Three Pools Have Near-total Control Over the Bitcoin Network

[Mike Hearn]

That's a rather huge over-generalization Erik.

Pooled mining is an activity with a very low barrier to entry. You need some software, some hardware and some kind of justification for your existence, and that's pretty much it. Some seed hash-power to get it started. Markets work well when barriers to entry are low and switching costs are low.

They work much less well when barriers to entry and switching costs are high. That's why there are only three CPU designers responsible for nearly all consumer CPUs, and of those three two of them (AMD and Intel) produce chips that run near-identical instruction sets. And these days it's really turned into a two horse race with Intel and ARM almost entirely controlling the high performance and low power markets respectively.

How easily monopolies form depends completely on the specific market.

Also whilst the situation now is better than it was, the previous situation was a complete perversion of Satoshis original vision. The point of mining is to vote, and what effectively happened was massive selling of votes in order to reap more predictable profits. The more miners the better as it makes it much harder for the rules to be changed out from underneath people (eg, a change to the inflation rate).

P2Pool should have fixed this but hasn't because it seems many miners would rather have simpler configuration than actually be first-class voters.

[hazek]

That's a rather huge over-generalization Erik.

Pooled mining is an activity with a very low barrier to entry. You need some software, some hardware and some kind of justification for your existence, and that's pretty much it. Some seed hash-power to get it started. Markets work well when barriers to entry are low and switching costs are low.

They work much less well when barriers to entry and switching costs are high. That's why there are only three CPU designers responsible for nearly all consumer CPUs, and of those three two of them (AMD and Intel) produce chips that run near-identical instruction sets. And these days it's really turned into a two horse race with Intel and ARM almost entirely controlling the high performance and low power markets respectively.

How easily monopolies form depends completely on the specific market.

Also whilst the situation now is better than it was, the previous situation was a complete perversion of Satoshis original vision. The point of mining is to vote, and what effectively happened was massive selling of votes in order to reap more predictable profits. The more miners the better as it makes it much harder for the rules to be changed out from underneath people (eg, a change to the inflation rate).

P2Pool should have fixed this but hasn't because it seems many miners would rather have simpler configuration than actually be first-class voters.

You couldn't be more wrong with both of your points.

First, your analogy with the CPU designers in no way applies because it does not exist in a free market such as the one Erik was referring to. No, that industry exists in the by the state controlled market where the state by using the threat of violence prevents competition from developing the same technology(enforcement of IP through the instrument of patents), forces the would competition to comply with expensive regulation and steals a good chunk of their profits. That is not a free market but a market with a central authority that promotes and defends monopolies which is why the cost of entry are so high!

In bitcoin where we actually have a market regulated strictly by market participants i.e. a free market there are no such costs. You just need to have the appropriate capital and you can begin. No mining pool is demanding their competition cease to use some part of their code because they thought of it first or that some portion of their profit be paid to some gang of thugs with guns. People are free to do as they please as long as the market i.e. the consumers want their goods or services.


Second you are oblivious to what mining is suppose to be and how mining pools work.

No one is forced to participate, it almost does not matter if miners mine individually or if they participate in a mining pool because they can always decide to stop participating. As soon as one of the pools turned out to in anyways misbehave they'd have the option to stop and leave. This is perfectly inline with Satoshi's whitepaper.

But also mining is not about voting because there is nothing to be voted on. Bitcoin is a protocol, you either use it or you don't. If miners stop using it by enforcing a different protocol with the same blockchain they wouldn't change Bitcoin, no they'd simply create a hard fork and begin using the new protocol. Anyone who decided to use the old protocol could still do so and they'd be the only one who could legitimately claim they are using Bitcoin.

[d'aniel]

...
This is perfectly inline with Satoshi's whitepaper.

But also mining is not about voting because there is nothing to be voted on.
...

The proof-of-work also solves the problem of determining representation in majority decision
making. If the majority were based on one-IP-address-one-vote, it could be subverted by anyone
able to allocate many IPs. Proof-of-work is essentially one-CPU-one-vote. The majority
decision is represented by the longest chain, which has the greatest proof-of-work effort invested
in it.

Couldn't resist doing that after you'd stated one of the lead client developers is

oblivious to what mining is suppose to be and how mining pools work.

[Gabi]

inform and promote p2pool

Tried it - rejects like crazy, much lower revenue in my 1-day trial.

Please note that a 7% like stale rate is fine on p2pool because p2pool stales are DIFFERENT from normal stales and are not useless.

[hazek]

...
This is perfectly inline with Satoshi's whitepaper.

But also mining is not about voting because there is nothing to be voted on.
...

The proof-of-work also solves the problem of determining representation in majority decision
making. If the majority were based on one-IP-address-one-vote, it could be subverted by anyone
able to allocate many IPs. Proof-of-work is essentially one-CPU-one-vote. The majority
decision is represented by the longest chain, which has the greatest proof-of-work effort invested
in it.

Couldn't resist doing that after you'd stated one of the lead client developers is

oblivious to what mining is suppose to be and how mining pools work.

Nice try but no.

First when you quote you should quote the entire paragraph making it clear what it is that you are quoting:

The proof-of-work also solves the problem of determining representation in majority decision
making. If the majority were based on one-IP-address-one-vote, it could be subverted by anyone
able to allocate many IPs. Proof-of-work is essentially one-CPU-one-vote. The majority
decision is represented by the longest chain, which has the greatest proof-of-work effort invested
in it. If a majority of CPU power is controlled by honest nodes, the honest chain will grow the
fastest and outpace any competing chains. To modify a past block, an attacker would have to
redo the proof-of-work of the block and all blocks after it and then catch up with and surpass the
work of the honest nodes.

What Satoshi refers to as voting here is actually enforcement of the rules in Bitcoin on the blockchain not on the protocol. It's a poor choice of an analogy on his part but proof of work is not voting, it's validation & enforcement in a distributed manner. It's how the network comes to a CONSENSUS about which blockchain is accurate.

Now you can call this voting and it sorta kinda resembles voting but it's not voting. In voting you have options with new rules and which ever option gets the most votes gets imposed on all the voters. In proof of work you have validation and enforcement of rules that are not subject to change or up to a vote, no, in proof of work nothing is decided for all, nothing gets imposed on all, the network merely discovers which version of the blockchain most members of the network have validated and enforced the rules on in order to discover a consensus. Per definition this consensus is not imposed on anyone after it's discovered.

Comparing proof of work to voting may have only served as an analogy to better explain the process but is not actually what is happening and therefor calling it voting is invalid. Yes, I actually am saying that even Satoshi can make a mistake. No man is infallible.

So I was correct. There is nothing to be voted on.

But also mining pools are perfectly inline with Satoshi's idea about how the network comes to a consensus about which version of the blockchain is accurate because all that mining pools do is they pool the validation and enforcement to a small chunk of the network first. What is important here is that this small chunk of the network is still validating the same blockchain and they are enforcing the same rules as they would if they were solo mining. Therefor as far as the proof of work concept is concerned the outcome is EXACTLY the same if miners participating in a pool would solo mine instead.

Btw using a developer disagreeing with me as proof of anything is a logical fallacy.

[Mike Hearn]

The reason you can't set up a company that competes with Intel is not due to patents. When was the last time you heard of a highly competitive CPU firm ending up in an Apple-v-Samsung style showdown? It doesn't happen.

In fact, AMD implements exactly the same sort of designs that Intel uses and Intel implements the AMD64 instruction set extensions. So they are both copying each others designs and there are no issues.

Do patents add to the cost? Quite likely, but it's not the sole reason the desktop/laptop CPU market has tended towards duopoly.

The actual reason it's hard to compete with these firms is because building a fab capable of producing chips competitive with Intels costs around $8 billion dollars. And that's assuming you have access to all the expertise, equipment and materials supply contracts you need. From cold start it'd certainly cost far more.

Mining absolutely is voting. As was already pointed out, it explicitly says that in Satoshis paper. You are casting a vote between multiple options, namely the heads of the chain. The outcome of these votes can determine not only which transaction wins in the case of a double spend, but also which types of transactions get included at all (IsStandard checks).

Also, not all users can validate all the rules of Bitcoin. Anyone using an SPV client can't and that will be most users in the coming years. If the majority of miners choose to vote on a chain that changes the inflation formula, SPV users will switch to that chain and accept payments based on it. Yet the inflation formula is one of the most fundamental and important rules of the system. So it's important that there be lots of voters to make it hard to set up a miner conspiracy to change the system rules. We currently lack that.

[hazek]

Please, take a look at: http://en.wikipedia.org/wiki/List_of_top_United_States_patent_recipients and tell me who for the year 2010 and 2009 is in the 8th spot of all US based companies ranked by who received the most patents? You think if I wanted to start a CPU factory I could just use their design without any costs? Not to mention all the regulatory costs and the tax burden the state lays on top.

Yes capital needed to produce CPUs is far more expensive than getting into mining but saying it's hard to get into producing CPUs because of that is plain idiotic because it ignores all the extra costs such a company would have to swallow and NOT push onto their customers if they wanted to be competitive compared to a by the state endorsed already entrenched monopoly.


Mining absolutely is not voting. You are seeing hoofs and you're thinking horses when in fact it's zebras. Just because it resembles voting it does not make it voting as I have demonstrated with an argument you didn't even attempt to dispute. Proof of work is validation and enforcement != voting. It's how the network comes to a consensus i.e. an agreement about which version of the blockchain is accurate not an election result that determines which version of the blockchain is forced on everyone. You can call this voting as much as you can call yellow the color yellow on your screen.

Iron clad argument why it's not voting: If it really was voting hard forks would not be possible because there would be no deviation from the election result.

Also all users do in fact validate the rules because even their light clients still need to know whether the transactions that they read from a remote server with the blockchain are legitimate or not or whether the transaction they are about to send out is going to be regarded as valid by the rest of the network, not to mention that some light clients like multibit or blockchain.info are open source. Of course how easy it would be to switch the light client's encoded rules without the user having to OK it is another question that mainly depends on the client's design but there is no question they have a say.


I can't understand how someone working with this software intimately wouldn't have a firmer grasp on what it is that they are working on.

[istar]

None the less its an important subject.

The original client should have a miner in it. That was how it was designed from start.

If Bitcoin reaches a 100 million users who all have the original client and let it mine in the background.
That would make Bitcoin way much much safer? Even if it was only 10% that would be 10% safer and thus 10% more value for free.
Because a huge part of the real value of Bitcoin is the cost of what it takes to own or brind down the network.

Problem is this also mean more competition for the big miners.
I know that some of the people working on the client are big miners.

I see a conflict of interest in that if its not easy for avarage people to mine. And there are certainly lots of gamers with new graphic cards who would like to mine if it was easy. But cant be bothered otherwise. Only a few percent can be bothered if its much trouble.

This gives them a greater piece of the cake.
Especially now in "early" times.

Its a matter of not only democracy/decentralization but also safety.

The vision was that anyone should be able to get some Bitcoins, by turning their computers on and having the client running in the background if the wanted, to help the network and for this sometimes get a reward.

Is now gone. Its no longer easy for an normal computer user to start to mine.
Its not in the best interest of Bitcoin, but in the best interest of a few single individuals.
How come?

In the long run, those percents really counts.
Especially if a goverment or private person where to attack the network by for example hacking some of the big mining sites.

An easy way to mine should be included in the mainstream client ASAP.

[kjj]

An easy way to mine is included in the default client.  Check out the RPC calls "getgenerate" and "setgenerate".