Module 3 · The Idea of Bitcoin / 3.1
A copy is not a payment.
Sending information is easy. Agreeing who can spend is the harder part.
You send a friend a photograph of your dog. A moment later, both of you have it.
That is exactly what you wanted. Digital information is easy to copy.
Now imagine trying the same thing with a payment. You have enough for one loaf of bread, but you send instructions paying both the baker and the miller. They cannot both receive the same spending power.
The file is not the scarce thing.
If you hand over a genuine banknote, you no longer hold that particular note. Sending a digital message does not have the same physical constraint.
But a bitcoin is not a picture of a coin in a file. The network keeps records of which amounts can still be spent. Copying wallet software or transaction data does not create new spendable bitcoin.
The problem is a double spend: two conflicting attempts to spend the same available amount. A working system needs rules that prevent both from becoming accepted payments in the same history.
Authorization is one check. A digital signature can show that an instruction was authorized by the relevant key. It does not, on its own, tell you whether the same available amount has already been spent.
Both recipients can keep the image. Sending it does not remove the original.
One operator can keep the official record.
A conventional payment service can check its records, accept one payment, and reject another that lacks funds. A well-built booking service can similarly reserve its last room without selling it twice. It does not need a person to sort out every collision.
When one organization has ultimate control over the record, we call that arrangement centralized. It can still run thousands of computers in different places. Its computing is distributed, but its control remains concentrated.
Central control can make coordination, customer support, recovery, and compliance easier. It also gives the operator power over access and creates dependence on its decisions, systems, and finances.
So “centralized” is not a synonym for useless or dishonest. It describes where a decision can ultimately be made.
One operator controls all four copies. The machines are distributed; their authority is centralized.
What changes when there is no official bookkeeper?
A decentralized arrangement spreads meaningful control across independent participants. In an open blockchain, people can run software that checks transactions for themselves rather than treating one company’s database as definitive.
That introduces a coordination problem. The baker’s computer might hear one payment first. The miller’s computer might hear the conflicting payment first. Messages take different routes and arrive at different times.
“Accept whichever arrived first” is not enough, because first at one computer need not be first at another. The participants need a procedure for converging on a shared, valid history. That procedure is part of consensus.
Consensus does not mean everyone knows every fact about the world or agrees at every instant. It concerns which permitted updates become the common record under the system’s assumptions.
Why one account, one vote is not enough.
An open system lets newcomers join. If influence came only from counting accounts, one person could create many identities and pretend to be a crowd. This is called a Sybil attack.
The problem is not voting in every possible form. It is treating cheap, unverified identities as independent people. A system needs some basis for influence or participation that an attacker cannot multiply for free.
Bitcoin uses proof of work as part of its answer. Other systems make different choices, including proof of stake. You will learn both, along with the rules they do and do not enforce.
Decentralization also has degrees. Who runs validators, writes widely used software, provides wallet access, or controls an application can matter. Ask which power is distributed, among whom, and how independently—not just how many dots appear on a network diagram.
A familiar coordination problem
The Byzantine Generals Problem studies how participants can reach agreement when some may send misleading information. Researchers developed many agreement protocols before Bitcoin. Different designs make different assumptions about membership, message delays, and faulty participants.
Bitcoin combined existing cryptographic tools with a particular way of ordering a public transaction history in an open network. It did not show that all earlier computer science was impossible, nor that consensus makes outside claims such as a hotel’s room count automatically true.
The idea to keep
A signature can authorize a request. A shared history decides whether its spending power is still available. An open network needs a way to maintain that history despite conflicting messages and fake identities.
That is the problem to keep in mind as we meet the ideas that came before Bitcoin.