The Release Announcement Read as a Product Launch

The message is in the public archive of the Cryptography mailing list at metzdowd.com, in the January 2009 folder. Subject line: “Bitcoin v0.1 released”. Sender: Satoshi Nakamoto, at a vistomail.com address. Timestamp: Thursday 8 January 2009, 14:27:40 Eastern.

It opens:

Announcing the first release of Bitcoin, a new electronic cash system that uses a peer-to-peer network to prevent double-spending. It’s completely decentralized with no server or central authority.

Two sentences, and they are the only two in the message that anybody quotes. What follows them is a release note, and reading the whole thing as a release note is the point of this piece, because the ordinary parts are doing more historical work than the famous parts.

What is actually in the document

In order, as it appears.

A pointer to a website for screenshots. A download link, to a .rar archive hosted on SourceForge. A platform note:

Windows only for now. Open source C++ code is included.

Then three bullets of installation instructions that would fit on a support page: unpack the files into a directory, run BITCOIN.EXE, it automatically connects to other nodes.

Then a request for help, phrased the way any small project phrases it:

If you can keep a node running that accepts incoming connections, you’ll really be helping the network a lot. Port 8333 on your firewall needs to be open to receive incoming connections.

Then a warning:

The software is still alpha and experimental. There’s no guarantee the system’s state won’t have to be restarted at some point if it becomes necessary, although I’ve done everything I can to build in extensibility and versioning.

Then the mining instructions, the two available methods of sending a payment, and the issuance schedule. Then a sign off.

That is the whole document. A file format, an operating system limitation, a firewall port, and an admission that the ledger might have to be thrown away and started again.

The tone contains none of the significance

Compare that to how the same event is described now, and the gap is the finding.

The announcement makes no claim to be founding anything. It does not mention banks, monetary policy, inflation, the financial system, sovereignty, freedom or the state. It contains no argument about why the thing should exist beyond the technical one in its first sentence: this prevents double-spending without a central authority. It offers no vision of adoption. It asks for people to open a port.

The most striking omission is confidence. “There’s no guarantee the system’s state won’t have to be restarted at some point” is not something you write about an immutable ledger. It is something you write about an alpha, because in January 2009 that is what it was, and the person who built it said so in the announcement in plain terms.

The .rar deserves a moment too. Not a .zip, not an installer, not a signed package: a proprietary archive format, on SourceForge, Windows only, with the source included in the same bundle. That is the distribution posture of a personal project, and it is completely at odds with the way the release is now narrated.

The parts that were consequential and were stated flatly

Two decisions in that message have shaped everything since, and both are delivered without emphasis, in the register of a specification.

The first:

Total circulation will be 21,000,000 coins. It’ll be distributed to network nodes when they make blocks, with the amount cut in half every 4 years.

Followed by the schedule, written out as four lines and an “etc…”. The single most quoted parameter in this entire history is announced in a sentence, with the arithmetic shown, and no argument at all for why that number or that curve.

The second is about what happens afterwards:

When that runs out, the system can support transaction fees if needed. It’s based on open market competition, and there will probably always be nodes willing to process transactions for free.

“If needed.” “Probably.” The long-run security question that the field has argued about ever since, what pays for the work once issuance stops, is addressed in two sentences and left as an expectation rather than as a design. That is not a criticism of the document. It is a release note, and release notes are allowed to leave things open. It is a criticism of every later account that reads the economics as fully worked out from the start, because the announcement plainly does not present them that way.

What happened next, as far as the archive shows

The angle here has to be strict, because reception is exactly where this material gets invented. So: only what is in the archive index.

The January 2009 folder of that mailing list contains eight messages under the subject “Bitcoin v0.1 released”. Three of them are from Satoshi Nakamoto, including the original. The others are from Hal Finney, who posted twice, and from Jonathan Thornburg, Bill Frantz, and a poster listed as dan at geer.org.

Eight messages. Five distinct participants. That is what the record shows for that list in that month, and it is all this piece will say about it. It is not a claim about how many people read the announcement, downloaded the archive, ran the program or thought about it, none of which is knowable. It is a count of messages in one archive, which is a different and much smaller fact, and the difference is the whole discipline.

Read against the paper

The whitepaper and this announcement are two documents with two audiences and two jobs, published about ten weeks apart, and running them together is how a modest engineering claim acquired a political one.

The paper argues. It has numbered sections, a threat model, a probability calculation and a conclusion. The announcement ships. It has a file, a port number and a warning about the alpha. Neither document says what the project is for in the grand sense, and the reason people believe both of them do is that the grand sense was supplied later, by other people, and read backwards onto whatever text was available. The sixty nine bytes in the first block carry the same retrofitted load for the same reason.

What was expanding, what was contracting

Answer the site’s question honestly for 8 January 2009 and the answer is almost comic.

Nothing was expanding. There was one implementation, on one operating system, in one archive format, with a network that had produced a handful of blocks and no second use. The capability had just come into existence and had not yet gone anywhere.

What was contracting was the set of assumptions that had governed every previous attempt in this history, and it contracted to zero on that day without anyone noticing. Before the release, every design for digital cash required somebody to operate it: an issuer, a bank, a company with a charter and an address. After the release, there was a running example that did not, however alpha it was and however few people ran it.

Who could tell? Effectively nobody, and the archive is the evidence. Eight messages, five people, on a specialist mailing list whose members were about as well equipped to evaluate the claim as any group on earth. The most credentialed possible audience read a competent release note, asked reasonable questions, and mostly moved on. The message immediately following the announcement in that archive is about weaknesses in hash functions.

That is not a failure of judgement and it should not be written as one. The document did not announce a threshold crossing. It announced version 0.1, Windows only, port 8333, still alpha. The significance was added afterwards, and measuring the distance between the announcement and the legend is the most useful thing this particular primary source can be made to do.