Topic: LN+segwit vs big blocks, levels of centralization. - page 6. (Read 8897 times)

We agree, but why did you mention then:


How could one interpret these "hard hitting instant, free payments" as suddenly increasing the velocity of money by many orders of magnitude (the same orders of magnitude that would bring bitcoin from a one-pizza economy to a world economy, say) otherwise than you thinking somehow that the speed of the transaction itself plays an important role??

What's so special about paying with your bitcoin salary versus paying with your fiat salary, that suddenly, people don't keep their bitcoins any longer than half a millisecond ?

How are you going to offset the absence of debasement, essentially and the growing by many orders of magnitude of the economy bought with the currency, just with the velocity ?

As this is not realistic, even to someone who clearly shows not to be an economist, the only conclusion can be that bitcoin is very severely deflationary, and hence, will plunge severely in a deflationary spiral, which it is doing right in front of your nose since it came out in fact.

Again, you keep ignoring how he literally says everyone should stick to the official version and how he doesn't believe in a second, compatible implementation of bitcoin because it will never be a good idea since 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. He is basically calling Bitcoin XT, Bitcoin Classic, BUcoin, and whatever isn't the satoshi client a menace to the network (and he obviously means this literally, his argument doesn't stop magically being true if someone else is developing it, *whoever is developing it, they are developing a different client so the point of all nodes being identical in order to not be considered a menace to the network still prevails* so don't try to twist what's obvious to meet your agenda)

Why do you keep ignoring facts?


No, because it's not the official satoshi client, and I trust Bitcoin Core to continue doing a good job in keeping trojan horses away.

If the satoshi client ever officially becomes a mess, then im dumping all of my BTC, but until then, bitcoin is showing how strong it to reject all this bullshit.

I can't possibly fathom how my words could be misinterpreted so severely

Obviously, I'm talking about the circulation velocity from an economic point of view which as I think I made pretty clear when I said that the increase in the velocity of money has the same effect as the increase in monetary base itself. And that could offset the deflationary factors up to a point where Bitcoin becomes inflationary (i.e. its price will go down even if no more coins get mined). I don't really know how that could have been construed in the way you did. As I have also noticed, you distort or misinterpret the arguments of your opponents to better fit your ideas or dogmas

Well, you were arguing that the velocity of bitcoin would go drastically up, to compensate its evolution from Q = 1 pizza to Q = world economy, and keep its value constant, not rising, because the LN network allows almost instantaneous payments.  The "time of a treatment of a payment" is hence almost instantaneously while the "time holding bitcoin on chain" is at least 10 minutes (one block period).  I thought that your argument was something like "we go from payments that take 10 minutes to payments that can take 20 milliseconds, so there is your velocity increase needed to undo deflation".

I didn't understand what you meant to say


Could you explain what exactly I'm confusing here?

Just in case, I meant specifically the velocity of money, i.e. "the rate at which money is exchanged from one transaction to another and how much a unit of currency is used in a given period of time", as per Wikipedia. In simple terms, the velocity of money shows how often a certain bill or other currency unit changes hands in a specific period of time (on average, obviously). Apart from that, when someone starts using convoluted terms and notions like "the velocity of the *treatment* of a payment", you can be pretty sure that he doesn't know what he is talking about. I don't know what you meant by this, either

No, the velocity does matter.  But let's first get the terms right.

Inflation/deflation has to do with the value of a monetary unit (say, the price of a bitcoin).  If the price of bitcoin rises, there is deflation, if the price of bitcoin falls, there is inflation.

Debasement has to do with the creation of monetary units (the increase of the monetary mass).

Fisher's formula, which is valid for any *currency*, goes as follows:

M * V = Q * P

Here, M is the monetary mass (the amount of coins)
V is the velocity (the number of times on average, a single coin is re-spend per year)
Q is the total economic value of the economy exchanged for the currency (in a "value unit" that is constant, say "a Big Mac") in one year
P is the price of that value unit (the number of coins you need to buy one Big Mac), which is essentially the inverse of the coin value.

As such, we have:

1/P = coin value = Q / (M * V)

In order to have no inflation or deflation, the monetary mass times the velocity needs to increase (debase) proportional to the increase in the amount of economic value that is bought with it (the increase in adoption).

What deisik says is theoretically right: one could start with Q = one pizza, and an ultra-low velocity (a coin is spent once in a year), and by the time that Q becomes the world economy, you never hold a bitcoin longer than half a milli second, and spend it immediately, and keep the value of a bitcoin constant.
But people don't spend their money faster because they have the technological ability to spend it !

In reality, if one wants the coin value to remain about constant, one has to augment the monetary mass (print money) with the increase of the economy it buys, and the velocity is a matter of "economic mood", but is not a very changing quantity by orders of magnitude.

It is the job of a central bank to try to keep M about in pace with the economic growth, but most central banks debase somewhat more and create a light inflation (diminishing of the coin value).  For instance, the European central bank has as a target, 2% inflation, and will print in principle enough money for this to happen ; however, the recent times, banks were not willing to accept so much new money, so the ECB doesn't achieve an inflation rate of 2%.

One can recognize weak arguments when they start with rhetoric statements about the intrinsic qualities of the counterparty


This is correct in principle, but most real economists know that this is a variable that doesn't adapt a lot, and you are confusing the velocity of the *treatment* of a payment with the monetary velocity, which is the harmonic average of people HOLDING the monetary asset between earning and spending.  In fact, if bitcoin were mainly used as a currency, then this velocity would most probably LOWER with increased adoption, still increasing the deflationary effect.  Let me explain why.

If bitcoin were only used as a currency (to buy stuff with), and not as a speculative HODL/greater fool asset, then the holding times that determine the monetary velocity would be essentially between acquiring coins on an exchange, and spending them on an item, because the economy is fiat-dominated, and there's no point holding bitcoin a long time.  The merchant would, most probably, also almost instantaneously convert his received coins into fiat.  This would imply a very high monetary velocity, because the holding times by the customer and the merchant are small.  There is indeed no reason to buy bitcoins on an exchange, long in advance of your buying occasionally something, and the merchant has no reason to keep bitcoins, and not convert them quickly to fiat.  You can buffer somewhat, to diminish the cost of fees by bulk buying and selling, but essentially, you only buy and sell them when you need them.

However, if ever bitcoin were to become a more generalized means of payment, then economic cycles would close more and more, and there would be less incentive to convert from and to fiat.  As such, the velocity of bitcoin would tend to the normal velocity of fiat, between earning and spending, because you would receive your salary in bitcoin, hold it in bitcoin, until you spend it.

The velocity would hence, the more economic cycles close, slow down from almost instantaneously to "normal", which is a small number of rounds per year.

Nope. Inflation is caused by increasing the money stock/supply, as in creating money out of thin air - fiat money. Velocity is just a guide of economic activity and does not cause inflation.

A buys ? for £10 from B, who then buys ? for £10 from C. Velocity is 2. C buys ? for £10 from D = velocity is 3 and so on. The higher the velocity = higher transactions made. Nothing to do with inflation.

It is obvious that you are not an economist

Since an economist (a real one, not some fresh graduate from a second grade university) would always keep in mind an "opposing force" which could potentially offset the deflationary nature of Bitcoin (at least, to a certain degree), and you'd better not argue with him in such matters. I'm talking about myself since I'm talking about circulation velocity. Increasing it is effectively equal to increasing a monetary base, i.e. how many coins are circulating in the economy. So even with a fixed supply (which Bitcoin has), you can get an inflationary currency in the end. To achieve that, you would just need the increase in the velocity of circulation (let's call it circulation acceleration) to exceed the expansion of the coin. In other words, inflationary forces should overwhelm deflationary ones, and you will get inflation as a net result. And this is where LN and instant, free of charge payments are intended to hit hardest

Bitcoin's built-in deflationary spiral, on which it is very proud, plus other more technical aspects, make that this stuff will not be a "unit of account" ever, because of the inherent instability of its price.  I agree with you that some initial speculative "push up from zero" was necessary to transform a token-transmitting system into a monetary belief system, but the total absence (on purpose built-in) to limit its value makes that this is going to be preferentially kept and traded, rather than used to buy groceries.  The emission model got it totally backwards if it were to be a currency ; but is perfectly right if it was to stimulate a bubbling speculative asset, to make early adopters rich on the back of the final layer of greater-fool adopters: namely a lot of emission when the thing is cheap, creating huge seigniorage, and when adoption kicks in, reducing emission and increasing rarity.  A stable currency would have exactly the opposite emission scheme, and was in fact very easily done, by NOT having automatic difficulty adjustments to limit the emission scheme, but rather by having slowly increasing, preprogrammed difficulty increase to keep the emission rate on par with the adoption, keeping the value almost constant with an adapting monetary mass.    But that wouldn't have given the dreams to become a millionaire to "adopters".   Which is the very proof that it is a speculative asset and not a currency.

If you'd need the currency - that is, if there was a true market demand for a means of payment - then you'd prefer the currency NOT to rise in value, because you are going to SPEND it and you don't want to have regrets of spending it instead of holding it.  The naked truth is that apart from some niches like dark markets, there is no market demand for a new means of payment, and bitcoin is not filling it.  Yes, it is used somewhat as a geeky gimmick "hey, I  paid my airplane ticket in bitcoin!" ; it is used for niche applications such as dark markets, where there was a strong demand for a means of payment ; and there is some usage as remittance in those places where banking doesn't work well.  But it is not its main application, and it doesn't get much traction there.  People citing bitpay show impressive numbers of *monthly* volume that is smaller than *daily* speculative volume on exchanges.

By far most people buying bitcoin do it because they want to be millionaires.   They don't buy it to use it as money.  Some do, but it is very little.  Because in most cases, credit cards and banking works better.

It doesn't matter whether bitcoin was said to be invented as a "currency on the internet".  It simply isn't mainly used for that, and its economic dynamics makes that logical, because it is designed as a hugely deflationary asset.  It only needs the story that it is "money" in order to give it some illusion of backing by economic utility.

The reason why bitcoin is not used a lot as a currency is in my opinion NOT technical (even though the technical side doesn't help).  It is because for most payments, there is simply no demand for a better payment system that is at the same time practical, legally protected, relatively cheap to use, and "no hassle".  It is only where fiat cannot go (as a payment system) that crypto has some interest as a payment system, and then you take the hassle, the legal insecurity and everything with it.
So the first reason is that apart from being geeky, there most of the time no *need* to have another payment system.

The second reason is that bitcoin's economic model is absolutely not adapted to be a stable unit of account ; on the contrary, it is going to be a high-volatility deflationary asset, with a lot of seigniorage in the hands of early adopters.

The third reason is that it is actually more involved, with more intermediaries, to use bitcoin as money, than to use fiat, because in practice, you *have* to use your bank, to make a wire transfer to an exchange, to buy your coins, to withdraw them, and finally, to be able to make a payment.  This can be worth it, if there's a reason.  But if there's no reason, then it is much easier, safer, and legally more protected, to do a direct payment from your bank.

Of course, the promise, the illusion of selling bitcoin as a currency, is simply by saying that the fiat system will crumble.  If you need the competition to kill itself before you think there will be a mainstream demand, it means you have essentially no market.  Apart from niche applications (mainly dark markets).

As to me, you are putting the cart before the horse

Indeed, in its early days (and maybe till very recent), Bitcoin was exactly that, i.e. "a vehicle for speculation" since otherwise (as a currency) it wouldn't be able to get any noticeable traction at all. In other words, if we had instant payments with low or no fees right from the start, everything would be essentially the same, till now. Personally, I'm heavily inclined to think that we are already past that point (Bitcoin only as a "speculative asset for the financial markets"). In other other words, the lack of instant payments and high fees have already started to affect the currency aspect of Bitcoin in a substantial way, and it could expand more specifically as a currency if these hurdles got removed somehow

Actually, I don't agree with that.  Scalling on-chain for VISA like levels of transactions is perfectly possible with the current actual decentralization level of bitcoin: only the mining pools need to have full nodes, and they have enough means to set up large data centres mastering such nodes, exactly as Satoshi considered it.  Some big exchanges might also invest in big nodes.

It is only if we want to maintain the illusion that non-mining nodes contribute a lot to the decentralisation of the power in bitcoin, that we cannot consider that, because one needs to keep the feeling intact that "Joe, with his node in his basement, is contributing to the network".  That's somewhat similar to the space shuttle, that had a funny procedure for the astronauts to land the vehicle: one had to switch to manual mode just before touch down.  There was no technical need for that, but at least, they had the illusion to be anything else but living cargo that way

In fact, even your motivated Joe could have a full node, but he will have to invest in it, with premium network service, and some data infrastructure.  

There's no escaping that increasing the size of a decentralized network increases the burden on every user that want to be decentralized and "check it for himself".  The real pain is that at a certain point, this burden becomes so important, and the utility of it becomes so low from the point of view of the user, that he prefers delegating his "trustless" trust to a trusted party, who gets power by that delegation.  In the end, you get banks.  Because that's what the market found as the optimal compromise between the cost of trustlessness, and the price of handing over power (and hence, corruption).

I guess what these scaling wars are about, are about the flipping point on the size of the network where the price of a decentralized network becomes too costly for the participant.  But actually, the technical cost is not the leading one.  The economic structure of bitcoin, with PoW, has already centralized the system far more than the technical burden.  So, why not accept this centralization, instead of wanting to maintain the illusion of a decentralized network which isn't, since quite a while, any more, and use that fact to reduce scaling costs, by delegating the burden to those entities that are already the central powers of the system ?

Also, the other side is the question whether a thing like bitcoin needs VISA like scaling one day.  In other words, is bitcoin really going to be a day-to-day currency, or is it more like a speculative asset for financial markets ?  Here too, all evidence points to the fact that crypto is mainly a speculative asset, and that its "currency" use is very minor.  Again, we are on the "image building" side of the illusion: bitcoin's illusion is "money for the masses".  If bitcoin were advertised as "speculative asset for the financial markets", it wouldn't get Average's Joe's enthusiasm "yet another financial toy people with too much money invented".  So to keep the illusion intact that bitcoin is, one day, going to replace fiat, the illusion that one day, it can replace VISA, is needed.  But there's absolutely no indication that bitcoin is going to do that and that you are going to pay your groceries in bitcoin.  People are "investing" in bitcoin, are "hodling" bitcoin, are day trading bitcoin.  And yes, a few people use it also to buy stuff.  But its main application is speculative finance.  These things don't need VISA style volumes, especially because trading needs to be fast, and is done on clearing houses (exchanges) with IOU.

Maybe control is the wrong word to use it when we talk about the developers. But in the history of Bitcoin there has always been one team that is being followed by the majority.

You are right, I am not a software engineer. As for the rest, let's then talk about facts only: as things stand today, scaling on-chain for Visa-like levels is indeed impossible.


Exactly, so it's not a big block, otherwise you and many others would explicitly support some kind of blocksize increase at some point. And yes, separating the witness does indeed allow to transact more, that much I could gather


1) I didn't say you implied that.
2) I was hoping a nonsensical exaggeration would get my point across, apparently not. Let's make those 10GB, 10MB instead. But I guess I made myself clear regardless.
3) At what cost? Attackers have been known to spam the network with a fair amount of funds so far and gave up without achieving a blockchain stall...



I guess this will probably be my last post here. Been reading the thread and it's been a bit derailed...

Your post wasn't that long, so I had no difficulty reading it

But I didn't pay particular attention to that sentence since I was intrigued what you could probably mean by quantum leap (in fact, exactly this phrase allowed me to recall that I did actually read this post in its entirety) since there is (was) a movie series by the same name. Anyway, you still made this statement conditional so it was obvious that you didn't link "a conservative increase" of the block size with the threat of centralization (which wouldn't go away). Otherwise, you would have just stated that we need "second-layer solutions" without specifying the conditions when we might not need them (thereby revealing that you don't care about the threat of mining centralization not going away). That's probably another reason why I ignored that part as inconsequential

I fully agree with that.  Imagine Tim Berners-Lee having said that in the interest of the user experience, there should only one single web browser.

http://info.cern.ch/hypertext/WWW/TheProject.html

But with bitcoin, moreover, if there's only one single software implementation, the distinction between software and the protocol is blurred.
If there's only one web browser, then of course that browser defines also HTML and HTTP.

By "validating all information" I do not necessarily mean that every individual (child) transaction needs to be validated: like with LN, and like with micropayments, there can be "a single transaction corresponding to a lot of transactions".   Sidechains are similar, but only up to a point.  Even though it is true that sidechains "mimick as a single transaction" on the main chain, and "represent a lot of transactions" on the side chain, and as a main chain user, you only need to see the single transaction, there is nevertheless the danger of an equivalent of "fractional reserve banking" on the side chain.  Now, in as much as these side chain tokens are not "fungible", there's no problem, but then we are back again to "several crypto tokens and a dynamical exchange".  However, in as much as these side chain tokens mimick as "original tokens", and they have a higher number as compared to the "locked coins", there's a danger of artificial debasement (of course, "when the side chain will settle" this will be done with, but that's exactly the same in FRB).

However, from the moment you get "cross payments" between these side chains, there's no choice but to verify everything, apart from those "books that have been closed again".  In a way, you could think of all these compartments as potentially independent and parallel block chains, but if you get random payments from one to the other, the user needs to check both chains and the crossing protocol.  In as much as you get a web of interacting payments, I don't see how you do not end up checking about everything.  Apart from those sidechains that *opened and closed again* and where you can forget about what happened on them, in the same way that you can forget about all the micropayments once the channel settles, and you can forget about any LN, once all the channels have settled, and only the settling transactions matter.

My statement only applies in the following sense: *for a given technology of processing transactions*, in a decentralized network, the burden per independent user increases with the size of the network.

The idea of "building a tree" doesn't seem compatible, to me, with the idea that transactions are a random mesh, that is to say, that just any user can transact with just any other user.  It only makes sense if there are "user communities" that transact a lot amongst themselves, and only rarely "cross borders".  Maybe there are smarter technologies, and I don't have a clear view on them.


This is absolutely, totally correct.  But it is only a security/privacy thing *to its owner*.  This is a bit like "running a Tor relay has the advantage of being able to use Tor without explicitly connecting to a Tor relay".  In fact, the only way to use, in a totally decentralized way, a crypto currency, is to be a full node, because it is the only way you can check for yourself.  All the rest is a matter of delegated trust.  But in fact, then one should also "write one's own node software implementing the protocol" because you cannot trust blindly the authors of the software.  At some point, you need to delegate trust.

im not appealing to satoshi. i can atleast look passed the literal and see the contextual
EG
"a second version would be a massive developmental and maintenance hassle for me"
well thats where he started realising he was being relied upon too much to develop and maintain it single handedly

things have moved on. we should not see a dictator with a single source of code as a good thing because it makes things easier for them to change code.. but as a bad thing because it makes things easier for them to change..

again for all them blockstreamers that REKT hearne.. ask yourself
if hearne had the only implementation that would work on the network, would you support it

diverse decentralised nodes are a good thing for security.

---

just look at the real loons
billy wanting one brand
and others thinking only pools should have full nodes of one brand. where users just be sheep with lite wallets.
.. thats bad precedent to push for in regards to decentralisation.

It seems you haven't read the post in its entirety But it's OK I mentioned it in a minor sentence. Here I quote it for you:


So I think we basically agree here.


I disagree a bit because we can also develop sharding or similar solutions where not all nodes have to validate _all_ information (like the solutions Ethereum and Ardor are developing) and that also are the essence of "sidechains" and "extension blocks". The goal here is to split the network in sections that are compatible one to another. I know there are many challenges but from what I know it's not impossible. RSK is one of the projects that I'm closely observing in the Bitcoin space, but to work in a decentralized manner, it needs a modification of Bitcoin Core's software that actually seems to be rejected by the Core developers.



I'm discussing this topic here in the "miner Intranet" thread that I just opened to know more about "Fibre": https://bitcointalksearch.org/topic/do-mining-pools-directly-connect-one-to-another-miner-intranet-1891848

An interesting point that was brought on there is that although maybe as blockchain sources we really only have 20 nodes, from the individual node operator's point of view a full node is always the most secure solution and that's actually a good argument against too large blocks.