Gwern's Reviews > Digital Gold: Bitcoin and the Inside Story of the Misfits and Millionaires Trying to Reinvent Money

Digital Gold by Nathaniel Popper
Rate this book
Clear rating

's review

really liked it

Popper delivers a whirlwind tour of almost all dramatis personae in the rise of Bitcoin over the past 5 years. He seems to have gotten access to and interviewed everyone, from the early coders to especially all the late-entering business and entrepreneur types and the incestuous Silicon Valley VC community. (He didn't get access to Ulbricht, for obvious reasons - even the NYT name can't open all doors - but the evidence filings make up for it.) Even I, someone who's watched the space in detail for years and made my own minor contributions to documenting Bitcoin history, learned a lot. (Karpeles had a Japanese wife & son who now live in Canada? I had no idea!) From the Winklevii opening the kimono to settle all their beefs with Charlie Shrem for bungling BitInstant into bankruptcy & personally into prison to Martti and Gavin and other early coders giving Popper Satoshi emails, he covers everything. Even the endlessly complicated story of SR1 gets a decent treatment (though necessarily not as thorough as Ormsby's Silk Road, and like it, somewhat outdated, and passing over the post-SR1 DNM history). As far as histories of Bitcoin up to 2014 go, I don't know of any better single source to consult right now, and the inside access means any future histories will have to look over it carefully as a primary source. (See also Satoshi on GPU mining & Martti & Satoshi discussing growth strategy).

If the book succeeds in capturing what a wide breadth of characters have been involved in Bitcoin (and yet, there are so many more things to cover - the MtGox leak, the ASIC scams, the DNM exit scams and wars, the Chinese market manipulation, the Cambrian explosion of altcoins with attendant pump-and-dumps, Ethereum's attempt to e all things to all people, the blocksize schism...), it perhaps does not succeed at offering any sort of overall synthesis or in giving closure to all the individual stories, or at least including a summary of where everyone and everything stood where the book closed. The description of growth can feel like just a chaos of events, one after another. (It's also fairly weak on explaining the technical aspects - I have to wonder if the lay reader comes away really understanding why Proof-of-Work works or what the Bitcoin blockchain really is.)


That said, as in any book touching on so many topics, there are some errors. Here are some corrections I noticed in material touching on particular interests of mine, the DNMs and Satoshi:

The nine-page PDF attached to the e-mail made it clear that Satoshi was deeply versed in all the previous efforts to create a self-sustaining digital money. Satoshi’s paper cited Back and Wei Dai, as well as several obscure journals of cryptography. But Satoshi put all these earlier innovations together to create a system that was quite unlike anything that had come before it.

'deeply versed'? It cited Dai only because Adam Back had told Satoshi to cite Dai. It also didn't cite any of Szabo's work, even though Finney had pointed that out on the mailing list before. Further, it did not compare or contrast Bitcoin in any meaningful way with all the previous work on digital currency like the whole universe of techniques and approaches based on Chaumian blinding. Altogether, it looks like the opposite of 'deeply'.

Ross didn’t know it at the time, but his downfall had not come through the sophisticated hacking techniques and leaking IP addresses that he had worried about so much. The Internal Revenue Service agent who finally identified Ross did so by searching on Google through old posts on the Bitcoin forum.

Everyone assumed from the inclusion in the complaint that the email was his downfall, but D-Y's testimony during the trial yielded the surprise (one of many) that he had found the email only shortly before the arrest and that the subpoenas had not yet come back with any information. They did help snag baronsyntax, but the actual cause was the FBI finding the Iceland server (thanks, presumably, to Tarbell hacking it), which had a VPN IP hardwired and had a clearnet backup server in Pennsylvania, both of which led back to Ross in San Francisco.

Most bizarrely, Nick altered the dates on his 2008 postings about bit gold to make it appear as though they had been published after Bitcoin was released, rather than before....Most bizarrely, Nick altered the dates: the dates that Nick later put on the posts are at the top of each post. But the URL addresses of the posts still show the original posting date. For instance, his post on “Bit Gold Markets” says that it was written on December 27, 2008, but the URL is

Nothing bizarre about it. As I've pointed out repeatedly since then, Szabo already in 2008 explained what the redating was about; he was re-running older posts: That's all.

Just a few months before Bitcoin was released, in April 2008, Nick had posted on his blog an item in which he talked about creating a trial model of bit gold and asked if anyone wanted to help him “code one up.”

This is evidence against Szabo being Satoshi! The prototype was a big piece of software with a ton of moving parts and low-level details, written in a low-productivity language, with a GUI, mockups for an online store and poker playing, and so on just in the first release; coding it up and debugging it to the point of a public release in just 8 months would be a pretty impressive feat all on its own, and worse, Satoshi says it took 'a year and a half' in November 2008, so he probably started around May 2007.

339“repeated use of ‘of course’ without isolating commas”: Skye Grey, “Satoshi Nakamoto Is (Probably) Nick Szabo,” LikeinaMirror, December 1, 2013,

Skye Grey's claims are BS; stylometrics doesn't work like that, and when people do run stylometrics, Szabo does not come out on top. (While not named in the article, I am told by an involved journalist that Szabo's writings were included but were a poorer match than Finney.)

An academic study of Silk Road later found that nearly 99% of all reviews gave the maximum score of 5 out of 5.

This is too high and was a mistake in that version of the paper. The percentage was biased upwards by a substantial amount because when you are scraping a site like SR1, you will only see a small fraction of the negative reviews from an exit scammer; if an exit scammer rips off 1000 people, he will be banned after a few dozen negative reviews, and then won't appear in your data at all. So as far as your analysis can tell, a 5-star seller just vanished overnight. For example, Tony76 could probably account for 1%+ of sales all on his own, yet his exit scam doesn't appear in the Christin data because they had scraping problems at the time and by the time they got another copy of SR1, that account was banned. Another issue is early finalization; to FE, you have to leave a review, which of course will be 5-stars, and then when you accept you've been scammed, you will probably never go back to update it to 1 star. So one of the changes made to the preprint version of Christin's paper was to address these issues, and the final version should be used instead: "Traveling the Silk Road: A measurement analysis of a large anonymous online marketplace".

(Also as far as this part of the book goes, it would be better to use Ross's own sales figures from the court evidence.)

tied to an Internet provider in California: Hal’s debug log showed that the IP addresses of the other user was reached through a Tor service that would have obscured the real IP address. But Tor generally routes users to nodes in the same geographic area, suggesting that the other user on Bitcoin’s first day was probably in California.

I'm a little annoyed to see someone else discovered this, but in any case, this is only partially correct. Freenode banned open proxies, Bitcoin only gained proxy support in the later version 0.2 in December 2009 (before, it couldn't've worked using Tor because it operated by running '/WHO' on other Bitcoin nodes and connecting straight to their IP), the Bitcoin prototype was designed to 'pay to IP', and in any case, the historical Tor exit node data for January 2009 do not list; of the 3 nicks in the Bitcoin IRC channel, 1 was obviously Finney's client, Satoshi was probably the Tor-cloaked user 'x93428606' in the log, and he was also almost certainly the final nick, the naked Bitcoin node 68.x, which resolves to a residential address in San Diego before 2009. (I looked into the one person I was able to link to that address, but unfortunately neither he nor any of his relatives or friends on Facebook look remotely like possible Satoshi candidates, so for non-state actors, that is a dead end.) Hence, I believe Satoshi was indeed in California that day and this was a rare OPSEC failure by him in exposing his real IP. Also, as far as I am aware, Tor doesn't, can't, and shouldn't 'route users to nodes in the same geographic area', as that would require exit nodes to know where the user is and defeating the point.

(Full disclosure: Popper offered a free copy of Digital Gold to me pre-publication to review, but I wound up not accepting because he was offering a physical book rather than an ebook. I also was a paid fact-checker on an earlier rival Bitcoin book, Dominic Frisby's Bitcoin: The future of money?.)
26 likes · flag

Sign into Goodreads to see if any of your friends have read Digital Gold.
Sign In »

Reading Progress

April 23, 2015 – Shelved
April 23, 2015 – Shelved as: to-read
November 11, 2015 – Started Reading
November 12, 2015 – Finished Reading

Comments Showing 1-10 of 10 (10 new)

dateDown arrow    newest »

Gwern Easier to get, easier to excerpt from, easier to store long-term.

Gwern I just use Calibre on my laptop.

message 3: by Atlas (new) - added it

Atlas Can also easier to pirate!

message 4: by Simon (new)

Simon Morgan Surprised anybody uses Calibre for actual reading. The included reader is a truly terrible piece of code.

Gwern FBreader kept segfaulting, so...

Krishaan Khubchand Green, do you have any recommendations on the history of Bitcoin + Crypto since 2014/15? Just finished reading this and it was great for adding context ‘till then, now I’m looking for something for the last 5 years 🤓

Gwern No, sorry.

message 8: by Andrei (new)

Andrei For the history of Bitcoin up until mid-2017 there is

The author is in the process of writing a sequel, see

Gwern I do not recommend anything on transhumanism or Bitcoin written by David Gerard.

message 10: by Ryan (new) - rated it 4 stars

Ryan Lackey I second that anti-recommendation.

back to top