Nick Szabo is probably not Satoshi Nakamoto
But the reclusive cryptographer has many suspicious parallels
A New York Times article recently announced that they’d proven that Adam Back is Satoshi Nakamoto, based on writing stylometry and other similarities. Some people don’t believe this, either because they’re not convinced by the stylometry analysis, or because they prefer some other theory for who Satoshi is.
Nick Szabo is one favorite candidate, among those other theories. A different New York Times piece, back in 2015, suggested Nick Szabo might be Satoshi. This article will go over the evidence for and against Nick Szabo as Satoshi, and make a few comparisons to the Adam Back theory.
There are actually dozens of theories for who Satoshi Nakamoto was, but there are 4 men who are usual suspects because they were pioneers of digital cash projects, before Bitcoin: Hal Finney, Adam Back, Wei Dai, and Nick Szabo.
Satoshi sent e-mails to Adam Back, Wei Dai, and Hal Finney.
There are no known e-mails between Satoshi and Nick Szabo. Could that be because they are the same person?
In 2005, Szabo proposed something called Bitgold, with the idea of creating a scarce cryptographic currency that sounds an awful lot like Bitcoin.
In one (now deleted) version of his Bitgold proposal, Szabo suggested using Adam Back’s hashcash, just as Bitcoin ultimately used hashcash:
The basic outline of the system I described was as follows:
(1) bits are generated using an algorithm of known computational cost, e.g. computing an n-bit hash collision (as proposed by Rivest/Shamir for MicroMint and implemented Adam Back with hashcash). In general there are challenge bits input into a benchmark function, which produces solution bits. The desiderata of a benchmark function is fundamentally different from that of a cryptographic one-way function, and I have developed a new formalization to reflect this. Anyone interested in studying such functions, please contact me.
(2) the bits which are the solution to the challenge are added to a multiparty hash chain (i.e., they are securely timestamped), then converted into the next challenge.
(3) Representations of ownership of these solution bits are stored in a public manner, e.g. in a distributed property title registry I also originally suggested (both off the list and in an impractical but privacy-protecting form I described on the list) that publically known levels of wealth can be represented by a system of publicly shared books. However, I think using this approach and discarding the solution bits raises several unecessary problems.
(4) Titles to the bits are traded, achieving a market price based on their estimated original cost of computation and continuing scarcity versus the demand for a secure long term store of value.
That sounds a lot like Bitcoin, except that step #4 is different.
Szabo wrote another post in April 2008, which talks about implementing this idea:
“The basic idea of bit gold is for “bit gold miners” to set their computers to solving computationally intensive mathematical puzzles, then to publish the solutions to these puzzles in secure public registries, giving them unique title to these provably scarce and securely timestamped bits. These titles to timestamped bits will be more secure and provably scarce than precious metals, collectibles, and any other objects that have ever been used as money.”
In the comments section, Szabo writes:
“Bit gold is always backed by the scarcity of solved puzzles, but the puzzle solutions in a prior week are not of equal supply, and thus not of equal value, to the solutions of the next week. If all of a sudden ten times as many puzzles are being solved in week 52 as in week 51, they will be priced at about 10% of the value of the puzzles in week 51, and it will take 10 times as many puzzles to create the standard pool. Supply of pools is based on weeks: week 2 doubles the supply of pools over week 1, but week 100 can only increase the overall supply of pools by 1%, regardless of how many puzzles are solved”
That sounds a bit like the way that Bitcoin adjusts difficulty as more miners join the network.
Szabo also wrote:
“it would greatly benefit from a demonstration, an experimental market … Anybody want to help me code one up?”
Did he make Bitcoin himself? Did he get the help he asked for, and make it with another person? This is about 7 months before Satoshi shares the Bitcoin code.
Note there is probably a timeline conflict here — in an early forum message, Satoshi actually said he’d been working on coding Bitcoin for a year and a half, so you might expect he started in spring of 2017, not spring of 2018. In a private email in July 2009, Satoshi says he “needs a break from it after 18 months development.” That would put the start a bit different, but now we’re thinking end of 2007 or beginning of 2008. Either way, Satoshi would have started before April 2008.
For comparison, there is a 3 year long gap in Adam Back’s online activity, starting in late 2007.
It is, of course, hard to prove when the development started. One file in the first Bitcoin code gave a version date of April 16, 2008. That file is part of the UI, so not likely the first thing that would have been written. It’s also unclear if that was the date the file was created or updated.
In an early forum post, Satoshi wrote:
“I actually did this kind of backwards. I had to write all the code before I could convince myself that I could solve every problem, then I wrote the paper.”
That sounds different than Szabo’s approach, where he’d already written about the Bitgold idea for years and then started asking for coding help.
That top line of Szabo’s Bitgold blog posts say they were from December 27, 2008 (i.e. after Satoshi’s whitepaper), but those dates have changed because he reposted them. The Bitgold post was from 2005, the “Bitgold markets” post was from April 2008, and then both blog posts were reposted in December 2008
To summarize the timeline:
Szabo proposed Bitgold in December 2005
Possible start of Satoshi coding Bitcoin: 2007
Szabo asked for Bitgold coding help in April 2008
Satoshi had a Bitcoin white paper draft ready by August 2008
Satoshi had some early Bitcoin code ready by September 2008
Satoshi publicly posted the white paper October 2008
Szabo reposted the Bitgold posts in December 2008
Satoshi launched the Bitcoin software in January 2009
The next thing to notice is a gap in Szabo’s blogging history
Normally Szabo was putting out a few blog posts every month. But then, on August 20th, 2008, he announced that he’s going to be posting “reruns” of old blog posts.
And then every post is a rerun, for a few months.
Okay, um… August 20th is the day that Satoshi e-mailed Adam Back. The reruns post is actually 4 hours after the e-mail from Satoshi to Adam Back.
August 18th is when Satoshi registered bitcoin.org.
2 days earlier, there was an empty directory that happened to show up on Szabo’s website:
Was he working on his alias?
As we dig more, we’ll find that Satoshi originally intended to call Bitcoin, “Ecash”, and he likely registered the Ecash domain on July 20th, 2008. He likely also registered Netcoin.org on August 17th, 2008.
Is Szabo’s absence of blog posts abnormal?
It looks abnormal, compared to his prior posting history, but not relative to his future posts. You could argue that he’d lost interest in blogging so often, first tried reruns, and then never really regained the enthusiasm. He also graduated from law school in 2007 and could have gotten more busy with work or other things.
There is still that time coincidence with the first Satoshi e-mails, and then Szabo posts the two Bitgold posts as two of his reruns, in late December 2008. Those are about a week before Bitcoin launches in January. Why was he so busy, from August to March? Releasing Bitcoin, maybe?
I should note it’s also possible Szabo stopped blogging in August because he got an e-mail from the real Satoshi some time around August 20th, the same time that Adam Back did, and he got interested and started helping.
In theory, Adam Back could have even forwarded that message to Szabo, on August 20th — Back would have had 4 hours to do so (this theory would assume that Satoshi is neither of these two men).
So we have 4 possible theories here:
• Nick Szabo, working alone, was Satoshi.
• Szabo worked with another programmer to release Bitgold as Bitcoin.
• Szabo was not Satoshi but he was contacted early and got interested.
• All the connections above are meaningless coincidences.
To summarize the timeline again:
Szabo proposed Bitgold in December 2005
Possible start of Satoshi coding Bitcoin: 2007
Szabo asked for Bitgold coding help in April 2008
Someone registers Ecash.org July 20, 2008
Someone registers Netcoin.org August 17, 2008
Satoshi registers Bitcoin.org August 18, 2008
Satoshi emails Adam Back August 20, 2008
Szabo announces reruns August 20, 2008
Satoshi posted the white paper October 31, 2008
Szabo reposted the Bitgold posts December 27, 2008
Satoshi starts Bitcoin mining January 9, 2009
Szabo didn’t react much to the launch of Bitcoin
Some of Szabo’s rerun posts had no mention of Bitgold and no obvious connection to Bitcoin, but 2 of Szabo’s rerun posts were about Bitgold, and a 3rd mentioned it less directly. A few of the posts also mentioned related concepts — his January 31, 2009 post was on cryptographic timestamping, which was an essential component of Bitcoin.
Some people note that one of his rerun posts, in late January 2009, was kind of a non-sequitor, showing cars trying to drive past toll barriers for busses and getting stuck. The text of the post is minimal:
And then it shares this video:
If you think about this a little harder, there might be a connection to Bitcoin. One key problem of digital money that Bitcoin solved was preventing double spending. This video is about drivers failing to drive through a checkpoint twice.
Some people have speculated that Szabo posted this to brag that the Bitcoin network was working against bad actors. That might be a bit of a stretch, I don’t know. Some of Szabo’s other rerun posts had nothing to do with Bitgold.
Szabo started writing new blog posts again in April, 2009.
While Szabo was very interested in creating Bitgold, he never actually implemented it. Perhaps he made Bitcoin, and that was Bitgold, or perhaps he accepted that someone else solved the problem and moved on.
Nick didn’t seem that excited when Bitcoin came out. He didn’t blog about Bitcoin until May 2009, a month after his reruns break ended. At that point it only gets a brief mention. The post was called “Liar-resistant government”:
I have extensively explored the technological codification of certain aspects of law — law that is “smart” as in “smart weapons”. These could implement in digital protocol important parts of law that are now processed by the mind via legal language or through physical enforcement. My focus has been on smart contracts, smart property, and other such technological reifications of private law. Here I explore the dangerous territory of extending these ideas to public law, especially to governmental forms and legal procedures. Officials of a wide variety of governments have too often over the course of history covered up or falsified evidence, destroyed or forged public records, and introduced other lies into legal and political processes. We need to protect future legal procedures and other governmental operations from such abuses.
The canonical problem explored by computer scientists in designing these protocols, the Byzantine Generals Problem, is itself an exercise about liars in government.
He went into more detail about something called the Byzantine general’s problem, which is basically a question about how to get messages across a network where some participants want to communicate while some nodes might be hostile. Near the end of the post, he briefly mentions Bitcoin solving this problem (Bitcoin solved this with the Blockchain and the longest chain rule):
Besides the obvious real property titles, domain names, and so, on, these registries could securely record and transfer the shares of a corporation. Bit gold, my sketch of an electronic currency that minimally relies on trust in any one person or organization, achieves this minimal vulnerability by using secure property titles. Satoshi Nakamoto has implemented BitCoin which very similarly uses a dense Byzantine fault tolerant peer-to-peer network and and cryptographic hash chains to ensure the integrity of a currency.
As far as I can tell, he doesn’t say anything more about Bitcoin until May 2011, in a post called “Bitcoin, what took ye so long?”
The short answer about why it took so long is that the bit gold/Bitcoin ideas were nowhere remotely close to being as obvious gwern suggests. They required a very substantial amount of unconventional thought, not just about the security technologies gwern lists (and I’m afraid the list misses one of the biggest ones, Byzantine-resilient peer-to-peer replication), but about how to choose and put together these protocols and why. Bitcoin is not a list of cryptographic features, it’s a very complex system of interacting mathematics and protocols in pursuit of what was a very unpopular goal.
While the security technology is very far from trivial, the “why” was by far the biggest stumbling block — nearly everybody who heard the general idea thought it was a very bad idea. Myself, Wei Dai, and Hal Finney were the only people I know of who liked the idea (or in Dai’s case his related idea) enough to pursue it to any significant extent until Nakamoto (assuming Nakamoto is not really Finney or Dai). Only Finney (RPOW) and Nakamoto were motivated enough to actually implement such a scheme.
Szabo also described a few ways that Bitcoin differed from Bitgold:
(3) Nakamoto improved a significant security shortcoming that my design had, namely by requiring a proof-of-work to be a node in the Byzantine-resilient peer-to-peer system to lessen the threat of an untrustworthy party controlling the majority of nodes and thus corrupting a number of important security features. Yet another feature obvious in hindsight, quite non-obvious in foresight.
(4) Instead of my automated market to account for the fact that the difficulty of puzzles can often radically change based on hardware improvements and cryptographic breakthroughs (i.e. discovering algorithms that can solve proofs-of-work faster), and the unpredictability of demand, Nakamoto designed a Byzantine-agreed algorithm adjusting the difficulty of puzzles. I can’t decide whether this aspect of Bitcoin is more feature or more bug, but it does make it simpler.
You could say it’s weird that Nick was so excited about this Bitgold idea but that he didn’t get more involved with the project and didn’t talk about it much. You could argue that shows Nick was Satoshi, that he was involved with the project from early on, but hiding his interest in it.
But, then again, maybe it wasn’t as obvious that Bitcoin would become so big, and the invention only looks so meaningful in hindsight.
There’s one other competing point here: Szabo didn’t even like the term cryptocurrency at first. In his 2011 post, Nick commented:
And what’s up with the term “cryptocurrency”? What is it supposed to mean? Cryptography is used to protect payments systems as radically different as credit cards, Chaumian digital cash, and Bitcoin. The term encourages the popular but profoundly naive view of Bitcoin as merely another form of digital cash.
When Satoshi first heard the word in 2009, he asked:
Someone came up with the word “cryptocurrency”… maybe it’s a word we
should use when describing Bitcoin, do you like it?
Satoshi also did not refer to Szabo
The Bitcoin whitepaper included references to Adam Back’s hashcash and to Wei Dai b-money, but it did not mention Nick Szabo’s Bitgold.
Hal Finney mentioned the Bitgold idea to Satoshi, in his very first response on November 8th, 2008. Satoshi responded to many parts of Hal’s message, but not to the part about Nick Szabo and Bitgold.
The first reference I could find is on Bitgold’s alpha release page. I see a reference to Bitgold on the March 3rd archive of the page, but not on the January 31st archive.
Satoshi also credited Bitgold the next year, in a July 2010 forum post.
Satoshi’s avoidance of mentioning Szabo could be seen as hiding the fact that he is Szabo. But it could also just show little knowledge of or interest in Szabo’s work.
Satoshi’s birthday
When Satoshi signed up for the P2P foundation, he listed his birthday as April 5th, 1975. Bitcoiners often speculate that this refers to the date that the US government banned ownership of gold (April 5th, 1933) and the year it became legal to own gold again (1975).
On August 8, 2008, a couple weeks before the reruns begin, Szabo referred to FDR’s ban on gold ownership:
As a collectible it is subject to less political risk than gold (recall FDR’s confiscation of gold in the U.S., for example)
That’s interesting but, of course, anyone interested in currency and gold would be aware of the same thing. The New York Times showed that Adam Back had mentioned the same gold ownership ban in his writing.
Wikipedia claims that Szabo himself was born April 5th, 1964. I don’t think that is correct, it might just have been put there by some Bitcoiners — the public records I found suggested Szabo was actually born in May, 1965 (possibly May 5th, 1965).
Background and ideology
Szabo’s father was from Hungary, and participated in the 1956 Hungarian uprising against Soviet occupation before immigrating to the US. One source says Nick grew up in California and went to high school in Los Gatos. He went to college in Seattle, getting a degree in computer science from the University of Washington, in 1989. He worked for JPL, in southern California. Near the age of 40, he went back to school to get a law degree from George Washington University, in Washington DC., where he finished school in 2007.
Wikipedia says he holds an honorary professorship, but Szabo denies being a college professor, or ever having been one.
Where was he in 2008–2011, the key years for Bitcoin’s creation?
I’m not sure — Szabo is reclusive and hard to gather many details on.
His blog posts from then give timestamps in Pacific time, but the public records I found for him put him in Washington DC from 2003 to September 2010, with an address 5 minutes away from George Washington University law school.
Nick shows up again in California, in Menlo Park, on April 2013. At the beginning of 2014, he joined Vaurum, a Bitcoin start-up in Palo Alto, working on building a better Bitcoin exchange.
I couldn’t find any records for the gap between 2010 and 2013, nor what exactly he was working on between law school and the job at Vaurum.
So, Nick’s Bitgold idea is very similar to Bitcoin, his blogging absence is suspiciously timed, and his ideology matches the project. That all makes him a candidate for being Satoshi, but it’s hardly proof.
If we wanted a match, I think we’d need a lot more than that. We’d want to know where Nick was and what he was doing, from 2007 to 2011. Since that’s hard to establish, the next best thing we can try is to match Satoshi’s writing with Nick’s.
Szabo doesn’t spell like Satoshi
If we can’t ID Nick as Satoshi by timeline, then we have to match his words. In this case, there are hits and misses.
His Bitgold proposal is double spaced, as are most of the blog posts I checked (you have to look at the page source to see this in HTML).
That’s a match, but he does not match Satoshi’s British spelling — Szabo always uses American, not British spellings of words. Crawling his Twitter account:
Szabo says “color” but never “colour”.
He says “gray” but never “grey”.
He says “realize” but never “realise”.
He says “apartment” but never “flat”.
There’s only one instance of “bloody” in the British context.
He never says “lad”.
He says “mate” once, but it’s in a weird, playful context.
Szabo joined Twitter in 2014, so theoretically he could have switched to American spelling. I didn’t search all of his earlier blog posts before Twitter to see if those were different, but I’d assume they aren’t — he grew up in America.
In theory, Satoshi’s British spelling could also be misdirection — you could just use a spellchecker to turn common words into British spelling. That would fix something like “gray” vs “grey”, but it wouldn’t help catch things like “cheque” vs “check”. It also wouldn’t direct you towards more subtle choices, like referring to “apartments” as “flats”, like Satoshi once did.
Cullen Hoback suggested looking at the words “sweet” and “retarded” as spontaneous utterances of Satoshi’s which might reflect his character.
Szabo doesn’t really use the word “sweet” in any of the ways
You can find better matches here with different candidates, like Adam Back or Peter Todd.
Szabo didn’t start using the word “retarded” on Twitter until 2025, but it’s not hard to imagine him using it privately, earlier than that:
These are interesting tells, but I’m not sure that they’re definitive.
Other people have noted Satoshi’s use of the word “hosed”.
I found one instance of Szabo saying this:
I couldn’t find this on Twitter, for Back or for Todd. I didn’t search other archives of their text.
Stylometry programs tend to look at more complex aspects of language like sentence structure or more systematic vocabulary searches. In theory, these should be more accurate than just looking at a few words. In practice, they might not work if Satoshi knew about these programs and tried to obfuscate his style.
Some stylometry studies claim there was a match
There are several linguistic studies that concluded that Szabo’s writing is a close match to Satoshi’s.
Anonymous blogger Skye Grey did 2 analyses which I can easily reproduce. The first analysis keyed in on various phrases that Szabo used which were also found in the Bitcoin white paper:
We can quickly check some of these on Twitter. Szabo frequently makes the first mistake:
But Adam Back also usually gets this wrong and Peter Todd often gets this wrong, too.
Szabo does frequently specify a level of something, but Adam Back also uses this a lot, and I’d imagine this is just a really normal part of the English language.
Szabo never says “preclude” on Twitter, but Adam Back does and Peter Todd does.
None of these 3 people ever start Twitter sentences with “it should be noted” or say “for our purposes” or “can be characterized”.
Nick Szabo does say “characterized” 4 times, Todd says it 4 times, Adam Back says it once.
Maybe we should be comparing their papers to the Bitcoin whitepaper, not their tweets.
None of these 3 phrases are in Adam Back’s hashcash paper.
But none are in Szabo’s Bitgold blog post, either.
Okay… I guess we need a larger corpus. Here’s a folder with 42 Szabo essays.
None of these say “can be characterized”
None of them say “it should be noted”
3 of the 42 essays say “for our purposes”
None of them use “precludes” or “preclude”.
Okay, let’s try Szabo’s blog:
Of all his blog posts, I see only 1 with “for our purposes”
2 with “it should be noted”
1 with “can be characterized” (actually zero exact matches, but I counted “can also be characterized” to be generous. There’s also 1 instance of “most often characterized”)
None of Szabo’s blog posts use the word “precludes”.
I think the total number of posts I’m searching here is 351. That’s a pretty weak signal, these aren’t common phrases.
Get Peter Miller’s stories in your inbox
Join Medium for free to get updates from this writer.
For comparison, I see zero of these phrases on Adam Back’s website.
I downloaded 13 of his 22 papers (the others weren’t publicly accessible).
2 of the 13 papers started a sentence with “it should be noted”
None of them said “can be characterized” (only one paper used the word “characterized” at all).
None of them said “for our purposes”. (his 1995 thesis did say “for the purposes” 5 times)
1 of his papers says “precludes”.
In summary, Skye Grey’s phrases aren’t a strong identifier for Szabo.
The only 2 proposed phrases that distinguish Szabo’s writing are “can be characterized” and “for our purposes”. Adam Back doesn’t use these phrases, but this isn’t really a fair comparison; we’re comparing 13 papers vs 351 blog posts.
The word “precludes” seems to actually point the wrong way — in the writing I searched, I could not find it used by Szabo, but I could find it used by Adam Back.
Writing histogram analysis
Skye Grey did a second analysis where, instead of looking at specific phrases, he looked at frequency of word lengths and characters:
I took extensive writing samples from Wei Dai, Adam Back, Hal Finney, David Chaum, and Nick Szabo. Samples were from 5k to 40k word long. I then computed histograms of word length frequency and character frequency, and compared these histograms with that of the original Bitcoin whitepaper. Here are my results. Units are arbitrary (smaller scores mean closer histograms).
Word length distribution
Diff Nick Szabo / Bitcoin: 0.160
Diff Wei Dai / Bitcoin: 0.241
Diff David Chaum / Bitcoin: 0.257
Diff Adam Back / Bitcoin: 0.337
Diff Hal Finney / Bitcoin: 0.510Character frequency distribution
Diff Nick Szabo / Bitcoin: 0.191
Diff Wei Dai / Bitcoin: 0.208
Diff David Chaum / Bitcoin: 0.228
Diff Hal Finney / Bitcoin: 0.284
Diff Adam Back / Bitcoin: 0.342
I was unsure of what text Skye Grey used in this analysis, so I downloaded all of the cryptography list that Bitcoin was first announced on. Adam Back posted on that list but Nick Szabo did not, so I gathered some of Szabo’s writing separately. I built histograms for character frequency and word length distribution for everyone on the cryptography list.
Neither Nick Szabo or Adam Back were anywhere near the top of this comparison. Out of about 1,000 people, Nick Szabo was #92 and Adam Back was #121.
The #1 match was fairly close to Satoshi’s writing, at least by this metric. He’s a more obscure candidate that you might not have heard of. I’ve seen him flagged as first place in 2 other stylometry studies, but I don’t think I’ll name him here unless I see more proof that he might be Satoshi, or unless I can study stylometry a lot more closely.
I have to conclude that this technique does not work to identify Nick Szabo as Satoshi.
Aston university study
A third study was done at Aston university:
To begin the search, scientists chose 11 personalities closely related to Bitcoin. The list was narrowed down to the following people: Dorian S. Nakamoto, Vili Lehdonvirta, Michael Clear, Shinichi Mochizuki, Gavin Andresen, Nick Szabo, Jed McCaleb, Dustin D. Trammel, Hal Finney, Wei Dai and the Neal King-Vladimir Oksman-Charles Bry trio.
Dr. Jack Grieve, the leader of the “Project Bitcoin” study, stated in his press release on the matter:
“The number of linguistic similarities between Mr. Szabo’s writing and the Bitcoin paper is uncanny, none of the other possible authors were anywhere near as good of a match. We are pretty confident that out of the primary suspects Nick Szabo is the main author of the paper, though we can’t rule out the possibility that others contributed.”
Major similarities found in both the texts of Satoshi Nakamoto and Nick Szabo blog entries include:
“… the phrases ‘chain of…’, ‘trusted third parties’, ‘for our purposes’, ‘need for…’, ‘still’, ‘of course’, ‘as long as’, ‘such as’ and ‘only’ numerous times, contractions, commas before ‘and’ and ‘but’, hyphenation, ‘-ly’ adverbs, the pronouns ‘we’ and ‘our’ in papers by a single author; fragmented sentences following colons and reflexive (-self) pronouns.”
It sounds like this search was done manually — Dr Grieve had a group of his students read and compare all the text.
This sounds too time consuming to reproduce manually, and their search did not include some of the favorite modern Satoshi candidates, like Adam Back or Peter Todd.
This could probably be automated, if someone has the time and the text, but I have not yet tried to reproduce this, and I’m not going to comment on whether they got this correct.
Perhaps my opinion would change if I could reproduce this, but I currently think that the stylometry pointing to Szabo is weak, and it sounds less impressive than what the New York Times did pointing to Adam Back.
The New York Times composed a list of all words without synonyms, then found that Adam Back shared the most with Satoshi. Then, they searched for:
• people who double space
• people who used British spelling
• people who mixed up “its” and “it’s”
• people who have trouble using hyphens correctly
• people who finish sentences with “also”
• people who say “bugfix” as one word
(They also mentioned spacing in the words “halfway” and “downside”, but I didn’t parse which way they were suggesting Satoshi used those)
You can search for many of these writing quirks on Twitter. For instance, Adam Back frequently ends a sentence with “also”. Nick Szabo never does that, and neither does Peter Todd. I did find two instances where Len Sassaman did this, but it’s rare for him.
Adam Back regularly mixes up “its” and “it’s”. Nick Szabo never did that, in the tweets I checked. Peter Todd almost never does, I found one mistake among dozens of tweets I checked. Len Sassaman almost never does, I also found one mistake.
I think the NYT argument sounds good, but if the key evidence is putting spaces in words, there’s at least one case where Satoshi doesn’t match Adam Back — Satoshi always says “cannot” as opposed to “can not”. Back mixes between those two. Szabo always says “cannot”. Todd almost always says “cannot”. Sassaman always says “cannot”.
That opens up the possibility that the NYT found only those cases where Satoshi’s spacing matches Back, ignored the cases where it does not match, and created a filter that looks stronger than it really is.
One way to test this might be to scan for all possible words which could have a space in the middle and count how many match. Another approach might be to do a sort of “sensitivity analysis”, where you see how many of the NYT’s filters you can remove before their arguments falls apart. Like, if you imagined that Satoshi was faking his British spelling, would your other arguments still be sufficient to identify Adam Back as the only possible candidate on the mailing list?
Counter-arguments against Szabo as Satoshi
Cullen Hoback wrote a thread on why Satoshi isn’t Nick Szabo:
Cullen offered a bulleted list with ~16 points. I numbered his bullets for convenience, and will evaluate each of them:
1. When Satoshi reached out to Adam Back, he was unaware of Wei Dai’s b-money. Cryptographers working on the digital cash problem in the late ’90s would almost certainly have known about b-money. So, if Szabo were Satoshi, this would have been deliberate misdirection.
This is true: Szabo should have known about b-money and this would be misdirection. That said, another competing theory is that Adam Back himself is Satoshi, so the level of deception for Szabo lying here is less than if Back was literally just e-mailing himself or if Back fabricated the e-mails after the fact.
Also, the hashcash paper literally mentions and cites b-money, so you could argue that anyone who read the paper should know about it.
Both Back and Szabo would have to be lying here, so you’re just picking which lie you prefer.
You could say this wasn’t deception by naming a more obscure candidate who didn’t know about b-money (like Cullen’s favorite candidate, Peter Todd, or perhaps some other obscure choice), and assuming they didn’t read that part of the hashcash paper.
2. Initially, Satoshi referred to Bitcoin as “e-cash.” The first centralized digital currency, launched by David Chaum, was also called Ecash. Nick Szabo worked for David Chaum, so why would he describe his own invention in the same way his boss did?
This seems kind of important — Nick talked about launching Bitgold for year after year, then begged for help coding it, and then he wants to launch Ecash? And then he considers Netcoin, before finally changing his mind to make it Bitcoin?
Okay, if Szabo was Satoshi, he can’t call it Bitgold because then it would immediately point to him, so he has to call it something else. Maybe the first thought would be that Bitcoin is too close to Bitgold? It’s hard for me to imagine the psychology here.
It’s also hard to imagine how this differs from all the other cypherpunks , who would have also known David Chaum or knew of him. Adam Back would definitely know about Chaum’s Ecash. Perhaps he wouldn’t want that name, either.
Cullen thinks that maybe this distinguishes his favorite candidate, Peter Todd, who was younger, but now you need some theory where Todd knew about enough cryptography stuff (hashcash, timestamping, etc) to invent Bitcoin but also didn’t know about things like b-money, Bitgold, and Ecash.
That’s possible, but I don’t know why it’s more or less likely than, “Satoshi was an older guy who was lying about what he knew”.
3. Satoshi was seemingly unaware of Nick Szabo’s Bit Gold proposal, later adding Szabo’s citation to the Bitcoin website in early 2009. Again, if Szabo were Satoshi, this would be further misdirection at a time when Bitcoin was still struggling to get off the ground.
Yes, Cullen, if Szabo was lying about his identity, he would lie.
4. Satoshi sent the whitepaper to the cryptography mailing list, not the cypherpunk mailing list. This might indicate that Satoshi wasn’t a cypherpunk in the ’90s — possibly someone younger.
Lots of these guys posted on both lists, I don’t get why this is a big deal.
5. Satoshi asked Adam Back to help with citations for the whitepaper. Szabo was well-known within the community and could have asked directly. It seems that Satoshi was hoping to get Adam Back to consider the invention and perhaps help elevate it. Early on, Satoshi struggled to get people on board with his idea.
Sure, it’s possible that Todd was using a pseudonym because he was young and not famous enough to publish. But it’s also possible that Szabo wanted to make a mysterious pseudonym for a decentralized currency, to avoid legal issues from creating a currency, etc. That’s likewise possible for Adam Back or any other Satoshi candidate.
6. If the Bitcoin whitepaper had been published by any known cryptographer, David Chaum told me it would have probably won “paper of the year.” If you have credibility, why not publish it openly? Here’s a rare interview with Wei Dai on the subject:
Szabo wasn’t young, but he may have had plenty of time and energy. He finished law school in 2007, it’s not obvious what he did afterwards. An August 2008 blog post said he had “personal cash flow needs”, so maybe he wasn’t working immediately out of school.
7. Hal and others were openly involved with and mining Bitcoin. Why did Satoshi need to stay anonymous? It would be years before Bitcoin had meaningful value. Writing a paper and implementing a protocol like Bitcoin isn’t a crime.
I don’t know if that’s true — creating a private currency is legal gray area at best, if not risky at worst. E-gold had big legal problems. Bitcoin faced lots of scrutiny when it was used for things like Wikileaks and silk road.
8. The 2005 Bit Gold blog entry on Szabo’s site, which “mysteriously” moved to December 2008, seems less mysterious if you go through his entries on the Wayback Machine. In August 2008, he announced that he would be doing “reruns” of his favorite entries. As a result, many old blog entries, including Bit Gold, were updated as he reposted his favorites.
As discussed before, the reruns started 4 hours after Satoshi’s first e-mail to Adam Back, so the reposting of Bitgold itself is not suspicious, but the timing of the reruns season is.
9. On his blog, Nick Szabo asks in the comments for help “to code something up” to bring his Bit Gold idea to life. This indicates that he believed he needed assistance with the technical aspect. Szabo is on record in emails from the ’90s reinforcing this belief. This suggests that if Szabo had led the effort to create Bitcoin, it would likely have required a group of people to code it.
This is true, in April 2008 Szabo asked for help coding Bitgold. Maybe he got help and didn’t release it alone?
10. However, the evidence supports the theory that the initial code and whitepaper were written by a single individual. Experts have suggested that the code appeared to be written by someone with an electrical engineering background, “close to the metal.” In other words, they would need to know “how the engine works to drive the car.” Many have confirmed that the code didn’t seem to be written by a professional developer, and it was coded on Windows.
I don’t know that this is true at all — we have no way of matching the Bitcoin code against the whitepaper, other than the fact that one account shared both on a mailing list, and Satoshi did work on both, later.
We don’t have a C++ code sample from Szabo (only a small javascript sample) so we can’t say it’s not his code. If he did get help coding it, then we don’t know who’s code we’re looking for. In general, this is a problem for most proposed theories — in the “Adam Back is Satoshi” theory, the NYT says that the writing matches Adam’s, but Bitcoin does not match Adam’s known coding style.
It does seem to be true that Szabo wasn’t the kind of guy who would obsess over the coding details on something like this. He was more of a theorist that liked to come up with designs than someone who would figure out all the intricacies of programming one.
11. A group should have had an easier time building support for Bitcoin in its early days. You would expect the approach to look very different if multiple people were involved.
I think this is probably true, but I don’t know why Szabo plus one coder would look so much different. I guess there would be an awkward process here, where Hal was finding bugs and relaying them to Szabo as Satoshi and then Szabo was passing them on to the coder.
12. Typical cryptography whitepapers include far more citations than Satoshi’s Bitcoin paper. For instance, Adam Back’s Hashcash had 19 citations, which is on the lower side of the average. Meanwhile, Bitcoin had only 8 total citations! Satoshi didn’t seem to know the proper way to write citations and likely mimicked the style of other whitepapers. It’s reasonable to assume they wanted to appear academic but didn’t hold a formal degree in the field.
Yes, this could be a sign that Satoshi was someone from outside of academia, but it could also be someone trying to keep their footprint small, as opposed to citing lots of other academics for a formal paper. It’s not a strong argument, either way.
13. Stylometric analysis has yielded varying results, depending on the samples, references, and a wide range of factors. All the main suspects use similar terminology. If Satoshi was copying others, you’ll get false positives. Therefore, it may be wiser to take stylometric analysis of Satoshi’s writing with a grain of salt, and instead focus on colloquial expressions and quick responses in informal settings, where slip-ups are more likely to occur.
We scanned every word Satoshi ever wrote, along with everything I could find from the main suspects. Satoshi casually used the specific colloquial term: “sweet.” We had something we called the “sweet test.” We focused on usage within a decade of Bitcoin’s launch, since language evolves, people attend conferences, and culture shifts. Peter Todd and Adam both used “sweet” during that time period. I couldn’t find Szabo using “sweet.” However, more recently, Szabo used “sweet” in 2021. All 3 have used “retarded” more recently.
We already talked about this before — I think Adam Back passes the “sweet test” most strongly, among these 3 candidates. Szabo does fail it, so that’s a strike against him.
14. Szabo was a great philosophical mind, particularly focused on the idea of collectibles being minted and traded online, as well as smart contracts. He uses this language in his Bit Gold proposal. However, Satoshi doesn’t use these terms. It’s only later, after discussions on forums, that Satoshi begins to discuss the idea of collectibles or smart contracts, mentioning them toward the end of his forum activity. For Szabo to be Satoshi, he would have needed to completely suppress his key philosophical fascinations during Bitcoin’s development.
I don’t know if this is a strong argument or not. My only question is: in a world where Szabo invented Bitcoin, would we have NFT’s earlier?
15. If you go through the recent Malmi email dump from the CSW court case, in 2010, Satoshi mysteriously received a $2,000 cash “donation,” which he wanted to send to Malmi in an untraceable way to keep the forum afloat. For obvious reasons, this “donation” was almost certainly self-funded. Bitcoin’s bootstrapping phase was incredibly fly-by-night. Malmi was counting down the remaining cash in various emails. The situation reads more like a college graduate scraping together pennies, not a professional like Szabo.
Szabo was selling his collectible bank notes for “personal cash flow needs” in August 2008, so it seems possible that he wasn’t rich. I don’t know where he was working from 2007 to 2014.
This point does seem to go against Adam Back, who would definitely have enough money to donate $2,000. It could also just be that Satoshi didn’t want to let on to Malmi that he was rich.
16. Szabo does seem closely connected to the origin story of Bitcoin. In a somewhat infamous photo, Szabo is seated next to Peter Todd. I wouldn’t be surprised if Szabo knows the true origin of Bitcoin and became involved at some point. However, I don’t believe he created Bitcoin.
Agreed, it sure does feel like Szabo could have some early connection to the project!
With respect to Cullen and all the research he’s done, I don’t find that anything in this list absolutely rules Szabo out. I think the strongest argument made here is just that Szabo was not enough of a coder to step up and solve all the necessary problems.
While Szabo is a great thinker on these topics, Satoshi was more of a practical coder. As we noted earlier, Satoshi claims that he designed the code before writing the whitepaper.
In another early message, Satoshi wrote:
“It’s very attractive to the libertarian viewpoint if we can explain it properly. I’m better with code than with words though.”
Szabo is, without a doubt, a great thinker and someone good with words. Many people think he would not dig into all the obscure details of coding a practical cryptocurrency.
Perhaps also relevant, Nick Szabo went to law school, while Satoshi once answered a question by stating:
“I am not a lawyer and I can’t possibly answer that.”
We could possibly rule out Szabo via stylometry. I’ve shown above that all the stylometry studies pointing towards him are weak.
The New York Times already says they’ve ruled him out, and proven Adam Back is Satoshi. If you trust their stylometry as definitive, then it’s case closed. I have not reviewed their methodology carefully enough to be sure.
If you don’t trust stylometry, then we need something more. To fully rule Szabo out, I would want either a very strong proof that someone else is Satoshi or more information on where Szabo was from 2007–2011: where did he live? what was he working on?
I wasn’t able to track down any of that.
I lean towards thinking Szabo is not Satoshi, but I do still wonder about the many overlaps with Bitgold and the timing of the rerun blog posts.
Time zone mismatch
I’m uncertain what Szabo was working on, from 2008 to 2011, but he was occasionally posting on his blog. I collected the timestamps from all of Nick’s blog posts and compared those to Satoshi’s activity:
It looks like Szabo almost never posts before noon (pacific time), while Satoshi wakes up and starts working before 8 (pacific time). There’s a somewhat poor match in summer 2009, where Satoshi is asleep by midnight (pacific time), while Szabo is posting between midnight and 4 AM. There’s an even worse match in spring 2011, where Satoshi is active from midnight to noon, while Szabo is active from noon to midnight.
I should note that the activity in 2011 is sparse, for both Satoshi and Szabo. I should also say that this is not as high quality of evidence as you might get from posts made on a public forum or mailing list — Szabo has control over his own blogging site, so the time stamps could theoretically be wrong. But his readers would have probably noticed if the timestamps were wrong when they read and commented, so that would be a bold strategy to fake all the timestamps to throw people off the trail, years later.
While there is a bit of uncertainty here, I think that Szabo is generally not a good time zone match for Satoshi.
Lifestyle choices
It’s widely seen as a mystery that Satoshi has not moved the million bitcoins that he mined, back in 2009–2010. Many people think that no one could hold on to that ~100 billion dollar fortune without selling some.
Some people think that this must mean that Satoshi is dead, so they favor candidates like Hal Finney or Len Sassaman.
I don’t think the Hal Finney theory helps explain this, at all — Hal mined some bitcoins in 2009 and he left them to his family. You would have to assume that his family is HODLing 100 billion dollars of coins, or that he didn’t leave them his main stash.
In the case of Nick Szabo, people sometimes argue that he’s just a humble man, drives an old car, and doesn’t really care about being rich.
I’d contrast this to Adam Back’s case, where he’s actively running a company to try to get the world to use Bitcoin. That would boost the value of his own holdings. It would be normal human behavior, for a billionaire — most of the other billionaires spend all their time trying to get richer.
It’s also worth noting that Adam Back avoided the Bitcoin forums for years after Bitcoin’s invention, but suddenly showed up in 2013, on the same day that someone analyzed the blockchain and discovered how much Bitcoin Satoshi had mined. One of Back’s first comments on the forum was about Satoshi’s hoard of coins:
No one knows how many bitcoins Satoshi mined, after May 2010. Why would Adam Back mention some other stash of coins? There has been some complicated blockchain analysis that’s come out, since then, showing that Satoshi might have sold coins from after 2010, while still holding the early ones. That would fit exactly with what Adam described here.
Back has since been a Bitcoin maximalist, promoting the use of the product, and critical of competitors like Ethereum, describing it as a dumpster fire. He says that all altcoins will go to zero and you should only buy Bitcoin.
Szabo has been more interested in crypto ideas and theories. He was initially positive about Ethereum, because it helped implement the kind of smart contracts he’d invented. He said Ethereum had the best ideas in Blockchain tech. He compared Bitcoin to a pocket calculator and Ethereum to a general purpose computer.
Back’s lifestyle and Bitcoin maximalist position match someone with a secret stash of a million bitcoins that he’s trying to pump. Szabo’s behavior matches the guy he’s always been — an early crypto theorist.
Evaluating coincidences
With every one of these Satoshi candidates, we can find interesting coincidences. Hal Finney had previously lived kind of near a guy who’s name (middle+last) was Satoshi Nakamoto. Is that a big deal? In a previous post, I showed that the odds are pretty high that you’d find that coincidence for one Satoshi candidate, if there are enough candidates and enough Satoshi’s in the world.
In this case, one of the biggest coincidences is that Nick Szabo’s blog started posting reruns 4 hours after Satoshi’s first e-mail to Adam Back.
Okay, that feels very weird!
But, suppose that the reruns had been on a different day. What if it was August 18th, the day that Satoshi registered Bitcoin.org? That’d be just as good a coincidence. What is it was July 20th, when ecash.org was registered, or August 17th, when netcoin.org was registered? What if the reruns started on or near October 31st, when Satoshi posted his whitepaper on mailing lists, or January 3rd, with the genesis block, or January 9th, when he started mining? Wouldn’t those still be equally weird coincidences? Suppose that Szabo had posted his reruns in 2007, around the time that Satoshi actually said that he was developing Bitcoin. In that case, you could still say that was evidence, maybe even stronger evidence.
Beyond that, note that we’re comparing a long list of “Satoshi events” with a long list of “Szabo events” and picking two that line up. Imagine if the first Satoshi email to Adam Back had not lined up with the “reruns” post, but it instead matched up to a post about the history of money, or a post about when FDR confiscated gold, or a post mentioning Bitgold. Wouldn’t each of those be interesting coincidences?
Now, remember you’re free to vary both of those parameters, so you just need to pick one important Satoshi date (first email, whitepaper date, launch date, genesis block, etc) and match it to one interesting Szabo date (some blog post that feels relevant).
I’m not sure this is a big enough coincidence to mean anything.
Still, there are other possibilities, besides this being a coincidence.
One parsimonious explanation is that Nick is not Satoshi, but he also got an email from Satoshi on or before August 20th. Maybe that email made Nick decide to get involved or pay attention to the project. That would neatly explain his subsequent posts.
I don’t know if that’s actually the simplest theory, though, because you’ve added another person to the conspiracy.
One last possibility is that the e-mails between Adam Back and Satoshi never happened, and Back was free to make up whatever date and time he wanted, when he shared these messages over a decade later.
In any case, I’m willing to accept the timing of the reruns could be pure coincidence. I don’t think that’s a big enough coincidence to confirm that Szabo is Satoshi.
The messy process of weighing evidence
After the New York Times article came out, someone published a Bayesian analysis giving 99.99% odds that Satoshi is Adam Back. Prior to that, someone else shared a Bayesian analysis giving 63% odds that Satoshi is Nick Szabo and a 2% chance he’s Adam Back. You can dig into both, if you’d like to see what evidence they see as important and how they weigh it.
If you’ve been following my work for a long time, you probably know how skeptical I am about Bayesian analysis. It’s highly sensitive to which inputs you put in, which numbers you put on them, and so on.
Most of the time, I don’t think it’s much better than just laying out all your evidence in a table. I might summarize the comparison like this:
I suppose that the Bayesian approach lets you weigh each factor differently, so that if one matching factor is more important, it outweighs weaker evidence which does not match.
Of course, in cases where one factor is extremely important, you can just point to that one factor and you don’t need Bayesian reasoning. When you just have many weaker factors, it’s hard to know the right weights to put on each factor, and it’s hard to know if you picked all the right factors
Conclusions
I think it’s unlikely that Nick Szabo is Satoshi.
This conclusion is weaker than the other articles I’ve written. Craig Wright is an obvious fraud. With Hal Finney, I showed that Satoshi was writing e-mails and sending bitcoins while Hal Finney was running a race. With Len Sassaman, I showed that Satoshi was active after Len died, and I showed that Satoshi’s activity did not match Len’s travel schedule when he was alive. I’m happy to call those other guys extremely unlikely candidates, < 1% choices.
With Szabo, It’s hard to decisively rule him out. I never found enough details on his activity to get an alibi. You can still speculate about possibilities where he got some coding help and made Bitcoin, or where he just stepped up to the plate and did it all. You’d have to imagine that he obfuscated his language, adopted Adam Back’s spelling mistakes, and so on, but none of that is 0%. I don’t know what odds I would actually put on it — less than 5%? 10% sounds too high. A closer look at stylometry would change my confidence level here. In any case, I would not rank him as the top candidate.
