It can be made a reality pretty easily if we co-locate the miners and nodes into a single data center - or a few data centers well interconnected with big fat lines, in a way that even 1gb blocks at 3000tx/sec would not be an issue. But then you don't really have a peer to peer bitcoin. You have a (centralized) client/server one with plenty of attack vectors, including physical raids.



Actually satoshi said we can raise it when we are closer to needing it. That was just an example.


With everyday technology, current software and p2p topology of nodes all over the world, it has a scalability issue, yes.

If you centralize it, it stops having the problem. If you wait for technology to evolve, so that it can also allow bitcoin to work in a decentralized way, you can get a scalable AND decentralized bitcoin.

The bloat cost creates

1) storage costs
2) bandwidth costs (as this 1gb or 10gb or 100gb must be downloaded and served over and over in the network - for decades)
3) processing costs / power costs (for decades)
4) I/O slowdowns (if the blockchain can't fit anymore in an affordable SSD you'll have to rely on a mechanical disk which is slow)
5) centralization cost as nodes have to drop out, increasing costs to the fewer nodes that remain. It's a vicious cycle where the less you have, the more the costs, plus you now become an easier target for DDOSing (more costs)

As for the problems that we may need to worry, these are what exactly? That txs without a few cents in fees won't go in in the next block? That's all there is to it.

As far as I know, there is no dev team (core or classic or unlimited or whatever) that has a plan that says "1mbforevah", so it's not like fees will always be in relation to a 1mb block.


Long-term vision can differ from short-term. Satoshi did say that bitcoin wasn't suitable for microtransactions (at least not without an aggregating mechanism), but he was sure that it would be in the future. Seeing that tech went 1000x in ram, hd, network, cpu etc from 1995 to 2015, I'd say his certainty is well founded.

He forgot the bloat part which is achieved with less cost when the network has ample capacity.

If normal use, is, say, 500-700kb and the rest is spam that goes in for cheap, then instead of a spammer filling 300-500kb of spam for free/near free, he gets to increase that to 1300-1500, which is 3-4x.

If I want to move money around, why do I care what it is, since it does my job quickly and cheaply.


Yes, well, gold is not a business either. It's an element: [Au] 79.

Sure, it can be business for miners, speculators, investors, but that doesn't mean it is a business.



Right now, we cannot have blocksizes that can do VISA-level txs - not in any decentralized way. Some day we will. This is not an economic decision. It's a technical one that arises from the differences of how VISA is set up (centralized / efficient) and how bitcoin is set up (decentralized / inefficient).

And even if one thinks that BTC's tech nerds are idiots, then is there any other cryptocurrency that can do what Bitcoin can not and scale to VISA levels, and do so in a decentralized manner? The answer is no.

Bitcoin is not a business.

From my perspective, the segwit upgrade is being done ahead of time as >1mb is not really needed right now, nor will it be in 2 months.

If it was getting closer to getting needed, we'd see some serious signs:

- you'd see pools use 1mb soft-limits instead of 250/750 - raising the avg block at something closer to 1, like 0.95mb, instead of 0.6-0.8mb (which have 20-40% more space)
- you'd see fee pressure building up and you would definitely not have 0.04-0.06$ for first block inclusion or 0.003$ to 0.02$ for any-block-inclusion.
- you'd see miners not mining empty blocks leaving money on the table (because there is no money on the table as the fees are too low)
- you'd see serious arguments from the opposition and not lies, propaganda or distorted statistics

Let's say I have far more belief in Satoshi and the mechanisms embedded in Bitcoin than those who are trying to "save" it.

Two mechanisms in particular:

a) The fee market
b) Value scaling

If Bitcoin "can't scale", with given network parameters intact since 2010, then perhaps someone can answer me how did it climb from a few dollars per day in transactions in 2010 to doing 150-230 million USD per day? It would seem impossible if we heard all the trolls.

And why is it that it can still scale from 200 million per day to 2 billion USD per day, doing over 600 billion USD per year and getting a volume that exceeds paypal and western union, multiplied by two (!), without even increasing the basic system by 1 transaction per second?

The compensatory mechanisms are extremely strong, even if there is currently a weakness in terms of transaction/sec capacity.

And since we are in an investment-related thread:

You can use the 250k transactions per day for buying coffees at 3$ apiece. You can even double or quadraple the blocksize to 4mb and get 1mn transactions of 3$ on average. Now tell me, do you feel better that you can do 12tx/s and you can buy 12 coffees per second? Do you feel better that you can only do 3mn per day or just 1 billion per year by doing microtransactions? Can you even justify BTC's marketcap of 6+ bn by doing that? The answer is no.

If, on the other hand, you are doing 250k or 1mn txs of 1000$ apiece, you are doing 250mn USD per day* to 1bn USD per day - or 91bn to 365bn per year, beating the likes of Western Union and Paypal, all the while making your marketcap grow astronomically because it's very under-rated at 6bn. You can choose whether you can go after microtransactions -where you will waste your enormous advantage- and harm your marketcap, or moving serious money and being the real deal. And the future is even brighter because we'll be able to do *both* (mass low value and high value transactions). It's technologically assured based on the tech progression curve.

* We are currently doing 150-230mn per day at 600-800$ per tx: https://blockchain.info/charts/estimated-transaction-volume-usd?timespan=30days&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

Using bitcoin for larger transactions is the way bitcoin can be utilized best - because that's where the biggest savings are.

Here is an example:

You have ~250.000 txs per day.

Now, you can do 250.000 micropayments per day, averaging, say 1$. So that's 250.000 USD per day in volume.

What have you accomplished, in the larger scheme of things? Nothing. Your annual transaction volume is a pitiful 91 million dollars, while having a marketcap exceeding 6bn USD. You can't even justify your marketcap with microtxs.

Even at 2mb blocks or 4mb blocks, or 10mb blocks, you are still at 180mn / 360mn / 910mn usd per year - again you can't justify your marketcap. It has to deflate and lose value.

Now you get those 250.000 txs per day and you attach to them 600$ in avg value*.

Suddenly you are moving around 200mn USD per DAY or 73 BILLION PER YEAR.

At 2mb or 4mb, you are hitting 146bn to 292bn per year.

Now you are making an impact**. Now you are saving people of billions in fees that would have gone to the western unions, paypals and credit cards. You can't save people those billions by throwing away your advantage and doing microtransactions, buying coffees.


* That's around how much we are doing right now, on avg, per tx: Currently ~250k txs can generate ~150-230mn USD per day in transaction volume. That's outputs minus change. https://blockchain.info/charts/estimated-transaction-volume-usd?timespan=30days&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

Some days it peaks to >800$ per tx - and in a few years we might be seeing averages in the tens of thousands - meaning that BTC will be able to pull annual transaction volumes in the trillions of USD - contrary to the propaganda that BTC can't scale (since value per transaction CAN scale 5-10-100x with no technological issues). Keep in mind that our current 600-800$/tx is currently done while fees for first block inclusion are typically at an extremely low 0.04$ to 0.06$ and while dust movement is still going on in the network because it is economically ...viable. Also keep in mind that western unions, paypals, banks, charge tens of dollars for moving around, say, 500 or 1000$ around the globe. A 250$ payment received through paypal cost me 20$ in paypal and bank fees and a western union transfer of 180$, cost me 25$. And people are complaining that BTC is ...expensive, or that it's excluding them. You have to lol pretty hard when you get a reality calibration versus the legacy payment systems.


** For comparison purposes:

Western Union[1] does 85bn in moving funds between customers and paypal[2] does ~230bn per year.

1. http://ir.westernunion.com/News/Press-Releases/Press-Release-Details/2016/Western-Union-Cross-Border-Platform-Connects-Consumers-to-over-One-Billion-Bank-Accounts/default.aspx

2. https://en.wikipedia.org/wiki/PayPal

As far as I'm concerned, I don't have any evidence that we actually need >1mb right now, although we will get >1mb... there is actually no plan that keeps bitcoin at 1mb.... core says 1.7 in a couple months, classic says 2mb with contentious/destructive hf, so... there you have it.

When you have tx fees ranging from 1 to 5 cents (depending the priority one wants), that means there is no overwhelming actual transaction demand which in turn has an underdeveloped fee market while also allowing spare capacity to be used for spamming and wasteful transactions.

The fact that some people have 1-2 cents in their wallet software and it takes a few hours when there is an attack, is not enough to hf bitcoin and it won't change their user experience once the 2mb are flooded the same way. Therefore they need a better wallet software. The system is working as intended per satoshi's instructions on how to bypass a flood attack (you outpay the spammer). If users don't know it, it's because their wallet software obviously doesn't do a good job of notifying them that fees are dynamic and not static.

As for the timing: Even if the code for HF was issued tomorrow it would still get activated months ahead so that users can get prepared. If the recent "transaction problems" are any indicator, then if people don't even know that they have to pay a few cents to transact, then they can't *really* be depended upon for upgrading in time. It will simply create a mess. So, TLDR, segwit deployment would be much faster. A HF should be planned well ahead, again per the instructions of Satoshi.

That's the plan: Segwit = 1.7mb to 3mb (depending use), plus solves tx malleability.

Obviously, all these "disagreements" aren't about the ~300kb difference between segwit and a 2mb hf.

There are some intel chips with 128MB L4 cache (shared cpu/gpu). Obviously the further "away" it is from the cores, the slower it will be, but still you'd expect it to be at least twice as fast as the ram.

The number is an example. You can only code it in satoshi/byte, so what is currently 0.08$ will be 0.25$ by the time we are back at the ATH of 1200+$. That will still be below PayPal / CCs and probably in line with "top of the microtransaction range". Of course, that's just my number, another one might say 1$, another one might say 0.1$ - who cares. Opinions are like a$$holes and everyone has one...  right?

What counts is the intent of transacting. If one wants to transact then they'd have no problem paying a few cents for their transaction. It's not any serious burden. If they want to spam thousands/millions of txs => they'll have to pay some serious fees. Right now it's typically costing the spammers 0 to 2 cents. That's the width of the spamming range.

If serious fees aren't used as a disincentive then it's an economic amplification attack against the system, in the sense that it costs me zero or near zero to burden the network and 7.000 nodes must then carry processing, bandwidth and storage costs for years and decades (which is not sustainable).

Anyway, all these and more, the devs already know. They have to balance things between abuse and expansion as well as science and populism. We'll see how it goes.

It means you only need to work with economic disincentives for those who are serious in transacting versus script kiddies.

One man making one-five-ten transactions with, say, a 0.08$ fee won't go "bust" by paying these low fees.

One man that doesn't really want to transact, but has the intention to spam the system with, say, 1 million transactions => that would be another issue altogether because he'd be required to pay a minimum of 80.000$. It would harm his wallet so to speak.

Still, you have large actors that could pay the 80.000$, or 8mn$, or 80mn$, but at least you got rid of the script kiddies.

Yes, I have. Starting with zx spectrum basic in the 80s and leading to assembly on PCs in the 90s.

However trying to understand the intent by code is unnecessary. You can see it and patch it in some cases (like dust movement), but it can be fooled with larger amounts. Anyway, the primary way to deal with the issue is simple: Fees. That's what all altcoins do when they get attacked. Raise fees. That's what the practice shows. The miners will typically transact anything that comes their way because they follow the system's rules. If the system hasn't the proper disincentives in place, in a way, it's not their fault.

You asked me about my definition of spam, not how we can limit it by using code.


I've been dealing with spam and dos since the mid-90s. I know pretty well when my worries are founded or not and when assurances are crap.

We disagree on the importance of the network to process as many transactions as it can - unconditionally (whether they are spam or not). The creator of the system didn't make the system to get abused and destroyed by spam.

I also disagree with the idea that other people have on what constitutes "growth", like if we make 10mb blocks and fill them tomorrow morning with crap, that it would "mean" that "bitcoin is somehow being used at 10x rate", which would in turn "mean" it's "network effect is better" and thus price should increase. This is bullshit. Spam isn't growth. Like hitting the reload in a web page 100.000 times isn't "page traffic" (compared to unique views).

I would definitely like the network to be able to process as many txs as it can, but only by paying some serious fee as not to have the system abused. That's the only rational thing to do.

Unfortunately, the political situation and animosity is such that I doubt core devs have the ability to propose fee increases to deal with spam. Perhaps they could create 2-3-5 block increase proposals that could go like that:

2mb effective blocks with unchanged fee structure
3mb effective blocks with min fee X
4mb effective blocks with min fee X+y%
5mb effective blocks with min fee X+z%.

...or something like it.

My definition of spam is when someone doesn't transact because they want to make an economic transaction, but rather to do something else, like filling blocks with crap or bloating the blockchain.

Apparently satoshi did take it seriously and had planned the fee mechanism as a countermeasure.


Obviously, fees won't work as a countermeasure for spam if they are too low. The whole point is to disincentivize the attacker. If the attacker doesn't feel sufficient pain he can easily proceed with the attack. Giving the attacker more space to conduct the attack, instead of more fee pressure, will simply amplify the attack and reduce the attackers cost.

The fee pressure might originate from either blocksize scarcity, miners setting a higher fee limit to what they process and/or devs setting higher default recommendations (or even network enforced fees - as long as the miners run the code).

There is also paid spam... for example if I pay 1 satoshi per byte, I can claim "I paid fees for my tx".

At a 1 satoshi per byte rate one can insert a billion bytes for a billion satoshi. That's ~4300$ for a gigabyte of spam. The blockchain is currently at 67GB, so spamming 67GB would cost just 288k USD. The costs for bandwidth, storage, processing etc of all that crap will exceed, over time, what was paid by the spammer, so, in a sense, it would represent a type of financial amplification attack, where the spammer incurs multiple costs to the ecosystem by spamming away - whether he does it for free or "pays fees" like 1 satoshi / byte. Even at 10 satoshi / byte the attack is feasible.

Who cares about "some devs" when Core has an official roadmap with segwit up ahead. The criticism, if you look at the expanded paragraph, is aimed directly at core for supporting things like LN instead of ever-ending block increases, all the while pretending they offer zero on-scaling solutions.

He writes at one point:


That's so blatant, as a lie, it's not even funny. Andersen's credibility tends to zero. Core's roadmap has a Segwit rollout in a few months, increasing tx/s and effective block size by ~70%. How can he pretend like there is absolutely nothing?