dwma]

By Moore's Law, we can expect hardware speed to be 10 times faster in 5 years and 100 times faster in 10. Even if Bitcoin grows at crazy adoption rates, I think computer speeds will stay ahead of the number of transactions.

Did it increase by tenfold in 5 years? Not even close. Satoshi did not have the adequate data here.

Satoshi certainly didn’t do much (if any) analysis of the scaling limitations of Bitcoin.

We need not appeal to authority. Just because Satoshi invented it, that does not mean that he knows the answers to everything.

I suspect Moore's law is dead and I'm not sure it ever existed for spinning drives or SSDs which will be the problem, not CPUs.

[AliceWonderMiscreations]

By Moore's Law, we can expect hardware speed to be 10 times faster in 5 years and 100 times faster in 10. Even if Bitcoin grows at crazy adoption rates, I think computer speeds will stay ahead of the number of transactions.

Did it increase by tenfold in 5 years? Not even close. Satoshi did not have the adequate data here.

Satoshi certainly didn’t do much (if any) analysis of the scaling limitations of Bitcoin.

We need not appeal to authority. Just because Satoshi invented it, that does not mean that he knows the answers to everything.

I suspect Moore's law is dead and I'm not sure it ever existed for spinning drives or SSDs which will be the problem, not CPUs.

In 1998 I paid $200 for a 2 GB Seagate SCSI drive so I could install MKLinux DR3 on my 233 MHz Beige G3.

Looking at newegg, I can get a 4 TB platter drive for about $200 or a 1 TB SSD (500 MB for top brand) for $200.

[sgbett]

The elephant in the room just will not go away though. The limit was a temporary fix.

You do realise that a block as big as 64 MB will take a lot longer than 10 minutes to verify on anything but the most costly hardware?

(so if Satoshi had "got it right" with the original block size then Bitcoin wouldn't be able to even produce blocks as quickly as 10 minutes and no-one but corporations would be able to afford to even run full nodes - but hey who cares about inconvenient truths such as those)

Premise:
- 64MB blocks are allowed
- 64MB blocks take over 10 minutes to verify

Question:
What financial incentive exists for a miner to produce such a block?

That's not even really the appropriate question. The real question you should be asking is "What would prevent someone who wants to destroy the network from producing such a block?"

But you don't ask that because ... (you tell me).

I do not ask the question because, an attacker intending to destroy the network would not produce such a block, because such a block does not destroy the network.

This transaction monopolization can only happen IF all the other mining pools choose to mine on top of that monopolizing mining pool's blocks. Yet, doing so will result in lower profitability for the other (the majority) of pools since they (having smaller validation capacity) must always mine 1-txn blocks and are therefore unable to reap transaction fees. This is an unstable situation - if a single mining pool chooses to ignore the large block and is able to find a small competing block while other pools are still validating a large block, it is in the other pools best interest to switch to this new sibling6. By switching, the other pools reduce the risk that they are mining on top of an invalid block, and can mine blocks with transactions. But if mining pools know that the majority will switch to a discovered sibling, it is rational for all pools except for the producer of the large block to search for a sibling rather than produce a 1-txn block.

Knowing that an expensive to validate block is mitigated as an attack vector, the question remains as to why somebody would produce such a block?

The implication being that if you have the money/hardware to produce blocks, your most profitable course of action is to just mine blocks honestly.

You are missing my point because you are assuming that the motivations of all participants is that of monetary profit. You can't just ignore the observation that decentralized cryptocurrency represents the only real threat to the central banksters in a very long time. The point is the same for 64MB as it is for 2MB, only more obvious in that the problems 64MB would create are much greater.

No such assumption was made. In the quoted text a method by which *other* miners who are motivated by money act to thwart this kind of attack by mining a smaller sibling block, building on that and ultimately orphaning the attacker's block.

There are simpler ways that a theoretical bad actor with no profit motivation and lots of money can destroy bitcoin. Setting up a mining operation to craft malicious expensive to validate blocks doesn't seem like a good angle to me!

[rizzlarolla]

I like CIYAM. He's grumpier than I am. Wow. 

I also agree with his ad sig attitude.
And I like his individualism, his strength of mind.
And I like his intelligence.

I have been trying to post in his support.
I failed. No post.

I just don't agree with his take on the way forward, any more, sorry.

I don't even want to get technical, CIYAM will wipe the floor with me.
I doubt CIYAM will therefore hold me in any regard. Unfortunately.

I didn't want a block size increase. I think I do now.
A limited increase. (1.5) 2mb. If it is "easily" doable.
I understand, in the scheme of things, it is.

Leave segwit for now. (and hopefully forever)
2mb is needed, safe, obvious, and simplistic?
(needed for the steady growth we have watched over the years, as I think David R said)

[monsanto]

Sure don't sound like Szabo.

[QuestionAuthority]

I like CIYAM. He's grumpier than I am. Wow. 

I also agree with his ad sig attitude.
And I like his individualism, his strength of mind.
And I like his intelligence.

I have been trying to post in his support.
I failed. No post.

I just don't agree with his take on the way forward, any more, sorry.

I don't even want to get technical, CIYAM will wipe the floor with me.
I doubt CIYAM will therefore hold me in any regard. Unfortunately.

I didn't want a block size increase. I think I do now.
A limited increase. (1.5) 2mb. If it is "easily" doable.
I understand, in the scheme of things, it is.

Leave segwit for now. (and hopefully forever)
2mb is needed, safe, obvious, and simplistic?
(needed for the steady growth we have watched over the years, as I think David R said)

I like CIYAM point of view as well. I don't think he's grumpy. He never used to be like this. He's just like me, he's just sick of the bullshit.

[SpiryGolden]

Meanwhile they are 66% of network left that are running Bitcoin Core code.

If it gets at 50% what that means? They've lost majority ?

[Lauda]

Meanwhile they are 66% of network left that are running Bitcoin Core code.

Do you not read what I write to you or is your brain unable to comprehend what I'm saying? I'm not certain which the case is. The number of nodes is not a reliable metric, especially not in a short time frame. This is because people can jump-start a lot of nodes in a very small amount of time. Additionally, it is possible to run fake nodes as well. 

If it gets at 50% what that means? They've lost majority ?

It means nothing. Exactly nothing, especially when a Sybil-attack is being promoted.

He's just like me, he's just sick of the bullshit.

Welcome to the club.

[BittBurger]

Can someone please explain to me how we have these Satoshi quotes and yet there is still any discussion / debate / disagreement on it?  

I don't understand how this can be viewed as anything other than what it is:  A complete deviation from the expected design of Bitcoin for the financial profit of a company that wants to take transactions off the main chain.  

And how the lack of urgency to scale is anything other than what it is:  A means to an end for ensuring the Lightening Network will make money.

This is an honest, genuine question.  Until I found the Satoshi quotes I was on the fence on this issue.  

Now im just like "what the actual fuck?"

-B-

[AlexGR]

Can someone please explain to me how we have these Satoshi quotes and yet there is still any discussion / debate / disagreement on it?  

The quote says one thing, reality says another.

Reality = right now you can't get to a blocksize where you can do visa level tx capacity.

And this is similarly true, not only for bitcoin, but bitcoin-based clones and other blockchain-based systems.

[madjules007]

The limit was a temporary fix.

Shortly before disappearing, Satoshi also repeated over and over that there were more ways that Bitcoin could be successfully DOS attacked than he could count.

The limit (in conjunction with transaction fees) were intended as a DOS attack and spam deterrent. Do you have any evidence that those risks have been mitigated? Do you have any evidence that suggests that 2MB blocks won't be filled to capacity right away? That would leave us where we are today -- with no scaling solutions, wondering how much more un-optimized throughput the network can safely handle.  

I don't believe a second, compatible implementation of Bitcoin will ever be a good idea.  So much of the design depends on all nodes getting exactly identical results in lockstep that a second implementation would be a menace to the network.

Satoshi's description of how to increase the block size limit was also clearly in the context of a software update. Not an incompatible implementation of bitcoin that essentially attacks the network and intentionally forks from all other client nodes. He even thought compatible alternatives were a bad idea.

One could make the argument that Satoshi thought we could raise the limit as/when needed (as if we needed his authority anyway). In that respect, the question of necessity is subjective and debatable, and it is not immediately clear that an increase to 2MB now, with no attempts to make throughput more scalable is necessary.

One could not make the argument that Satoshi thought we should increase the block size limit through a contentious hard fork.

All of this is moot. We should be talking in terms of "what is best for bitcoin" -- with respect to users, nodes and miners -- not talking past each other with interpretations of things Satoshi said 5-6 years ago. Not only was Satoshi wrong about some things, but the state of bitcoin has changed a lot. Hell, if we left the codebase as he originally wrote it, those 184 billion bitcoins from the August 2010 value overflow incident would still be with us. Better to work towards bitcoin's principles than aimlessly trying to fit some arbitrary interpretation of Satoshi's words.

[AliceGored]

The limit was a temporary fix.

Shortly before disappearing, Satoshi also repeated over and over that there were more ways that Bitcoin could be successfully DOS attacked than he could count.

Well then, as he talked about "other ways" should we disregard this:

It can be phased in, like:

if (blocknumber > 115000)
    maxblocksize = largerlimit

It can start being in versions way ahead, so by the time it reaches that block number and goes into effect, the older versions that don't have it are already obsolete.

When we're near the cutoff block number, I can put an alert to old versions to make sure they know they have to upgrade.

Also, while not conclusive at all, notice the hypothetical block number he used, and possibly why he didn't pull 1 million out of that hat.

The limit (in conjunction with transaction fees) were intended as a DOS attack and spam deterrent. Do you have any evidence that those risks have been mitigated? Do you have any evidence that suggests that 2MB blocks won't be filled to capacity right away? That would leave us where we are today -- with no scaling solutions, wondering how much more un-optimized throughput the network can safely handle.  

It's been somewhat mitigated, in the sense that back then, it cost exactly nothing to solve a block, or a trivial amount. Today it takes massive investment, located in massive farms, which are not incentivized to kill Bitcoin by DoSing it with malicious blocks, just for the lels. Blocks that the other miners might as well orphan(stale) anyway. The disagreement is about who determines what is spam. One group thinks it should be the miners deciding the block sizes they produce and the fees they require for admission. The other group thinks it should be an insular group of friends on irc, with a blatant COI.

I don't believe a second, compatible implementation of Bitcoin will ever be a good idea.  So much of the design depends on all nodes getting exactly identical results in lockstep that a second implementation would be a menace to the network.

Satoshi's description of how to increase the block size limit was also clearly in the context of a software update. Not an incompatible implementation of bitcoin that essentially attacks the network and intentionally forks from all other client nodes. He even thought compatible alternatives were a bad idea.

One could make the argument that Satoshi thought we could raise the limit as/when needed (as if we needed his authority anyway). In that respect, the question of necessity is subjective and debatable, and it is not immediately clear that an increase to 2MB now, with no attempts to make throughput more scalable is necessary.

One could not make the argument that Satoshi thought we should increase the block size limit through a contentious hard fork.

All of this is moot. We should be talking in terms of "what is best for bitcoin" -- with respect to users, nodes and miners -- not talking past each other with interpretations of things Satoshi said 5-6 years ago. Not only was Satoshi wrong about some things, but the state of bitcoin has changed a lot. Hell, if we left the codebase as he originally wrote it, those 184 billion bitcoins from the August 2010 value overflow incident would still be with us. Better to work towards bitcoin's principles than aimlessly trying to fit some arbitrary interpretation of Satoshi's words.

Which (possibly) makes you wonder if there is some reason it hasn't been released as a software update... Could there be anyone with their fingers on the central planning levers that might have a conflicted interest in a somewhat expensive main chain, via offering the medicine for that disease? Has an extremely well connected group taken $76 million of other peoples' money... to do exactly that?

[AliceWonderMiscreations]

The disagreement is about who determines what is spam.

Nodes can decide what they relay and do not relay, providing some level of spam control.

For transactions that make it to a miner, really only the miner should decide. It's their block.

Now maybe we could have consensus rules and I support consensus rules for a maximum TX rate but minimum, no, if a miner wants to include some cheap transactions it should be up to them.

That's my opinion.

The only reason I want rules for maximum is to protect users from bugs in clients (or mistakes) and prevent an easy mechanism of money laundering through TX fee.

[n0ne]

It is possible to grow larger than the visa network with the current user rate in accordance to the time of existence.

I would say the bitcoin network definetly could get bigger then the visa network,  and the time of existance is relativly small so it still has time to grow as well.
And if you look on blockchain.info and you see al the transactions already being made.
You realize this is really big already,  the amount of people all over the world that are using it,  we could form a Bitcoin country already!

No need to form such a country to establish our bitcoin. A lot of transactions are being used with blockchain but visa usage is high due to the universal acceptance. But if bitcoin gets universal acceptance, it will be controlled by someone which causes a backing in growth.