The Genesis Block Carries a Headline, and That Is Almost All We Know

The artefact first. Here is the coinbase input script of the only transaction in block 0, pulled from a block explorer while writing this, in hexadecimal exactly as it appears:

04ffff001d0104455468652054696d65732030332f4a616e2f3230303920
4368616e63656c6c6f72206f6e206272696e6b206f66207365636f6e6420
6261696c6f757420666f722062616e6b73

Seventy seven bytes. It decodes into three separate pushes, and the third one is the famous part.

04 ffff001d pushes four bytes, ffff001d, which is the same difficulty target that sits in the block header. 01 04 pushes a single byte with the value 4. Then 45 pushes the next sixty nine bytes, and those sixty nine bytes are ASCII:

The Times 03/Jan/2009 Chancellor on brink of second bailout for banks

That is the message. It is not hidden, it is not encrypted, and it is not in a special field. A coinbase input has no previous output to reference, so the space where a signature would go is free, and the first block put a sentence in it.

What the bytes establish

Exactly one thing, and it is a floor rather than a date.

Whoever built that transaction had the string in front of them. The string contains a date, 03/Jan/2009, and the standard attribution is that it is the front page headline of The Times of London for that day. If the attribution holds, the coinbase could not have been constructed before that newspaper existed.

That is a real and useful constraint, because it is one of a very small number of facts about the earliest period of this history that does not rest on somebody’s memory. It is also considerably less than what the message is usually asked to prove.

The rest of the block, which nobody quotes

The block header records a timestamp of 1231006505, which is 3 January 2009 at 18:15:05 UTC. That figure is an assertion by whoever produced the block rather than an observation, and the rules that keep such an assertion inside a tolerable range are a separate piece. The single transaction pays 5,000,000,000 satoshis to a pay-to-public-key script, and explorers render the corresponding address as 1A1zP1eP5QGefi2DMPTfTL5SLmv7DivfNa. Because there is only one transaction, the block’s merkle root is that transaction’s own identifier, 4a5e1e4b...fdeda33b.

And then there is the thing that genuinely is odd, which almost nobody mentions in the same breath as the headline: that fifty coin output cannot be spent. Not by convention, and not because a key was lost. Because the software refuses to process the block’s transactions at all. From src/validation.cpp in Bitcoin Core, in the function that connects a block to the chain state:

// Special case for the genesis block, skipping connection of its transactions
// (its coinbase is unspendable)
if (block_hash == params.GetConsensus().hashGenesisBlock) {

The genesis block is short circuited before its outputs are ever added to the set of spendable coins. The consequence is that the supply everyone quotes is slightly wrong, by fifty, permanently.

Is that deliberate? The comment in the modern source says the coinbase is unspendable, which is a statement of the effect rather than of the intent. The honest reading is that the original implementation special cased the first block because the first block has no parent to validate against, and this fell out of that. Whether it was designed as a gesture or emerged as a quirk is not something the code tells you, and the difference is exactly the kind of thing that gets narrated with confidence by people who did not open the file.

The six day gap

Block 0 is timestamped 3 January 2009 at 18:15:05 UTC. Block 1 is timestamped 9 January 2009 at 02:54:25 UTC. Both numbers are on the chain and either can be checked in a minute.

That is roughly five and a half days of nothing between the first block and the second, on a network whose difficulty at the time was the minimum the protocol allows. It is a genuinely unusual gap, and it is contested. The explanations in circulation include a deliberate pause, a period of testing, a timestamp chosen rather than observed because the genesis block is hard coded into the client rather than mined by it, and simple non operation of the software in between.

I am not going to pick one. None of them is established by a document, all of them are consistent with the two numbers, and the block timestamps are in any case assertions made by whoever produced the block rather than observations made by anything outside it. What is worth saying plainly is that a gap this large is the sort of detail that ought to make a reader more careful about the certainty of the surrounding story, not less.

The load the sentence cannot carry

Sixty nine bytes are routinely made to establish a motive, a political programme, a monetary philosophy and a founding grievance. They do not.

What can be said: somebody chose a headline about bank rescues, at a moment when the banking system was extremely visibly in trouble, and put it where the first block’s spare bytes were. What cannot be said is why. The bytes are compatible with a timestamp proof and nothing more, with a comment on the news, with a joke, with a signal to a specific reader, and with several motives at once. Choosing among those requires evidence that is not in the block, and the evidence usually offered is a retrofit: the interpretation is read back onto the bytes from things said years later by other people.

The same discipline applies to the paper. The whitepaper is an engineering document that argues about double spending, incentives and network assumptions. It does not contain a monetary manifesto, and a headline in a coinbase field does not supply one on its behalf. Treating the two together as a declaration of intent is the single most common move in writing about this period and it is not supported by either document.

What was expanding, what was contracting

The interesting reading here is not about markets, because there was no market. It is about evidence.

What was expanding, from block 0 forward, was the amount of history that would be recorded in a form nobody could quietly revise. Everything in this period that we know from forums, mailing lists and recollection is subject to editing, deletion, hosting decisions and memory. The chain is the exception. Sixty nine bytes were committed on 3 January 2009 and they are still byte for byte checkable by anybody who wants to check them, which is why this piece could open with hexadecimal instead of with a paraphrase.

What was contracting, at the same instant and for the same reason, was the space for a certain kind of claim. Not immediately, and not visibly, but from that block onward there existed a class of assertion about crypto history that could be settled rather than argued.

Who could tell at the time? Nobody, and that is not a criticism. On 3 January 2009 there was one block, one participant capable of producing another, and no observers. The property that makes the genesis block valuable to a historian is not the message. It is that the message is in a structure that cannot be quietly changed, and the people who did the most interpreting of that message have generally been the least interested in the structure.