Dreaming Systems: Let Me Buy You an Ice Cream

So far, this series has been about perception: Where we are when we look at screens, why systems feel inevitable in hindsight.

Now, let’s stop thinking. Instead, I propose we actually do something.

It was 2013 when we left off. Or was it 2012? My memory is pretty sketchy for a memoirist. You’ll have to live with that.

Anyway: now it’s 2016.

I’m still sitting at my laptop. Don’t worry - I did get up at some point. My laptop is in a different place now.

I’ve been self-employed for a while. I think I’ve found a good compromise between my professional background and my interests. For the moment, I am pretty happy about the status quo. I write about medicine and technology, and the overlap between the two.

An inquiry lands in my inbox, from a US-based blog platform: would I be interested in writing something about applications of blockchain technology in medicine?

Sure, I can do that, I reply.

Then I google: what was that blockchain thing again?

Ah.

No - actually, not really ah. I read a few articles and still don’t quite understand it.

Blockchain is the technology that Bitcoin is built on, an electronic means of payment. I remember my first fleeting encounter with it a few years earlier. Bitcoin hadn’t sparked my interest back then - the technology and I, we had gone our separate ways.

So blockchain is a kind of file in which Bitcoin transactions are stored. You could, I read, also store other things in it: for example, information about patients.

I’m not entirely sure what that’s supposed to achieve.

I could already complete the assignment at this point. I could write a fluffy article about blockchain in healthcare and submit it. But I don’t want to - I want to understand it properly before I explain it to others.

I read a few more articles, and then I start watching YouTube videos about the topic. That’s how I learned programming in Python back then, when the first freely accessible online courses appeared: Massive Open Online Courses from well-known and lesser-known universities, lecture videos and assignments.

But this time I don’t want to pay, don’t even want to sign up anywhere. Everything I need is freely available out there. YouTube explains blockchain to me, and at some point, I get it.

I love that moment - it feels a bit like learning to read, for the first time, every single time.

So: blockchain. What is it, and what is it for?

It’s called a blockchain because it is a chain of blocks. The blocks are, of course, digital - that is, stored on a computer.

Inside these blocks, transactions are encoded and written down. A transaction, in the most well-known application of the technology, is a transfer of currency from one address to another.

Let’s say you are the owner of my favourite ice cream parlour. I want to buy an ice cream from you: two scoops, chocolate and pistachio, no whipped cream, in a cone. I create a transaction in which I transfer 0.000013 Bitcoin from my address to your address.

How can you know that the 0.000013 Bitcoin have actually arrived?

That you can hand me the ice cream without being left with the expenses (pistachios ain’t cheap)?

Where do you check?

If the transaction were stored only on my computer, you would have a problem. I could show you that the transaction is there, you give me my ice cream, and then I delete the transaction again. I’m happy about my free ice cream - but you’ll be out of business soon.

So we need a way to ensure that both of us agree that the payment has taken place.

Traditionally, that’s what banks are for. Even when I pay for my ice cream in cash, the central bank is involved in the background.

The bank stands between us, as an intermediary: you are protected by the bank from me and my potentially fraudulent intentions. But I am also protected from you (you could claim I had eaten the ice cream without paying in valid currency).

This requires that both of us trust the bank - that you assume the bank is not secretly on my side.

And it creates costs, which you or I or both of us have to bear - or someone else does for us. Your house bank’s elegant office in the pedestrian zone (or the advertising campaigns of your direct bank), the salaries of the employees - none of that pays for itself. No free lunch and all.

These are the two most important arguments that have always been made for cryptocurrencies and other blockchain-based ventures: decentralization removes the need to trust a central authority - and it reduces costs.

Back to the blockchain and my 0.000013 Bitcoin: you and I don’t want a bank to approve our transaction. We don’t want to pay for the tiles in the bank’s lobby, and perhaps we come from a country where banks - maybe even the central bank - are not always reliable or trustworthy. Whatever the reason: we want to handle this ourselves, no outsiders involved.

So we start by storing the blocks with the transactions on multiple computers. The more copies there are, the more likely it is that manipulation of one copy will be noticed. These copies can - and should - be public: the more eyes that can verify the contents, the better.

If we really want to create an electronic currency on this basis, there’s another complication: it shouldn’t just work for the two of us. Ideally, everyone should be able to use it - or at least as many people as possible.

That makes things messy. Anyone could try to modify copies of the blockchain and invent their own transactions. Transfer money to themselves from other participants who never agreed to it.

So we need a deterrent mechanism - something that makes fraud unattractive.

In Bitcoin, this is called proof of work.

But first, we need a rule: a transaction is only valid once it has become part of a block, and that block has been added to the existing, currently valid blockchain. When any participant creates a transaction, it is (after a small plausibility check) placed on a waiting list. Every computer in the network has a copy of this list.

In principle, any computer in the network can take transactions from this list, assemble them into a complete block, and add that block to the blockchain.

But it must overcome a hurdle, and this hurdle is proof of work.

Before a computer is allowed to add a new block, it has to solve a mathematical puzzle. This is a kind of computational problem whose solution is very difficult to find - but once found, it is easy for others to verify.

The computer (or node) that solves the puzzle first wins the right to add the next block to the chain. Everyone else can easily check whether the solution is correct. If it is, they adopt the new block into their own copy of the blockchain.

That is proof of work - a “proof of effort.” The computer must prove that it has invested work into solving the puzzle before it can add a new block.

To complete the security of the blockchain, there are two additional rules:

A block, once valid, is immutable - its transactions are permanently recorded.
And: if there are different versions of the blockchain, the longest one is considered the valid one, because it contains the most accumulated proof of work.

The way proof of work and these rules interact is called a consensus mechanism. Other blockchains use different mechanisms, but the principle is similar: to ensure that all participants agree on the valid state of the blockchain and that manipulation is unattractive.

Let’s assume I wanted to manipulate the blockchain.

I’ve eaten my ice cream, but the cone was a bit soggy. I’m dissatisfied. I want to change the block that contains my transaction of 0.000013 Bitcoin to you - to undo it, so you’re left covering the cost of the ice cream.

We already know I can’t change a valid block after the fact. So I would have to create a new valid block. And since the longest chain is always the valid one, I would also have to recreate all the blocks that came after it - containing all the transactions that happened after our ice cream purchase.

That would require an enormous amount of additional proof of work.

Forgery would therefore be extremely expensive - so expensive that I would spend far more on computational power than my 0.000013 Bitcoin are worth.

That is the blockchain in its basic form.

I’ve left out a few details - for example, that I have to sign my transaction with a cryptographic key to prove that I initiated it. This is also why cryptocurrencies are called cryptocurrencies - not because they are inherently hidden (crypto comes from the Greek kryptos, meaning hidden or secret), but because cryptography, mathematical encryption, is a central part of their underlying algorithms.

Another detail that isn’t essential here: creating new blocks - so-called mining - is itself rewarded with Bitcoin.

And we’ll ignore the practical problems for now - for example, that using Bitcoin can be more expensive than using a bank. Or that a Bitcoin transaction can take so long that the ice cream would have melted by the time it is confirmed.

For now, let’s just note:

I have bought an ice cream from you - and no bank, no government was involved.

And now, congratulations: you most likely now know more about blockchain than you ever wanted to.

I understand - I feel the same way.

We deserve a break, don’t we? Next week, I’m going to introduce you to a woman of shady origin and questionable reputation. You’ll like her. Stay tuned.