Wei Dai

Aaron van Wirdum of Bitcoin Magazine wrote that in the 1990s Wei Dai was an undergraduate in computer science at “Washington University” and was drawn to Timothy C. MayTimothy C. MayTimothy C. May (1951–2018) was a physicist and writer who, while at Intel, identified alpha particles as the cause of soft errors in dynamic memory chips; he wrote “The Crypto Anarchist Manifesto” in 1988, co-founded the Cypherpunks mailing list with Eric Hughes and others in 1992, and wrote “The Cyphernomicon,” a document in FAQ form.Open the full entry’s crypto-anarchy.⁠[1]

In the mid-1990s he was a prolific contributor to the Cypherpunks mailing list, discussing digital reputation systems, game theory, privacy and anonymity in digital cash systems, and proposing schemes that included trusted timestamping, an encrypted TCP tunneler and a secure file-sharing system. Van Wirdum wrote that even then no one knew much about him personally; Timothy C. May recently said that no one knew even whether Dai was male or female.⁠[1]

He later recalled on LessWrong that in the 1990s he spent a few years thinking about how a group of anonymous people on the Internet could pay each other with money without outside help, work that culminated in the publication of b-money in 1998; he wrote: “We (or at least I) weren’t trying to destroy government, but just create new virtual communities that aren’t ruled by the threat of violence.”⁠[2]

Crypto++, a free C++ class library of cryptographic schemes, was originally written by Wei Dai; the compilation copyright notice in the library’s licence file reads “Compilation Copyright (c) 1995-2019 by Wei Dai.”⁠[3][4]

The version history records that version 1.0 was the first public release, was withdrawn at the request of RSA DSI over patent claims and had a serious bug in its RSA key generation code; version 2.0 added a completely new, faster multiprecision integer class; and version 3.0 placed all names into the “CryptoPP” namespace.⁠[5]

On his LessWrong profile he lists creating Crypto++ among his main “claims to fame,” describing it as “the first general purpose open source cryptography programming library.”⁠[6]

Van Wirdum wrote that Dai casually announced the idea of b-money in November 1998, just after graduating from university.⁠[1]

The b-money essay opens: “I am fascinated by Tim May’s crypto-anarchy.” It states that a community is defined by the cooperation of its participants and that efficient cooperation requires a medium of exchange (money) and a way to enforce contracts; traditionally these services have been provided by the government or government-sponsored institutions and only to legal entities, and the essay sets out to describe a protocol by which they can be provided to and by untraceable entities. It describes two protocols. In the first, every participant maintains a separate database of how much money belongs to each pseudonym; anyone can create money by broadcasting the solution to a previously unsolved computational problem, the number of monetary units created being equal to the cost of the computing effort; money is transferred by a broadcast message signed by the payer, which is ignored if it would create a negative balance in the payer’s account; a valid contract must include a maximum reparation in case of default for each party and a party who will perform arbitration should there be a dispute, and when it takes effect each party’s account is debited by its maximum reparation and a special account identified by a secure hash of the contract is credited. The essay calls the first protocol impractical because it makes heavy use of a “synchronous and unjammable anonymous broadcast channel.”⁠[7]

In the second protocol, the accounts of who has how much money are kept by a subset of the participants (servers); each server is required to deposit a certain amount of money in a special account to be used as potential fines or rewards for proof of misconduct, and each participant should verify that his own account balances are correct. The appendix, “Appendix A: alternative b-money creation,” opens: “One of the more problematic parts in the b-money protocol is money creation.” It explains that the protocol requires all account keepers to agree on the cost of particular computations, information that may be unavailable, inaccurate or outdated because computing technology tends to advance rapidly and not always publicly; it therefore proposes an alternative money-creation subprotocol in which the account keepers instead agree on the amount of b-money to be created each period, with the cost of creating that money determined by an auction: anyone who wants to create b-money broadcasts a bid giving the amount he wants to create and an unsolved problem from a predetermined problem class, and after the solutions are broadcast each account keeper accepts the highest bids in terms of nominal cost per unit created and credits the bidders’ accounts.⁠[7]

He later wrote on LessWrong that b-money was not yet a complete practical design and that he did not take any steps to code it up, adding: “I had actually grown somewhat disillusioned with cryptoanarchy by the time I finished writing up b-money.”⁠[2]

In November 2006 Ted Krovetz of CSU Sacramento published the Internet-Draft “VMAC: Message Authentication Code using Universal Hashing,” describing VMAC, a message authentication algorithm designed to have exceptional performance in software on 64-bit CPU architectures while still performing well on 32-bit architectures; the April 2007 revision (draft-krovetz-vmac-01) adds Wei Dai as an author, with the affiliation Bitvise Limited. The draft is an individual draft without IETF endorsement and expired in October 2007.⁠[8][9]

On August 29, 2007, Wei Dai and Ted Krovetz posted the paper “VHASH Security” to the Cryptology ePrint Archive; it describes VHASH as an almost-delta-universal hash family designed for exceptional performance on computers and states that the changes it details improve both security and performance over the original 2006 version. Crypto++ version 5.5 added VMAC.⁠[10][5]

The comments of the file vmac.h in the Crypto++ repository on GitHub state that the file was originally written and placed in the public domain by Wei Dai, and that VMAC is a block cipher-based message authentication code algorithm using a universal hash proposed by Ted Krovetz and Wei Dai in April 2007; they link to draft-krovetz-vmac-01 and state that the implementation is based on Ted Krovetz’s public-domain vmac.c.⁠[11]

On August 22, 2008, Satoshi NakamotoSatoshi NakamotoSatoshi Nakamoto is the pseudonymous creator of Bitcoin, who published the Bitcoin white paper in 2008, released the first Bitcoin software and launched the blockchain in 2009, and stopped public activity around 2011; the person’s true identity has never been confirmed.Open the full entry emailed Wei Dai under the subject “Citation of your b-money page,” opening with “I was very interested to read your b-money page.” He wrote that Adam BackAdam BackAdam Back is a British cryptographer and computer scientist, the inventor of the proof-of-work system Hashcash, and a co-founder and the chief executive of the Bitcoin technology company Blockstream.Open the full entry had pointed him to Dai’s site, that he was getting ready to release a paper that “expands on your ideas into a complete working system,” and that he needed the year of publication of the b-money page for the citation in the paper; his draft citation gave the year provisionally as “(2006?),” and he included the abstract of the draft “Electronic Cash Without a Trusted Third Party.”⁠[12]

Wei Dai replied that b-money had been announced on the Cypherpunks mailing list in 1998, gave links to the archived post and to discussion of it, and said he would take a look at the paper and let Nakamoto know if he had any comments or questions.⁠[12]

The introduction of the Gwern.net page describes Wei Dai as “one of the few cryptographers to even dabble in e-currency speculation at the time” and states that he was not particularly interested but mentioned Nick SzaboNick SzaboNick Szabo is an American computer scientist, legal scholar and cryptographer who developed the term and concept of smart contracts and in 1998 conceived bit gold, a design for digital money often described as a precursor of Bitcoin; he has repeatedly denied being Satoshi Nakamoto, and in January 2025 he became chief scientist of JAN3.Open the full entry’s uncited Bit Gold to Nakamoto. Wei Dai’s reply as published on the page does not mention Bit Gold or Nick Szabo, and the page includes no reply from him to the email of January 10, 2009. Bitcoin Magazine’s van Wirdum, by contrast, wrote that Dai did not respond to Satoshi’s email and in retrospect wished he had.⁠[12][1]

Reference 1 of the Bitcoin paper, “Bitcoin: A Peer-to-Peer Electronic Cash System,” is W. Dai’s “b-money” (1998), which section 2, “Transactions,” cites where it states that transactions must be publicly announced; the references also include Adam Back’s Hashcash.⁠[13]

On January 10, 2009, Satoshi Nakamoto wrote to Wei Dai again, saying that he had just released Bitcoin v0.1, the full implementation of the paper he had sent Dai a few months earlier, and adding: “I think it achieves nearly all the goals you set out to solve in your b-money paper.” The email also quoted Hal FinneyHal FinneyHal Finney (1956–2014) was a software engineer and cryptographer, an early developer of the PGP encryption software, who released the reusable proof-of-work system RPOW in 2004; in January 2009 he ran the Bitcoin software released by Satoshi Nakamoto and received the first Bitcoin transaction.Open the full entry’s overview of Bitcoin from the cryptography mailing list.⁠[12]

On August 13, 2009, Wei Dai published “Towards a New Decision Theory” on LessWrong. It opens by noting that current decision theories have deficiencies that show up as various paradoxes; since there seemed little hope that Eliezer YudkowskyEliezer YudkowskyEliezer Yudkowsky is an American AI researcher and writer who in 2000 founded the organization later renamed the Machine Intelligence Research Institute (MIRI) and is regarded as an early founder of the field of AI alignment; he has long warned that superintelligence could cause human extinction and calls for a worldwide halt to large AI training runs, views that many researchers dispute.Open the full entry would publish his Timeless Decision Theory soon, Dai wrote, he had decided to try to synthesize some of the ideas discussed on the forum. A note added later at the end states that the idea came to be called “Updateless Decision Theory.”⁠[14]

LessWrong’s wiki page states that Updateless Decision Theory (UDT) was invented by Wei Dai and first described in that post, and that it is meant to deal with the problem of dynamic inconsistency; UDT and Timeless Decision Theory (TDT) have very similar motivations, Wei Dai having formulated UDT as something like a guess about TDT because Eliezer was slow to publish its details; the page adds, attributing the statement to Nate SoaresNate SoaresNate Soares is an AI safety researcher and president of the Machine Intelligence Research Institute (MIRI). He works on decision theory and AI alignment and co-authored If Anyone Builds It, Everyone Dies with Eliezer Yudkowsky.Open the full entry, that Wei Dai does not endorse functional decision theory’s (FDT) focus on causal-graph-style counterpossible reasoning.⁠[15]

On July 20, 2010, Satoshi Nakamoto wrote in a BitcoinTalk forum post that “Bitcoin is an implementation of Wei Dai’s b-money proposal,” noting that the proposal had appeared on the Cypherpunks mailing list in 1998; the same post also mentions Nick Szabo’s Bitgold proposal.⁠[16]

On February 16, 2011, the LessWrong user Clippy posted “Making money with Bitcoin?”, introducing Bitcoin and asking whether the “Wei Dai” mentioned in it was the same as the LessWrong user Wei_Dai. Wei Dai replied in a comment that he did not create Bitcoin “but only described a similar idea more than a decade ago”; his understanding was that Satoshi Nakamoto did not even read his article before reinventing the idea himself, and learned about it afterward and credited him in his paper, so that his connection with the project was quite limited. He wrote that, thanks to the discussion, he had bought a Radeon 5870 and started mining himself, since it looked likely that he could at least break even on the cost of the card; he had heard about Bitcoin earlier, but in his version of the idea nobody could make large profits by mining or minting, so he had not looked into the possibility. He also warned that he had not analyzed the security of the protocol, that to his understanding not many other cryptographers had either, and that in his experience such protocols often have flaws that are not found until that has happened.⁠[17]

Commenting under “Bitcoins are not digital greenbacks,” a LessWrong post published on April 19, 2013, Wei Dai wrote: “I would consider Bitcoin to have failed with regard to its monetary policy,” because the policy causes high price volatility that imposes a heavy cost on users, who have to take undesirable risks or engage in costly hedging; Bitcoin, he wrote, might have been coded with a monetary policy that emphasized price stability instead of a fixed supply of money, and this may have been partially his fault, because when Satoshi wrote asking for comments on the draft paper, “I never got back to him.”. Bitcoin Magazine, quoting the comment, gives the following sentence: “Otherwise perhaps I could have dissuaded him (or them) from the ‘fixed supply of money’ idea.”⁠[18][1]

On January 12, 2014, the LessWrong user Will_Newsome opened the thread “AALWA: Ask any LessWronger anything,” in which Wei Dai answered questions, opening with the remark that he had been getting an increasing number of interview requests from reporters and book writers. In his answers he wrote that he thought Satoshi was probably one person; on Satoshi’s motivation he guessed in one reply that it was probably a mixture of technical challenge and wanting to change the world, and wrote in another that he was not sure Satoshi would even consider himself a crypto-anarchist and that Satoshi might have been motivated more by a distrust of financial institutions and government monetary authorities and wanted to create a monetary system that did not have to depend on such trust.⁠[2]

Replying to gwern, he clarified that he did not mean that Nick Szabo does not know C++, but that Szabo is not known to have programmed much in C or C++; he added that Nick considers his ideas to be at least an independent invention from b-money.⁠[2]

On March 2, 2014, The Sunday Times published Andrew Smith’s article “Desperately seeking Satoshi; From nowhere, bitcoin is now worth billions. Where did it come from?”, which quoted Wei Dai as being “pretty sure” that Nick Szabo is not Satoshi.⁠[12]

On March 17, 2014, Gwern Branwen published on his website the three emails exchanged between Wei Dai and Satoshi Nakamoto in 2008–2009; the page states that Wei Dai provided them publicly to support his judgment that Satoshi is not Nick Szabo, his reasons including that Satoshi’s early emails show no awareness of Szabo’s ideas and that Nick “isn’t known for being a C++ programmer.”⁠[12]

An Aston University release published on April 17, 2014, states that its researchers compared the language of the Bitcoin paper with the writing of eleven individuals, Wei Dai among them, along with Dorian S. Nakamoto, Nick Szabo, Hal Finney and others; the study’s lead, Jack Grieve, said that out of the list of people regularly referred to as possibilities, they were pretty confident that Nick Szabo is the main author of the paper.⁠[19]

In June 2015 Wei Dai turned the Crypto++ library over to the community, an event the Readme file calls “Changing of the guard”; the first community release was version 5.6.3, which the version history records as having 12 unique contributors and fixing CVE-2015-2141.⁠[3][5]

As of November 2026 the README of the Crypto++ repository on GitHub is headed with version 8.9, dated October 1, 2023, and its version history lists 8.9.0, of October 1, 2023, first; the README states “The Crypto++ library was originally written by Wei Dai.” and “The library is now maintained by several team members and the community.”⁠[20]

On June 14, 2018, Bitcoin Magazine published Aaron van Wirdum’s article “The Genesis Files: If Bitcoin Had a First Draft, Wei Dai’s B-Money Was It.” The article judges that the first version of b-money, in which every participant keeps the ledger, is roughly how Satoshi Nakamoto designed Bitcoin ten years later, and that the second version loosely resembles what would today be called a proof-of-stake system; it adds that Dai included an early version of a smart contract solution and that the value of b-money was to be coupled to the value of a theoretical basket of goods.⁠[1]

The article states that b-money was never implemented, that Dai himself expected it to be at most a niche currency or contract-enforcement mechanism, and that he believes Bitcoin’s inventor came up with the idea independently. Van Wirdum also wrote that it is probably more accurate to consider Nick Szabo’s Bit Gold “Bitcoin’s first draft.”⁠[1]

The README of the Crypto++ repository on GitHub states “Crypto++ Library is a free C++ class library of cryptographic schemes.” and lists among its contents authenticated encryption schemes, high-speed stream ciphers, AES and the AES candidates and other block ciphers, block cipher modes of operation, message authentication codes, hash functions, public-key cryptography, elliptic curve cryptography, and insecure or obsolescent algorithms retained for backwards compatibility and historical value; the version history records that version 5.6.2 changed the licence to the Boost Software License 1.0, and the licence file states that all individual files in the compilation are placed in the public domain by Wei Dai and other contributors.⁠[20][5][4]

The version history records that version 5.03 created a DLL version of Crypto++ for FIPS 140-2 validation; the Readme file states that the library and the DLL are no longer considered validated and that the DLL should no longer be used, and the README of the GitHub repository states that the library was moved to the CMVP’s Historical Validation List.⁠[5][3][20]

An article in The Motley Fool states that the EthereumEthereumEthereum is an open, general-purpose smart contract blockchain that launched in 2015, and its native asset is ether (ETH). The project was proposed by Vitalik Buterin and advanced jointly by eight co-founders; the network now uses proof of stake and centers on the EVM, the account model, a rollup scaling roadmap, and the open EIP process.Open the full entry team named the smallest unit of ether “wei” as a nod to Wei Dai; each of the seven units of ether has a nickname based on a notable figure, babbage, for example, being named for Charles Babbage.⁠[21]

Van Wirdum wrote that Wei Dai is active on LessWrong, where he philosophizes on artificial intelligence, ethics, epistemology and other topics, and that he has drawn praise from the AI researcher Eliezer Yudkowsky and repeated invitations to speak at the Machine Intelligence Research Institute (MIRI, previously known as the Singularity Institute).⁠[1]

His LessWrong posts include “A tale from Communist China” (October 18, 2020), “Morality is Scary” (December 2, 2021) and, in 2025, “Legible vs. Illegible AI Safety Problems” (November 4), “Problems I’ve Tried to Legibilize” (November 9), “Please, Don’t Roll Your Own Metaethics” (November 12), “Relitigating the Race to Build Friendly AI” (December 3) and “A Conflict Between AI Alignment and Philosophical Competence” (December 27).⁠[6]

Van Wirdum wrote that The New York TimesThe New York TimesThe New York Times is an American morning daily newspaper published in New York City. Founded in 1851, it is published by The New York Times Company.Open the full entry once described him as an “intensely private computer engineer.”⁠[1]

In LessWrong’s AALWA thread the user riceissa asked him how he decides which organizations to associate with, noting that he had never joined a rationality-community or effective-altruism organization. He replied that he has an independent source of income and that it is costly to move his family (“we’re not near any major orgs”), so he would have to join in a remote capacity, and listed the pros and cons of joining an organization, its name redacted, that had recently tried to recruit him.⁠[2]

On October 18, 2020, he published “A tale from Communist China” on LessWrong, about his grandparents’ failure to emigrate from Communist China before it was too late. He wrote that none of this was revealed to him when he was a child living in China, and that by then his parents had grown quite disillusioned with Communism and the Communist Party. By his account, his grandparents were all supporters of the Communist Party at the beginning, seeing it as potential saviors or liberators of the Chinese people and nation, and were treated well at first: one of them, being one of few people in China with a higher education, was given a high official post, and another retained a manager position at the factory he used to own and operate. The latter side of the family had considered moving away from China before the Communist victory, since they would be classified as “capitalists” under the new regime, but were reassured by high-level party members that they would be treated well if they stayed and helped to build the “new China”; most of their property was later confiscated or nationalized, and they were eventually forced to share the house they used to own with about four other families, left with a single room to live in. The other side were more straightforward “true believers,” part of the educated class, who supported Communism at an early stage: his grandmother on that side even tried to run away from her family to join the revolution, and later wanted to move away from China and got the required visas; his grandfather was given an important post and went around making important speeches, was later demoted and moved from city to city as the Party assigned him to various jobs, and, after a political dispute near the start of the Cultural Revolution led his opponents to dig up an incident in his distant past, was persecuted in various ways, including being confined in a makeshift prison for ten years; he died shortly after the Cultural Revolution ended and he was released, just before the author was born.⁠[22]

Personal website: https://www.weidai.com/ (opens in a new window).⁠[23][13]

Crypto++: Official website: https://cryptopp.com/ (opens in a new window).⁠[20]

LessWrong: Public profile: https://www.lesswrong.com/users/wei_dai (opens in a new window).⁠[6]