This is an essential collection of forum exchanges and writings of Bitcoin's founder Satoshi Nakamoto, written circa 2008-2010. Of the 3 books on Bitcoin, I've read thus far, including "Digital Gold", and "The Age of Cryptocurrency", this one is by far the most informative, focusing purely on the motivation and thinking behind the construction and early refinement of the Bitcoin architecture. The problem with many contemporary tellings of the Bitcoin story is many authors (mostly journalist) who have attempted it, seem to inevitably attempt to ape the "Social Network"/"Big Short" heist-like narrative of a group of miscreants/outcasts, raging (figuratively) against the machine to birth something "great".
This perspective is not only false in many cases but has become tired. Although lacking in formal narrative what the "Book of Satoshi" does well is to show the reader the step-by-step struggles of puzzling through the notion of a decentralized non-trust based currency system, many of the details, the conceptual point/counterpoints, and how they were resolved, and despite the subject-matter of cryptocurrency falling well within the domain of computer science and/or applied discrete mathematics, the exchanges of Satoshi Nakamoto and his colleagues is easy enough to understand for a reasonably intelligent person to grok listening while jogging, with little rewind/playback.
Anyone who's even thought a little about BItcoin has thought immediately about possible exploits, and "The Book of Satoshi" thankfully introduces the most famous of these exploits early on in the narrative (which I believe is ordered chronologically), the "51% attack". Since Bitcoin is a non-trust based decentralized currency, Nakamoto explains that "CPU time" (or GPU time as was the case after 2010 and 2011) is an instrument for trust, and he goes into intricate detail on how the blockchain works within the context of a peer-2-peer system when there are potentially many different versions of the "current" ledger, which involves a validation by voting across the network. In this context, the natural exploit many have come up with is the 51% attack, the idea that if enough systems were commandeered by a malicious group or antagonist nation-state, enough computing cycles could be deployed to alternative (and wrong) ledgers that could pierce the veracity of the record, allowing wholesale faulty redistribution of capital between the current accounts, the so-called "double-spend" conundrum.
The exchange between Nakomoto and his colleagues here is great because it shows that good ideas do not spring fully-formed from minds of inscrutable geniuses, but are often half-formed and often the founders themselves had incomplete solutions to challenges they faced, but Post Hoc the solution that eventually came about seems magically simple and exhaustive, and hence, genius. Nakomoto here gave a feasibility argument that practically no group of exploiters, that is those who often run 'zombie farms' could wield enough run-cycles to overturn the true ledger in such a voting scheme on account of scale in order of magnitude, since these groups often wield farms of 100k commodity boxes vs 10 million or more. Further, Nakomoto argues, any such entity/group that could yield this kind of compute-resources would just deploy it to mining bitcoin, since cycle-time deployed to that activity would yield more revenue than attempting to cheat the system. This is a surprisingly "economist"-like argument answers the question by not answering the question. Yes, assuming all agents within the network had a pure profit-objective, this is "probably" correct. It is also the case that the scope of actions possible by the "51-percent" holder would be limited in that they would not be able to introduce new transactions to the ledger, yet it's clear that such an exploiter could deal a lot of damage to the system if their objective was pure malice with a political and/or other non-profit based motivation. Clearly, this won't be the last word on this matter. Adversarial algorithms are just starting to lift off in the field of machine learning, and I imagine cross-pollinating applications in that to these kinds of exploit challenges.
Other topics covered by the book include Nakamoto's thoughts on transaction costs, inflation, the mechanics of adding blocks to the block-chain, including some exploration on cases where two competing, but single-spend blocks are being added to the ledger, how this ledger system, the "proof of work" resolves the Byzantine's Generals Problem" from discrete mathematics and networking etc. There is also some minor discussion on more traditional "currency and money" topics including some discussion on central banking, the problem of regulations and money supply challenges that Bitcoin could help solve in a "fair" manner. What's clear though is Nakotmo was/is a technologist, and his views on these topics are sparse (at least from his writings). Contrary to what some would also believe, Nakotmo also seems to have very tentative readings into traditional libertarian writings, like those of Hayek and Von-Mises, though both are mentioned in some of his posts.
More interesting are the later posts, where the Cryptography group discuss how this bourgeoning system would interact with traditional currency/transaction systems in-place in the world today. Some topics here include what would happen if there was a war between China and the US, and the bitcoin ledger was bifurcated? In 2010, a fantastical notion, but potentially more prescient with respect to today's headlines, at least in the financial/economic/technological realms. There also includes a series of posts dealing with versioning and updates, which are less relevant to a more generalist audience, but still interesting as a historical text. Others still are superfluous to most, but again interesting as a historical curiosity, like the groups attempt to prevent the "Bitcoin" on Wikipedia from being deleted in the early days of the 2010s.
This is great reading. I've since gone on to start working through a formal textbook published by PUP on this subject, Arvind Narayanan et. al. text "Bitcoin and Cryptocurrency Technologies". Which should be read first, or perhaps concurrently, I can't say for sure at the moment. However, if one wants to read just one book on Bitcoin, this is the one I'd pick up if it's just 2, I'd check out this one and "The Age of Cryptocurrency". Highly recommended