The Bitcoin blockchain is famously promoted as “immutable.” Is the structure of Bitcoin itself immutable?
Is the 21 million BTC limit out of the control of any individual? Is Bitcoin a decentralised entity of its own, the essence of its operating parameters unalterable by mere fallible humans? Is Bitcoin truly trustless?
Well, no, obviously. Even though bitcoiners literally argue all the above — most notably Saifedean Ammous in The Bitcoin Standard, and Andreas Antonopoulos in his books on Bitcoin — and these ideas are standard in the subculture.
Decentralisation was always a phantom. At most it’s a way to say “can’t sue me, bro.” Every process in Bitcoin tends to centralisation — because Bitcoin runs on economic incentives, and centralised systems are more economically efficient.
Trustlessness is also a phantom. Bitcoin had to create an entire infrastructure of trusted entities to operate in the world.
Something called “Bitcoin” will be around for decades. All you need is the software, the blockchain data, and two or more enthusiasts. But Bitcoin’s particular mythology and operating parameters are entirely separate questions.
Bitcoin’s basic operating parameters are unlikely to change in the near future — but this is entirely based on trust in the humans who run it. Their actions are based in whether changes risk spooking the suckers with the precious actual-dollars.
The Bitcoin Cash debacle destroyed any hope of substantive change to Bitcoin for a while, leaving Bitcoin as just a speculative trading commodity with nothing else going for it. But if a new narrative is needed, all bets are off.
Social convention is entirely normal, and how everything else works — but it’s not the promise of immutable salvation through code. That was always delusion at best.
Image by Mike In Space
How Bitcoin is marketed
Bitcoin is not about the technology. It’s never been about the technology. Bitcoin is about the psychology of getting rich for free.
People will say and do anything if you tell them they can get rich for free. You don’t even have to deliver.
Bitcoin also has an elaborate political mythology — which is largely delusional and literally based in conspiracy theories.
The marketing pitch is that the actual-money economy will surely collapse any moment now! And if you get into Bitcoin, you can get rich from this.
If you want to get rich for free, take on this weird ideology. Don’t worry if you don’t understand the ideology yet — just keep doing the things, and you’ll get rich for free!
That the mythology is so clearly at odds with reality is a feature, not a bug — it just proves the world’s out to get you, and you need to stick with the tribe.
So the key ingredient in Bitcoin is mythology. And job number one is: don’t spook the suckers.
Also, say “21 million” a lot. It’s a mantra to remind the believers to keep the faith.
The centralisation of mining
When Satoshi Nakamoto designed Bitcoin, he wanted the process of generating new bitcoins to be distributed. He distributed coins to whoever processed a block of transactions. (These blocks were chained together to form the ledger — which was called the “block chain” in the original Bitcoin source code.)
Just giving away bitcoins to anyone who asked wouldn’t work because of the sybil problem — where you couldn’t tell if a thousand people asking for coins were really just one guy with a thousand sockpuppets.
Satoshi’s way around this was to require some form of unfakeable commitment before you’d be allowed to validate the transactions, and win the coins. He came up with using an old idea called “proof-of-work” — which is really proof of waste. You waste electricity to show your commitment, and the competitors win bitcoins in proportion to how much electricity they waste.
This is called “bitcoin mining,” in an analogy to gold mining. Miners guess numbers and calculate a hash as fast as they can; if their guess hashes to a small enough number, they win the bitcoins!
Satoshi envisioned widely distributed Bitcoin mining — “Proof-of-work is essentially one-CPU-one-vote.” [Bitcoin white paper, PDF]
The problem here is that mining has economies of scale. The bigger a mining operation you have, the more you can optimise the process, and calculate more hashes with each watt-hour of energy. This means that proof-of-work mining naturally centralises. And this is what we see happening in practice.
For the first year, Satoshi personally did a lot of the mining — accumulating a stash of around a million bitcoins, that he never moved. More individual CPU users joined in over 2009 and 2010; but by late 2010, people started mining on video cards, which could calculate hashes much faster — to Satoshi’s shock. [thread]
When you have a tiny share of all the mining, rewards are sporadic — you might not see a bitcoin for months. The solution was to join together in mining pools, who would share the rewards. The first was Slushpool in late 2010. The Deepbit pool ran from 2011 to 2013; at its peak, it controlled 45% of mining.
By late 2013, application-specific ICs (ASICs) that did nothing but mine bitcoins as fast as possible were being deployed; you couldn’t compete without using ASICs. The most successful early ASIC manufacturer was Bitmain — who also controlled a large mining pool.
The doomsday scenario in early Bitcoin was a “51% attack” — if you had 51% of mining, you could block anyone else’s transactions and accept only those you wanted, and the Bitcoin blockchain would read the way you wanted it to. If anyone achieved 51%, it was game over!
GHash.io achieved 51% in July 2014. [Guardian, 2014]
The GHash.io pool promptly split apart, to calm the upset Bitcoin fan base. Nobody spoke of the 51% problem in Bitcoin again. (Though altcoins have frequently suffered 51% attacks.) Even Ammous talked about 51% attacks in The Bitcoin Standard and somehow forgot to mention that this had already happened.
But from 2014 on, Bitcoin mining was indisputably centralised. In 2015, the men controlling 80% of Bitcoin mining stood on stage together at a conference. Three or four entities have run Bitcoin mining since then. The only thing preventing miner misbehaviour is wanting to avoid spooking the suckers — it’s completely trust-based. Bitcoin now uses a country’s worth of electricity [Digiconomist; CBECI] for no actual reason. You could do the transactions on a 2007 iPhone.
Controlling the mining chips also controls mining. You can’t afford to piss off Bitmain — as Sia found out, when they couldn’t get chips for their minor altcoin fabricated in China because Bitmain didn’t like it.
The centralisation of development
Bitcoin’s operating parameters get tweaked all the time. There’s even a standard process for suggesting improvements.
Peter Ryan’s essay “Bitcoin’s Third Rail: The Code is Controlled” details how the Bitcoin development process works in practice: you submit changes to a core group, who then decide whether this is going in. The core group then decides what goes in and how it will work. [Ryan Research]
The thing is, this is completely normal. This is how real open source projects work. The essay presents this basic reality in a straightforward fashion; I’d take issue only with the headline, which paints this as in any way surprising or shocking. It really isn’t.
There are multiple bitcoin wallets — the official bitcoin.org wallet and many, many others. So the wallet software is nicely varied. You are, of course, trusting the developers — because approximately zero Bitcoin users are capable of auditing the code, let alone doing so. Hardware wallets such as Ledger have also been exploited plenty of times.
What if you don’t like what the developers do?
The fundamental promise of free software, or open source software, is that you have the freedom to change the code and do your own version, and the original developers can’t stop you. If you don’t like the official version of Linux, you can go off and do your own.
In the world of cryptocurrency, this means you can take the Bitcoin code, tweak a few numbers, and start your own altcoin. And thousands did.
What this doesn’t get you is control of the existing network that runs a crypto-token that’s called “BTC” and sold for a lot of actual money on crypto exchanges. That’s the prize.
The centralisation of trading
Crypto exchanges also benefit from economies of scale — there’s more liquidity and volume on the biggest exchanges. However, there’s now a reasonable variety of exchanges to choose from — not like 2014, when everyone’s coins were in Mt. Gox. You can even pick whether you want a pro-regulation exchange like Gemini, or a free-for-all offshore casino!
Exchanges collectively hold one important power: what they trade as the token with the ticker symbol “BTC.” As we’ll see later, this power needs consideration.
Crypto remains utterly dependent on the US dollar. Bitcoin maxis who profess to hate dollars will never shut up about the dollar price of their holding. (Unless number goes down — then they’re suddenly into Bitcoin for the technology.)
The point of crypto trading is to cash out at some point — and actual dollars have regulation. FinCEN are interested in what you do with actual money. Dollars pretty much always pass through the New York banking system, which creates a number of issues for crypto companies.
A few exchanges let you get actual dollars in and out. Many more — including some of the most popular — are too dodgy to get proper banking. These use tethers, a stablecoin supposedly worth one dollar.
Crypto trading solved the tawdry nuisance of dollars being regulated by using tethers instead, and cashing out your bitcoin winnings through Coinbase or Bitstamp as gateway exchanges.
If you can believe the numbers reported by the tether exchanges — which, to be fair, you probably can’t without a massive fudge factor — then the overwhelming majority of trading against Bitcoin, or indeed any other crypto, is in tethers.
This becomes a systemic issue for the crypto markets when Tether’s backing turns out to be ludicrously questionable.
The crypto market pumpers seem to think Tether is in trouble, and now the USDC stablecoin is issuing dollar-equivalent tokens at a rate of billions a month. USDC’s accountant attestations recently changed from saying that every USDC is backed by dollars in a bank account to saying that they may also be backed by “approved investments” — though not what those investments might be, or what the proportion is. [March attestation, PDF] I’m sure it’ll be fine.
Bitcoin doesn’t work at scale, and can’t work — because they took Satoshi’s paper-and-string proof of concept and pressed it into production. This approach never goes well. But insisting the fatal flaws are actually features has pacified the suckers so far.
Bitcoin is not very fast. It can process a theoretical maximum of about 7 to 10 transactions per second (TPS) — total, world-wide, across the whole network. In practice, it’s usually around 4 to 5 TPS. For comparison, Visa claims up to 65,000 TPS. [Visa, PDF]
In mid-2015, the Bitcoin network finally filled its tiny transaction capacity. Transactions became slow, expensive and clogged. By October 2016, Bitcoin regularly had around 40,000 unconfirmed transactions waiting, and in May 2017 it peaked at 200,000 stuck in the queue. [FT, 2017, free with login]
Nobody could agree how to fix this, and everyone involved despised each other.
The possible solutions were:
- Increase the block size. This would increase centralisation even further. (Though that ship really sailed in 2013.)
- The Lightning Network: bolt on a completely different non-Bitcoin network, and do all the real transactions there. This only had the minor problem that the Lightning Network’s design couldn’t possibly fix the problem.
- Do nothing. Leave the payment markets to use a different cryptocurrency that hasn’t clogged yet. (Payment markets, such as the darknets, ended up moving to other coins that worked better.)
Bitcoin mostly chose option 3 — though 2 is talked up, just as if saying “But, the Lighting Network!” solves the transaction clog.
But, the Lightning Network!
The Lightning Network was proposed in February 2015 as a solution to the clog that everyone could see was coming. It’s really clearly an idea someone made up off the top of their heads — but it was immediately seized upon as a possible solution, because nobody had any better ideas.
Lightning doesn’t work as advertised, and can’t work. This is not a matter of a buggy or incomplete implementation — this is a matter of the blitheringly incompetent original design: prepaid channels, and a mesh network that would literally require new mathematics to implement.
Users have to set up a pre-funded channel for Lightning transactions, by doing an expensive transaction on the Bitcoin blockchain. This contains all the money you think you’ll ever spend in the channel.
This is ridiculously impractical. You’re not going to send money back-and-forth with a random coffee shop — you want to give them money and get your coffee. Lightning’s promise of thousands of transactions per second can obviously only work if you have large centralised entities who almost everyone opens channels with. You could call them “banks,” or “money transmitters.”
Lightning originally proposed a mesh network — you send money from A to B via C, D and E, who all have their own funded channels set up. But routing transactions across a mesh network, from arbitrary point A to arbitrary point B, without a centralised map or directory, is an unsolved problem in computer science. This is even before the added complication of the network’s liquidity changing with every transaction. This basic design parameter of Lightning requires a solution nobody knows how to code. Again, this drives Lightning toward central entities as the only way to get your transactions across the network.
There are other technical issues in the fundamental design of Lightning, which make it a great place to lose your money to bugs, errors or happenstance. [Reddit]
Lightning is an incompetent banking system, full of credit risks, set up by people who had no idea that’s what they were designing, and that they’re still in denial that that’s what Lightning is.
The only purpose Lightning serves is as an excuse for Bitcoin’s miserable failure to scale, with the promise it’ll be great in eighteen months. Lightning has been promising that it’s eighteen months from greatness since 2015.
A good worked example of Lightning as an all-purpose excuse can be seen in the present version of El Salvador’s nascent Bitcoin system. Strike loudly proclaims it uses Lightning — but their own FAQ says they don’t pass transactions from the rest of the network. Bitcoin Beach uses Lightning — but reports indicate it doesn’t reliably pass money back out again, except via the slow and expensive Bitcoin blockchain.
With sufficient thrust, pigs fly just fine. [RFC 1925] The Lightning fans strap a rocket to Porky and then try to sell you on his graceful aerobatics.
The Bitcoin Cash split
So why not just make the Bitcoin blocks bigger? If Bitcoin sucks, we can fix it, right?
Bitcoin Cash, in 2017, was the last time there was a serious attempt to fix Bitcoin’s operating parameters.
Bitcoin developers could see the blocks filling in early 2015. Some proposed a simple fix: raise the size of a block of transactions from 1 megabyte to 2 or 8 megabytes.
But the Bitcoin community was sufficiently dysfunctional that even this simple proposal led to community schisms, code forks, retributive DDOS attacks, death threats, a split between the Chinese miners and the American core programmers … and plenty of other clear evidence that this and other problems in the Bitcoin protocol could never be fixed by a consensus process. “Trustlessness” just ends up attracting people who can’t be trusted. [New York Times]
This didn’t make the problems go away. Finally, in late 2017, large holder Roger Ver, in concert with large mining pool and mining hardware manufacturer Bitmain, promoted Bitcoin Cash, with large blocks, as the replacement for the deprecated Bitcoin software.
This wasn’t just starting a fresh altcoin — it was a fork of the blockchain itself. So everyone who had a large Bitcoin holding suddenly had the same holding of Bitcoin Cash as well. Free money!
The Bitcoin Cash split was a decentralised Judgement of Solomon — wherein an exasperated Solomon says “All right, we’ll just cut the baby in two then,” and the mothers think that’s a great idea, and start fighting bitterly over whether to slice the kid horizontally or vertically.
Bitcoin Cash launched in September 2017, and wanted to take over the “BTC” ticker on exchanges. This failed — it had to go with BCH. But Bitcoin Cash did have a chance at the “BTC” ticker for a while there; exchanges were watching to see if Bitcoin Cash became more popular.
Bitmain mined BCH furiously instead of mining BTC — and BTC’s block times went from ten minutes to over an hour. As it happened, Bitcoin was in the middle of a bubble — so nobody much noticed or cared, because all the number-go-up action was happening on exchanges, and not on chain.
Ver owned bitcoin.com, and furiously promoted BCH as a Bitcoin that you could use for cash transactions. This was a worthy ambition — but nobody much cared. Mostly, the few retail users of Bitcoin would get confused between the two, and end up sending money to an address on the wrong chain.
Bitcoin Cash completely failed to gain traction as a retail crypto. Most glaringly, it failed to get any takeup in the darknet markets — the first real use case for Bitcoin, and the people having the most practical trouble with Bitcoin’s transaction clog.
The BTC version of Bitcoin also lost all traction as a retail crypto — the average transaction fee peaked at around $55 in December 2017.
It was around this time that Bitcoin advocates stopped trying to pretend that Bitcoin would ever work as currency. They went in hard on the “digital gold” narrative, and pretended this had always been the intention — and never mind that the Bitcoin white paper is literally titled “Bitcoin: A Peer-to-Peer Electronic Cash System.”
There were other completely ridiculous fights to the death for insanely low stakes around this time — Segwit (which was eventually adopted, making blocks slightly larger), Segwit2x (which wasn’t), and UASF (where non-miners thought they could sabotage protocol changes they didn’t like). These were different in technical terms, but not different in bitterness or stupidity.
None of the disputes were really technical — it was all the politics of who got to make money. Everyone involved hated everyone else, and characterised their opponents as working in bad faith to sabotage Bitcoin. They figured that if they shouted enough abuse at each other, they’d get rich faster, or something.
Bitcoin Cash fell flat on its face. Bitmain ended up with 1 million BCH “inventory” — that is, a pile of coins there was no market for — and fired its entire BCH team in late 2018. [CCN]
BCH continues as just another altcoin with hopes and dreams — and not as any serious prospect to take over from BTC. The main thing keeping it alive is that it’s listed on Coinbase.
Also, you can still send BCashers into conniptions just by calling Bitcoin Cash “BCash” — the BCashers decided that calling BCH-Coin “BCash” was a slur. Though their real objection was that this term leaves out the word “bitcoin.” It’s unfortunate that BCash’s graphic designer never told them about the effects of putting a transparent logo PNG on a white background, so it looks like it says “B Cash.” [bitcoincash.org, archive]
Jonathan Bier has written a book about this slice of Bitcoin history: The Blocksize War (US, UK). Peter Ryan has reviewed the book; he thinks the chronology is correct, but the presentation is horribly slanted. [Twitter] I concur, and it’s glaring right from the intro — it’s a book-length blog rant about people who really pissed Bier off, and it’s incomprehensible if you weren’t following all of this at the time. It’s on Kindle Unlimited.
THIS IS WHY YOU GET A PROPER GRAPHIC ARTIST, AND PAY THEM.
But the blockchain’s immutable, right?
The blockchain is probably immutable — there’s no real way to change past entries without redoing all the hash-guessing that got to the present.
But the blockchain is data interpreted by software.
In 2016, Ethereum suffered the collapse of The DAO — a decentralised organisation, deliberately restricted from human interference, running on “the steadfast iron will of unstoppable code.” (Bold in original.)
The DAO was hacked — taking 14% of all ether at the time with it.
How did the Ethereum developers get around this? They changed how the software interpreted the data!
Immutability lasted precisely until the big boys were in danger of losing money.
There’s already legal rumblings in this direction — Craig Wright, the man who has previously failed to be Satoshi Nakamoto, has sent legal letters to Bitcoin developers demanding that they aid him in recovering 111,000 bitcoins that he doesn’t have the keys for. [Reuters]
Has history ended, then?
There are people who would quite like the 21 million Bitcoin limit to change: the miners. The 2020 halving dropped the issuance of bitcoins from 12.5 BTC per block to 6.25 BTC. In 2024, that’ll drop again, to 3.125 BTC.
The question is power — miners have a lot of power in the Bitcoin system. That power is shaky at present, because so much mining just got kicked out of China. Can they swing a change to Bitcoin issuance?
The bit where proof-of-work mining uses a country’s worth of electricity to run the most inefficient payment system in human history is finally coming to public attention, and is probably Bitcoin’s biggest public relations problem. Normal people think of Bitcoin as this dumb nerd money that nerds rip each other off with — but when they hear about proof-of-work, they get angry. Externalities turn out to matter.
Ethereum is the other big crypto that’s relatively convertible to and from actual money. If Ethereum can pull off a move away from proof-of-work, that will create tremendous pressure on Bitcoin to change, or be hobbled politically and risk the all-important interfaces to actual money.
(That said, Ethereum just put off the change from proof-of-work to … about eighteen months away, where it’s been since 2014. Ah well.)
Bitcoin mythology has changed before, and it’ll change again. “21 million” will be broken — if number-go-up requires it.
The overriding consideration for any change to Bitcoin is: it must not risk shaking the faith of the most dedicated suckers. They supply the scarce actual-dollars that keep the casino going.
This article was a commission: to write about “the fallacy that Bitcoin is immutable, when apparently a vote can increase the 21M cap or do anything else.” If you have a question you really want in-depth coverage of, and that I haven’t got around to doing a deep dive on — I write for money!
Your subscriptions keep this site going. Sign up today!