What a blockchain is, and what it isn’t

Is a spreadsheet the right metaphor?

There are four differences from an ordinary shared sheet, and they’re all in the rules. Everybody holds a full copy, rows already written can’t be edited, everybody sees everything, and to add a row you follow a mathematical procedure instead of asking permission from whoever created the file.

That’s where the name comes from too. Rows don’t go in one at a time but in groups, the blocks, and each block carries the fingerprint of the one before it: blocks in a chain, blockchain. Remembering there’s a table underneath keeps you from being dazzled, because the remarkable part is the rules they built around it.

Who decides a row is true?

If the majority says it’s fine, the row goes into the block and becomes final. If somebody tries to write that they sent money they don’t have, the others have the whole history in front of them and reject it. Nobody needs to investigate: it’s enough that the numbers don’t add up.

That’s why the number of copies matters more than how powerful any one of them is. A network with few copies in few hands is faster but also more fragile. One with many copies spread around is slow and hard to bend. You can’t have both, and that’s why blockchains are so different from one another.

How do you earn the right to write?

The second is proof of stake, which Ethereum has used since 2022: whoever wants to take part puts up tokens as collateral, like a deposit. Validate honestly and you get paid; cheat and you lose the deposit. The whole thing uses a fraction of the energy of the first.

Both do the same job, making cheating expensive without having to trust anyone. And both have a price, energy on one side and concentration of capital on the other, which you’ll find compared point by point in the guide devoted to them.

What are the four properties, and what do they cost?

Immutable means nobody rewrites the history, and neither do your mistakes get rewritten. Transparent means everything is checkable by anyone, including what you do. Distributed means the network doesn’t go down, not that power is spread out. Always open means no opening hours, and it also means the risk doesn’t sleep.

It’s immutable: what’s written stays, and nobody can rewrite the history. That goes for your mistakes too, because if you send money to the wrong address it’s gone, for good, with no appeal. It’s transparent: anyone can download the whole history and check it back to day one. That goes for your activity too. Addresses don’t carry names, but anyone who cares to look can see what you do, and it doesn’t take much to put two clues together and get to the person.

It’s distributed: if a piece goes down the network carries on, and to switch it off you’d have to switch off the internet. That doesn’t stop power from concentrating, though, as you’ll see in a moment. And it’s always open: no hours, no weekends, no holidays. That’s handy, and it also means a position can get liquidated at three in the morning while you sleep.

How distributed is it, in practice?

In bitcoin the copies of the ledger number in the thousands, but the computing power is concentrated in a few mining pools. In Ethereum anyone can validate, but it takes thirty-two ether to do it alone and about 25% of the stake goes through a single operator. On the fast networks the validators are few by construction.

Ethereum has the opposite problem. In theory anyone can take part. In practice you need thirty-two ether to do it alone, so almost everyone goes through somebody who does it for them, and a single operator, Lido, handles about 25% of all the stake today. The faster networks, from Solana down, have few validators by design: that’s the price they pay for speed.

The useful takeaway is that “distributed” tells you how hard a network is to switch off, and nothing about how the power is shared. They’re two different things and worth looking at separately, not least because the first is almost always true and the second almost never.

What is it actually good for, and what not?

These hold up: transfers of value between countries, agreements carried out by a contract, ownership of digital goods. These don’t: supply-chain tracking, because a false entry at the start becomes permanent; voting, because a stolen key votes in your place irreversibly; medical records, which don’t improve by being public and impossible to correct.

Where the problem is something else, it’s no use. Tracking tomatoes from the field to the supermarket doesn’t become reliable because the ledger is immutable: if someone writes something false at the start, that falsehood becomes permanent, garbage in, garbage forever. Voting from home doesn’t become safe, because anyone who stole your key could vote in your place, and there’d be no undoing it. And medical records don’t get better by becoming public and impossible to correct.

Before every “let’s put it on the blockchain”, the question to ask is always the same: do I need an immutable, public ledger here, or just a well-built database? In the second case an ordinary database costs a thousand times less and can be corrected, which most of the time is exactly what you need.

Why is it talked about as a revolution?

It’s underrated because under all that noise there’s something that didn’t exist before: a ledger nobody controls that anyone can write to by following the rules, and programs that carry out an agreement without anyone having to authorize them. That opens up new possibilities, and decentralized finance is the first and most concrete of them.

Two questions are enough to find your way. The first is which middleman this thing is cutting out, and if there’s no answer, you don’t need a blockchain. The second is what breaks if the code is badly written, because in a system where what’s written stays written, a mistake doesn’t get fixed. You pay for it.