Inflation and the end of 50 BTC per block (from technical discussion)

[kiba]

I'm not sure why there would necessarily be price inflation given that the marginal increase in btc is relatively small and can in any case be forecast.

Sorry for not being clear. Before the bitcoin economy, I was talking about your proposed monetary system.

[hugolp]

The max block size is not imposed by fiat, but by agreement to a convention.  You could change that in your client right now and produce blocks just fine, but unless at least 50% of generators agree with you, whenever your client produces a block that defies this agreed convention it will be ignored by those who do.

By fiat, I don't mean to suggest that someone dictated it with an iron fist.  But I do mean to point out that it makes the currency dependent on a central authority.

No, its not dependent on a central authority. And fiat means imposed using force or thread of force. Please dont change the meaning of the definitions because it makes discussing things harder and tedious.

Sure, it is open source, but it is meant to benefit humankind, not just developers... over 99.9% of its users will have no clue how to participate in the management of this convention, which at this point appears to be designed to require it.

No. If someone wants to change it, you just need one developer changing the source and making it available in the Internet. Things are always going to be like this, people have different knowledge. The developers are not a unit, they are different people that can take different decisions. The software is open source. A group of people could even pay a developer to implement the changes they want. Anyone can make what they want with bitcoin. Bitcoin works by voluntary agreement.

Wouldn't it make sense to change the convention to something that will automatically cope with growth

I dont understand what you mean by "cope with growth" (growth of currency, growth of prices, growth of GDP, economic growth, ...?).

That said you wont be able to make a software that is 100% versatile for any change you can imagine. A software is always limited and actually its bad if you try to make it too versatile. My experience shows that it is better to keep a software simple and that solves a practical task, not some range of remote posibilities.

, much like the difficulty system adjusts to new computational pressure?  With a hard limit like this, the threshold at which it suddenly grinds reliable transaction processing to a halt at a reasonable price (until everyone upgrades their client at the direction of the developers) will, in all likelihood, be exactly at the moment where it sees a viral surge in new popularity that suddenly brings it past that point and to its knees.  That may bring it a lot of unnecessary questioning and criticism by the public and the media, particularly about just how truly viable and independent it is.

I have read repeatedly that the achilles heal of Bitcoin is that 50% CPU threshold.  I have never heard of another one I believe might exist - the possibility that the Bitcoin community gets very large, and then has factions, the block chain forks, and the system crumbles in the face of FUD while two or more camps argue which leg is correct (sort of like a country that started out with a Constitution or a religion that started out with a Bible or Koran and now has two large factions eternally asserting their particular interpretation of it).  Two religious factions can go their own way and live their own separate lives on separate lands, but if they must trade while they can't agree on what constitutes the existence of money, the money may as well not exist.  If Bitcoin is in its "constitution" phase, I would submit that the time to address something like this is now.

Competing currencies is very common and it does not have to undermine bitcoin. If bitcoin forks it will just strengthen the system. There will be exchange places where you will be able to buy or sell one bitcon currency for another if you need to, etc... In fact, I have said since the beginning that the its inevitable that different bitcoin chains appear. I think the division will be by physical territory, but who knows really.

[casascius]

I dont understand what you mean by "cope with growth" (growth of currency, growth of prices, growth of GDP, economic growth, ...?).

Growth of transaction volume specifically, and its corresponding burden on all the users in the form of the block chain.

However I find creighto's explanation satisfactory, that services like mybitcoin.com will be depended upon and available to take transaction volume off the main peer-to-peer network, and in a "de facto" sense, create a hierarchy that insulates every iPhone from needing to receive a transaction message each time anyone anywhere buys a soda with bitcoin.

[ribuck]

... I was impressed that (largely) one man created this system, assuming that Satoshi is actually a person, and not an avatar for some greater organization ...

A committee could never have made a design so well-thought-out, with so few rough edges. Also, committees don't format their white papers using LaTeX. Nor could a committee show so many signs of caring deeply about their project.

I have no doubt that Satoshi is one person.

[davout]

@creighto : you never back that "only a nation could overpower the network" statement. Let me throw some numbers in.

At current difficulty it takes 80 Gh/s to generate a block evey 10 minutes (source : bitcoin generation calculator)
Last 24h, the network generated 7.71 blocks/h (bitcoin watch).

That leads us to a network crunching roughly at (80*7.71/6) => 102.3 Gh/s which is roughly 171 Radeon 5970's crunching at full speed.

You can fit two of those in a PC, so if bought in bulk we can assume we could buy a PC with two Radeon 5970 for, let's say $2.5k

So you'd need $213,750 to match the networks power, round it to $250,000 to have a comfortable margin.

It's nothing remotely comparable to the resources a nation, or even a reasonably sized corporation has at hand.

[nanotube]

davout: well, you're a little undershooting it... at current diff, it's closer to 90ghps for 10 minutes. and the current 'next estimate' is based on the average of the generation speed in this chunk, which, if it is growing, is under the actual speed at the tip.

just fyi. your basic argument about the ballpark cost being relatively small is still valid.

[MoonShadow]

... I was impressed that (largely) one man created this system, assuming that Satoshi is actually a person, and not an avatar for some greater organization ...

A committee could never have made a design so well-thought-out, with so few rough edges.

Yes, it's downright elegant.

 Also, committees don't format their white papers using LaTeX. Nor could a committee show so many signs of caring deeply about their project.

I have no doubt that Satoshi is one person.

Well, one very well rounded genius level intellect would be the most likely explanation, I admit.  But I wasn't thinking of any kind of committee.  More like a tightly knit workgroup.

Yet, many open source projects are designed by a kind of committee, the 'bazaar', such a GNU/Linux OS's.  Of course, Linux prior to 2.x.x was far removed from elegant, and certainly couldn't have gotten there with only Linus doing the coding.  By his own admission, Linux in beta was ugly.

[MoonShadow]

@creighto : you never back that "only a nation could overpower the network" statement. Let me throw some numbers in.

Hmm.  I stand corrected. I'm not going to attempt to check your math.

So at what difficulty level would we be at the level that I thought?  What difficulty level will we be comparable to #10 on Wikipedia's top ten supercomputers list?

[MoonShadow]

I feel doubtful that a nation state is what it would take to control 50% of the CPU power, given plenty of disincentives for mass crowds to generate blocks.  Didn't someone recently say that an investment in $140,000 of video cards would be enough to presently overtake the 50% mark?

I don't know who might have said that, but they were wrong.

I said that, http://bitcointalk.org/index.php?topic=2227.msg29538#msg29538, I don't believe that I was wrong. It is true that bitcoin is under generated atm. 

However I do not expect that to remain true for long, it is just too profitable!

My apologies.

[casascius]

The way I understand it so far, comparison between computational power for Bitcoin versus the power of a supercomputer might be a little bit murky, because of the different kinds of workloads make it a little bit like comparing apples and oranges.

In other words, the question of "how many flops make a hash?" sounds difficult to answer, and may vary widely based on the intricacies of any given GPU...   would be like rating a bunch of muscle cars primarily on how quickly they can get you to a grocery store one block away.

Mining doesn't require the sophistication of floating point at all.  When a floating point powerhouse is busy mining, I suspect many of its transistors are busy doing absolutely nothing, since they perform functions irrelevant to SHA256.  (e.g. a GPU that can compute a mighty fast logarithm or square root probably won't do SHA256 a shred faster just because of it).

If I were a chief engineer looking to build a bitcoin "mining army" (and a large established and thriving Bitcoin economy justified the expense), I would look to build and mass-produce a special-purpose chip especially optimized for SHA256 hashing, much like the MPEG codec chips now commonplace.

There would be no limit to how big of a hashing "supercomputer" I could build, because brute force hashing is so easily parallelized and scaled, there's no engineering difficulty on how to make all the chips interact with one another, because they never have to.  The maximum power of such a supercomputer would be limited merely by the quantity of raw materials available to crank out copies of that chip.  Naturally, it would be worthless for anything but hashing.

My guess is that SHA256 is so simple to implement in silicon that it's within the grasp of undergraduate-level engineers.  It is orders of magnitude less complex than MPEG.

Such a chip would blow GPU's away on two fronts - one on the hash count, second on the cost, because all the wasted resources and engineering that normally goes into a supercomputer useful for solving really complex scientific problems (or a GPU to solve problems necessary for 3d graphic rendering) could be entirely skipped. 

[jgarzik]

My guess is that SHA256 is so simple to implement in silicon that it's within the grasp of undergraduate-level engineers.  It is orders of magnitude less complex than MPEG.

Well not "so simple" but basically correct.  There's even a LGPL'd core you can download already: http://opencores.org/project,sha_core

Such a chip would blow GPU's away on two fronts - one on the hash count, second on the cost, because all the wasted resources and engineering that normally goes into a supercomputer useful for solving really complex scientific problems (or a GPU to solve problems necessary for 3d graphic rendering) could be entirely skipped. 

Up to a point, buying a bunch of mass-market GPU's with ALUs is quite comparable to the up-front costs of even a small run of ASICs.  FPGAs won't get you nearly the speed needed to keep up with modern GPUs.

[FreeMoney]

There would be no limit to how big of a hashing "supercomputer" I could build, because brute force hashing is so easily parallelized and scaled, there's no engineering difficulty on how to make all the chips interact with one another, because they never have to.  The maximum power of such a supercomputer would be limited merely by the quantity of raw materials available to crank out copies of that chip.  Naturally, it would be worthless for anything but hashing.

My guess is that SHA256 is so simple to implement in silicon that it's within the grasp of undergraduate-level engineers.  It is orders of magnitude less complex than MPEG.

Such a chip would blow GPU's away on two fronts - one on the hash count, second on the cost, because all the wasted resources and engineering that normally goes into a supercomputer useful for solving really complex scientific problems (or a GPU to solve problems necessary for 3d graphic rendering) could be entirely skipped. 

Interesting. This seems to imply that mass hashing has never before been needed or else this would already be in production. Was the "hash over and over until you get below the target" idea really a new creation and not borrowed form somewhere?

[jgarzik]

Interesting. This seems to imply that mass hashing has never before been needed or else this would already be in production. Was the "hash over and over until you get below the target" idea really a new creation and not borrowed form somewhere?

Hashing mostly-the-same data twice, over and over again, is not remotely a common activity

[acrylicist]

Let me recalibrate my understanding.

Every computer will be essentially generating bitcoin. Essentially, it is hyperinflation. However, this distort economic calculation to a really bad level. Everything you invest in looks like a profit, when it's actually not.

There is still no incentive to save, because your saving is becoming worthless every single day due to inflation. Basically, it rewards the biggest generator at the expense of everybody else, except those who borrows money.

In the bitcoin economy, the generators still benefit disapportionately, but it will also benefit those who save. In other words, everybody who save win, except those who borrows money.

Investors in business will be forced to be more conservative or efficient at finding economic opportunity.

There won't be hyperinflation with BitCoin because of the algorithms--the faster BitCoins are generated, the harder they become to generate (and thus require ever more amounts of electricity or hardware). My issues with BitCoin come from the unfairness of money creation, which is essentially a technocratic central bank issuing currency. If you have enough CPU-power and energy, you can originate most significant amount of new BTC into the market. While the rest either hoard it because of its deflation or spend it occasionally because of "time preference."  As people deflate BTC into smaller and smaller units, in absolute terms, everyone is playing with "more and more units of currency," the only difference being the zeros are being added on the opposite side of the decimal point. With all of the hype and attention for BitCoin in recent press, it could be heading towards hyperdeflation, where "People postpone buying indefinitely because they expect prices to fall further."  In BitCoin, there will be no incentive to spend, which is kind of ironic and self-defeating for a proposed new money.   

If I had 50BTC and thought that it would deflate incredibly, I'd be waiting until 0.00005BTC would pay my rent before I'd use it, I'd have a million "units" to play with.

[kiba]

My issues with BitCoin come from the unfairness of money creation, which is essentially a technocratic central bank issuing currency. If you have enough CPU-power and energy, you can originate most significant amount of new BTC into the market.

Technocratic central bank issuing currency? WTF?

You buy CPU cycles in the hope that you make more money than you would in loss. It's not technocratic but based on economic calculus. Bitcoin doesn't exactly grow on trees. That is the fairest way we can think of distributing the currency.

The experts aren't issuing the currency, bitcoin speculators who sunk tons of money into their mining rigs are.

[jgarzik]

My issues with BitCoin come from the unfairness of money creation, which is essentially a technocratic central bank issuing currency. If you have enough CPU-power and energy, you can originate most significant amount of new BTC into the market.

Technocratic central bank issuing currency? WTF?

acrylicist is being hyperbolic, but he's right.

bitcoin is a software-fiat currency, where the "central bank" is a collective agreement on software rules.

USD is a human-fiat currency, with a traditional central bank.

And it has always been true that the more CPU power you have, the more BTC you will receive.

[kiba]

bitcoin is a software-fiat currency, where the "central bank" is a collective agreement on software rules.

I'll concede that based on the fact that central banks are responsible for the control of the money supply. However, Satoshi and the core developers have no power beyond the agreement of all the miners. They are checked by bitcoiners in general and the right to fork. They do not control or regulate the Bitcoin banking industry, amongst other things.

Certainly, they have less power compared to a nation-state's central bank.

And it has always been true that the more CPU power you have, the more BTC you will receive.

Nobody is disagreeing with that fact.

[ribuck]

And it has always been true that the more CPU power you have, the more BTC you will receive.

You say that as if everyone's CPU power is pre-ordained.

The system is "fair", in that each person has the choice of how much CPU power they wish to devote to bitcoin generation.

There's no enforced cartel, no generating license needed, no official that needs to be bribed, no extra taxes for generators in some states, no need to "know the right people", no pre-booking needed. It's about as fair as anything could possibly be.

[jgarzik]

And it has always been true that the more CPU power you have, the more BTC you will receive.

You say that as if everyone's CPU power is pre-ordained.

No, I say that as if nothing.

It is a simple functional statement of fact.  It is a basic assumption of the bitcoin system.  If you contribute more CPU power, you are highly likely to receive greater rewards.  That is the way bitcoin is supposed to work.  That is the incentive for mining (and for scaling up your mining operation).

[ribuck]

It is a simple functional statement of fact.

OK, I see the misunderstanding. I thought you meant "have" as in the amount of CPU power that you own or control, but it makes sense if "have" means the amount of CPU power that you devote to mining.