Be your own bank

An article for normal people to understand what this Bitcoin thing is all about.
Be your own bank

A strange bank

“Nodes” and “miners” don’t mean anything to most people. So forget those words for a minute and picture a bank. The Bank of Bitcoin.

It’s a strange bank. Nobody runs it. There’s no CEO, no headquarters and nobody who owns it. It doesn’t have a vault either. The bank never holds your money, so it can’t freeze your account and it can’t reverse a payment.

What it does have is a ledger. One book with every payment written in it since the doors opened in 2009. There’s also a rulebook that everyone checks the ledger against. Two kinds of workers keep the ledger honest, tellers and bookkeepers. Anyone in the world can take either job today without asking anyone.

The easiest way to understand it is to follow one payment through the building. Say you want to send some bitcoin to a friend.

What’s in your pocket

Before you walk in, know what you’re actually holding. It isn’t bitcoin. Nobody holds bitcoin. Coins only exist as entries in the ledger.

What you hold is a key. The ledger says certain coins can only move with a signature from that key. To pay your friend you sign an instruction that says these coins, from my key, go to theirs. That signed instruction is a transaction and it’s the only thing you ever bring to the bank. You also attach a fee for whichever bookkeeper ends up writing your transaction into the ledger. We’ll get to them.

Whoever holds the key controls the money. There is no account recovery desk. Lose the key and the coins sit in the ledger forever and nobody can spend them. Let someone copy it and they can spend your coins before you do. Holding your own keys means backups and security are on you.

Now compare that to keeping bitcoin on an exchange. The exchange holds the keys. You never talk to a teller. You ask the exchange to send a payment and the exchange decides if it wants to pass your request along. Your “balance” is a promise written in the exchange’s own private ledger.

That’s what “be your own bank” actually means. You aren’t running a bank. You’re the only one who can sign for your money, and you walk up to the window yourself.

At the window

You hand your signed transaction to a teller. Tellers are what Bitcoin people call nodes.

The teller’s job is simple. They check your transaction against the rulebook every teller has agreed to. Does the signature match the key that controls those coins? Were those coins already spent? Does it create money that shouldn’t exist? That’s it.

The teller doesn’t get an opinion. There’s no manager to call and no “let me see what I can do.” The software says valid or invalid and that’s the answer. A teller can’t approve a bad transaction any more than a calculator can decide 2 + 2 = 5.

Tellers also don’t take anyone’s word for anything. On day one a new teller replays the entire history of the bank, every page back to 2009, and checks every entry themselves. Only after that do they start checking yours.

Tellers don’t get paid and you don’t have to be one. Most people never are. Your wallet hands your transaction to someone else’s teller, who checks it and passes it along like any other. There are thousands of independent tellers all enforcing the same rulebook, so trusting that network is reasonable for most people. A teller you don’t run could lie to your wallet about what happened. It can’t fake the bookkeepers’ work though, and good wallets ask more than one teller.

Running your own teller is an upgrade, not an entry fee. It’s also cheaper than most people think. Once a teller is caught up the work barely registers. Checking one new page every ten minutes takes almost nothing in computing power or bandwidth. Old laptops, retired desktops and cheap single board computers all handle it fine. The only heavy part is the start, when a new teller downloads and checks the whole history. On old hardware that can take days but you only do it once.

Storage is the real cost. The full ledger is several hundred gigabytes and still growing. If you’re short on space you can run a pruned teller. It still checks every page back to 2009, then throws away the old pages and keeps the recent ones plus the current record of who owns what. That fits in a few tens of gigabytes. A pruned teller enforces the rulebook exactly like a full one. It just can’t hand old pages to newcomers.

So what do you get for running one? No outside teller learns which coins are yours every time your wallet checks a balance. Nobody can quietly misreport your payments to you. You never need anyone’s help getting a transaction to the back offices, because your teller is a teller like any other and the bookkeepers treat it that way. And you’re one more teller enforcing the rules you actually signed up for.

The back office

Your transaction checks out. The teller files it in their back office, which Bitcoin people call the mempool. It’s a stack of valid transactions waiting to be written into the ledger.

Every teller has their own back office. Every teller also has assistants who pass verified transactions to the other tellers they’re connected to. Each of those tellers checks it again, files it and passes it on. Within a few seconds your payment is sitting in the back office of nearly every teller in the world.

Tellers do get some say here. Each one sets house policy for their own back office. They can skip transactions with tiny fees, turn away certain kinds of data, and so on. House policy isn’t the rulebook though. It only decides what that one teller keeps on file.

House policy can’t lock out a valid transaction. If no teller you know will take yours, become a teller, file it yourself and let your assistants spread it. Somewhere along the line other back offices will accept it and the bookkeepers will find it.

The bookkeepers

Your payment is filed but it isn’t in the ledger yet. Writing the ledger is a separate job done by bookkeepers. Bitcoin people call them miners.

The ledger gets written one page at a time. Each page is a block. To earn the right to write the next page, a bookkeeper has to burn energy hashing. Think of it like a lottery where you print tickets as fast as your machines can go. A winning ticket is hard to find and dead easy to check. More tickets per second means better odds.

The rulebook sets how hard it is to win. Every 2,016 pages the difficulty adjusts so the whole room of bookkeepers together finds a winner about once every ten minutes. Tellers don’t pick that number. They just reject any page that doesn’t meet it.

The rulebook also caps how much fits on a page, so space is scarce. This is where your fee comes in. Bookkeepers usually fill their page with whatever pays the most. Usually. It’s their page and they can put what they want on it as long as it follows the rules.

The winner gets paid twice, for now. They get every fee on their page plus newly created coins. That second part is called the subsidy and it’s the only way new bitcoin ever comes into existence.

The 21 million coin cap is one of the rules nobody touches, and the subsidy is how the coins get handed out without going over it. In 2009 the winner of each page got 50 new coins. Every 210,000 pages, about every four years, that reward gets cut in half. Today it’s 3.125 coins. Around year 2140 it rounds down to zero and the last sliver of the 21 million is out. No bookkeeper can pay themselves one coin more than the schedule allows because the tellers would throw the page out.

This makes Bitcoin’s distribution different from almost any other money. Every coin that exists first entered the ledger as a bookkeeper’s reward for real work. There was no premine, no insider allocation, no investor round and no foundation treasury. Early bookkeepers got coins cheap, but under the same rules anyone could have joined. Everyone else got theirs by trading for them.

When the subsidy runs out, bookkeepers get paid entirely from fees on transactions like yours. That’s on purpose. Bitcoin is a push system. You send the payment and you pay the people securing the ledger directly. Traditional finance is a pull system. You authorize an institution to take money out of your account, and the cost of running the whole thing is buried in spreads, card fees and account charges.

The energy bookkeepers burn isn’t wasted either. It’s the reason the ledger is so expensive to rewrite, and your fee is what pays for that security.

A bookkeeper can be a guy in his garage, a pool of thousands, a public company, or anything in between. Anyone can join and nobody gets out of following the rulebook.

Into the book

A bookkeeper pulls your transaction from the back office, puts it on the page they’re working on and starts printing tickets. Eventually somebody wins. If your fee was competitive, your payment is on that page.

The winner takes the page to the tellers, and this is where the tellers’ power shows up. Every teller checks every line on the page against the rulebook and checks the winning ticket. One bad line and the whole page gets thrown out no matter how much energy went into it. A cheating bookkeeper doesn’t get punished. They just don’t get paid.

If the page holds up, every teller adds it to their copy of the ledger, chained to every page before it. Your friend has been paid.

Every new page written on top makes yours harder to rewrite, since rewriting it means redoing all that work. That’s why people wait a few pages before treating a big payment as final.

Nobody runs this bank

Go back through what just happened and look at who actually had power over your payment.

The tellers could check it, but they couldn’t approve or deny it because they felt like it. Their software follows the rulebook. They can’t fire a bookkeeper who plays by the rules either. All they can do is throw out pages that break them.

The bookkeepers could choose whether to include it, but they couldn’t force anything invalid past the tellers. Bookkeepers don’t work for the tellers. They work for whoever pays fees, and really for themselves. They’re in it for profit.

These aren’t separate tribes either. Every serious bookkeeper runs their own teller because they need to know what’s valid before betting energy on a page. Exchanges, merchants and wallets run tellers too.

Counting tellers isn’t counting votes. Spinning up a teller is cheap. Ten thousand tellers nobody spends money through matter less than one teller at an exchange settling billions. Influence follows the people actually using the bank.

Nobody owns the Bank of Bitcoin. You join it by choice, and its rules hold because the people who hold, spend, check and write all keep choosing the same ones.

What if one bookkeeper gets too big?

Some bookkeepers do get big. Most of the big names are actually pools, thousands of individual machines combining their lottery tickets and splitting the winnings. The pool operator usually picks what goes on the page, but newer methods let the individual miners build the page themselves. That’s another article, and it’s one of a long list of improvements landing all the time. Either way the people who own the machines pick which pool to point them at.

A huge bookkeeper can do less than you’d think. Even one with most of the hashing can’t spend coins from your key, can’t print coins past the schedule and can’t break any other rule. Every teller throws those pages out no matter who brings them.

What a dominant bookkeeper can do is narrower but real. They can refuse to write certain transactions. With a majority of the hashing they can ignore everyone else’s pages and rewrite recent history to reverse their own recent payments. That’s the famous 51% attack, and it’s why big payments wait a few pages before counting as final.

So why doesn’t it happen? A bookkeeper gets paid in the bank’s own coin and has sunk a fortune into machines that are useless for anything else. Openly attacking the bank tanks the value of the coin they earn and the equipment they own. Being honest pays better.

It also tends to fix itself. Pool operators answer to the machine owners, and machine owners can switch pools in minutes. In 2014 the pool GHash.io crept up toward half of all hashing. Miners left in droves and it shrank within weeks. And a transaction one bookkeeper refuses is still sitting in the back offices, waiting for any other bookkeeper who wants the fee.

If a bookkeeper ever really went rogue, the tellers could fire them. The only way to do that is changing the rulebook though, and only if the bank’s users wanted it. The bluntest version is changing the lottery itself, the proof of work, which fires every bookkeeper running today’s machines all at once. That’s the nuclear option. It means a new rulebook, a split bank, and betting that the users follow. Ten million tellers announcing new rules doesn’t make it happen just because they said so.

So a rogue bookkeeper isn’t impossible. The damage they can do is limited, the incentives push hard toward honesty, and the fix is in the hands of the people using and securing the bank. Not some authority.

When the rulebook splits

House policy is one thing. What happens when tellers disagree about the rulebook itself?

Say a group of tellers wakes up tomorrow running a new rulebook that prints extra money. Their software now accepts pages everyone else rejects. The moment one of those pages gets written, the ledger splits. Now there are two banks with the same history up to that page and different history after it.

Each teller follows the ledger with the most total work that’s valid under their own rulebook. A ledger with more work but broken rules isn’t even in the running for them. It’s noise. So each group sees its own bank as the real one.

So which one is the Bank of Bitcoin? The tellers don’t settle that and neither do the bookkeepers. Everyone who held coins before the split now has coins in both banks. They can keep both, sell one for the other, or ignore one completely. Bookkeepers follow the fees and the price. Merchants take whatever their customers pay with.

The bank people keep using, valuing and securing is the one everyone calls the Bank of Bitcoin. The other one still exists. It’s just a different bank with different tellers, different bookkeepers and a different crowd paying for the work.

Rule changes do happen, but nobody gets to just declare one. A change gets proposed in the open, bookkeepers signal support on the pages they write, and everyone watches the count.

2017 is what it looks like when it works. An upgrade called SegWit stalled because large bookkeepers wouldn’t signal for it. Users answered with a user activated soft fork, BIP148. Tellers set a date after which they’d reject any page that didn’t support the change. Exchanges and businesses backed it, the bookkeepers fell in line before the deadline, and SegWit activated. The big block holdouts left and opened their own bank, Bitcoin Cash.

BIP-110 is what it looks like when it doesn’t. Its tellers started rejecting pages that didn’t signal for its data restrictions in August 2026, even though only about 2.5% of pages had signaled and it needed 55%. The ledger split at page 961,632. The BIP-110 side wrote two pages and stalled while the main bank kept going, and no major exchange backed it. Its supporters say it hasn’t failed. But the bank people kept using ran on the rulebook they were already using.

Both used the same mechanism. What made the difference wasn’t how many tellers ran the new rules. It was whether the people holding, spending and accepting the money came along. Tellers decide what their bank is. Everyone else decides which door to walk through.

Why people keep walking through this door

Anyone can start their own bank. Copy the ledger, change the rulebook, run the software. The code was never the hard part. Getting people to walk through your door is.

People keep walking through the door of the bank whose rules don’t change because a small group wants them to. Nobody with savings in a bank volunteers to get diluted. A fixed supply that no teller or bookkeeper can change on their own is the whole point of the Bank of Bitcoin.

So hold your own keys, and run your own teller if you can. That’s what being your own bank is.


Write a comment