Crypto-Anarchy and Virtual Communities

By Timothy C. May


Modern Cryptography

The past two decades have produced a revolution in cryptography (crypto, for short) the science of the making of ciphers and codes. Beyond just simple ciphers, useful mainly for keeping communications secret, modern crypto includes diverse tools for authentication of messages, for digital time stamping of documents, for hiding messages in other documents (steganography), and even for schemes for digital cash.

Public key cryptography, the creation of Diffie and Hellman, has dramatically altered the role of crypto. Coming at the same time as the wholesale conversion to computer networks and worldwide communications, it has been a key element of security, confidence, and success. The role of crypto will only become more important over the coming decades. Pretty Good Privacy, PGP, is a popular version of the algorithm developed by Rivest, Shamir, and Adleman, known of course as RSA. The RSA algorithm was given a patent in the U.S., though not in any European countries, and is licensed commercially.  [2]

These tools are described in detail in various texts and Conference proceedings, and are not the subject of this paper.  [3] The focus here is on the implications of strong crypto for cyberspace, especially on virtual communities. Mention should be made of the role of David Chaum in defining the key concepts here. In several seminal papers (for example,  [4]  [5]), Chaum introduced the ideas of using public key cryptography methods for anonymous, untraceable electronic mail, for digital money systems in which spender identity is not revealed, and in schemes related to these. (I make no claims of course that Chaum agrees with my conclusions about the political and socioeconomic implications of these results.)

Virtual Communities

Notes: cyberspace, Habitat, VR, Vinge, etc. Crypto holds up the "walls" of these cyberspatial realities. Access control, access rights, modification privileges.

Virtual communities are the networks of individuals or groups which are not necessarily closely connected geographically. The "virtual" is meant to imply a nonphysical linking, but should not be taken to mean that these are any less community like than are conventional physical communities.

Examples include churches, service organizations, clubs, criminal gangs, cartels, fan groups, etc. The Catholic Church and the Boy Scouts are both examples of virtual communities which span the globe, transcend national borders, and create a sense of allegiance, of belonging, and a sense of "community." Likewise, the Mafia is a virtual community (with its enforcement mechanisms, its own extralegal rules, etc.) Lots of other examples: Masons, Triads, Red Cross, Interpol, Islam, Judaism, Mormons, Sindero Luminoso, the IRA, drug cartels, terrorist groups, Aryan Nation, Greenpeace, the Animal Liberation Front, and so on. There are undoubtedly many more such virtual communities than there are nation-states, and the ties that bind them are for the most part much stronger than are the chauvinist nationalism emotions. Any group in which the common interests of the group, be it a shared ideology or a particular interest, are enough to create a cohesive community.

Corporations are another prime example of a virtual community, having scattered sites, private communication channels (generally inaccessible to the outside world, including the authorities), and their own goals and methods. In fact, many "cyberpunk" (not cypherpunk) fiction authors make a mistake, I think, in assuming the future world will be dominated by transnational mega-corporate "states." In fact, corporations are just one example of many of such virtual communities which will be effectively on a par with nation states. (Note especially that any laws designed to limit use of crypto cause immediate and profound problems for corporations/countries like France and the Philippines, which have attempted to limit the use of crypto, have mostly been ignored by corporations. Any attempts to outlaw crypto will produce a surge of sudden "incorporations," thus gaining for the new corporate members the aegis of corporate privacy.)

In an academic setting, "invisible colleges" are the communities of researchers.

These virtual communities typically are "opaque" to outsiders. Attempts to gain access to the internals of these communities are rarely successful. Law enforcement and intelligence agencies (such as the NSA in the U.S., Chobetsu in Japan, SDECE in France, and so on, in every country) may infiltrate such groups and use electronic surveillance (ELINT) to monitor these virtual communities. Not surprisingly, these communities are early adopters of encryption technology, ranging from scrambled cellphones to full blown PGP encryption.  [8]

The use of encryption by "evil" groups, such as child pornographers, terrorists, abortionists, abortion protestors, etc., is cited by those who wish to limit civilian access to crypto tools. We call these the "Four Horseman of the Infocalypse," as they are so often cited as the reason why ordinary citizen units of the nation-state are not to have access to crypto.

This is clearly a dangerous argument to make, for various good reasons. The basic right of free speech is the right to speak in a language one's neighbors or governing leaders may not find comprehensible: encrypted speech. There's not enough space here to go into the many good arguments against a limit on access to privacy, communications tools, and crypto.

The advent of full featured communications systems for computer-mediated virtual communities will have even more profound implications. MUDs and MOOs (multi-user domains, etc.) and 3D virtual realities are one avenue, and text-centric Net communications are another. (Someday, soon, they'll merge, as described in Vernor Vinge's prophetic 1980 novella, True Names.)

Observability and Surveillance

An interesting way to view issues of network visibility is in terms of the "transparency" of nodes and links between nodes. Transparent means visible to outsiders, perhaps those in law enforcement or the intelligence community. Opaque mean not transparent, not visible. A postcard is transparent, a sealed letter is opaque. PGP inventor Phil Zimmermann has likened the requirement for transparency to being ordered to use postcards for all correspondence, with encryption the equivalent of an opaque envelope (envelopes can be opened, of course, and long have been).

Transparent links and nodes are the norm in a police state, such as the U.S.S.R., Iraq, China, and so forth. Communications channels are tapped, and private use of computers is restricted. (This is becoming increasingly hard to do, even for police states; many cite the spread of communications options as a proximate cause of the collapse of communism in recent years.)

There are interesting "chemistries" or "algebras" of transparent vs. opaque links and nodes. What happens if links must be transparent, but nodes are allowed to be opaque? (The answer: the result is as if opaque links and nodes were allowed, i.e., full implications of strong crypto. Hence, any attempt to ban communications crypto while still allowing private CPUs to exist....)

If Alice and Bob are free to communicate, and to choose routing paths, then Alice can use "crypto arbitrage" (a variation on the term, "regulatory arbitrage," the term Eric Hughes uses to capture this idea of moving transactions to other jurisdictions) to communicate with sites - perhaps in other countries - that will perform as she wishes. This can mean remailing, mixing, etc. As an example, Canadian citizens who are told they cannot access information on the Homolka-Teale murder case (a controversial case in which the judge has ordered the media in Canada, and entering Canada, not to discuss the gory details) nevertheless have a vast array of options, including using telnet, gopher, ftp, the Web, etc., to access sites in many other countries or even in no country in particular.

Most of the consequences described here arise from this chemistry of links and nodes: unless nearly all node and links are forced to be transparent, including links to other nations and the nodes in those nations, then the result is that private communication can still occur. Crypto-anarchy results.

Crypto-anarchy

"The Net is an anarchy." This truism is the core of crypto anarchy. No central control, no ruler, no leader (except by example, reputation), no "laws." No single nation controls the Net, no administrative body sets policy. The Ayatollah in Iran is as powerless to stop a newsgroup - alt.wanted.moslem.women or alt.wanted.moslem.gay come to mind - he doesn't like as the President of France is as powerless to stop, say, the abuse of French in soc.culture.french. Likewise, the CIA can't stop newsgroups, or sites, or Web pages, which give away their secrets. At least not in terms of the Net itself...what non-Net steps might be taken is left as an exercise for the paranoid and the cautious.

This essential anarchy is much more common than many think. Anarchy - the absence of a ruler telling one what to do - is common in many walks of life: choice of books to read, movies to see, friends to socialize with, etc. Anarchy does not mean complete freedom - one can, after all, only read the books which someone has written and had published - but it does mean freedom from external coercion. Anarchy as a concept, though, has been tainted by other associations.

First, the "anarchy" here is not the anarchy of popular conception: lawlessness, disorder, chaos, and "anarchy." Nor is it the bomb throwing anarchy of the 19th century "black" anarchists, usually associated with Russia and labor movements. Nor is it the "black flag" anarchy of anarcho-syndicalism and writers such as Proudhon. Rather, the anarchy being spoken of here is the anarchy of "absence of government" (literally, "an arch," without a chief or head).

This is the same sense of anarchy used in "anarcho-capitalism," the libertarian free market ideology which promotes voluntary, uncoerced economic transactions. [ 6] I devised the term crypto-anarchy as a pun on crypto, meaning "hidden," on the use of "crypto" in combination with political views (as in Gore Vidal's famous charge to William F. Buckley: "You're crypto-fascist!"), and of course because the technology of crypto makes this form of anarchy possible. The first presentation of this was in a 1988 "Manifesto," whimsically patterned after another famous manifesto.  [7] Perhaps a more popularly understandable term, such as "cyber liberty," might have some advantages, but Crypto-anarchy has its own charm, I think.

And anarchy in this sense does not mean local hierarchies don't exist, nor does it mean that no rulers exist. Groups outside the direct control of local governmental authorities may still have leaders, rulers, club presidents, elected bodies, etc. Many will not, though.

Politically, virtual communities outside the scope of local governmental control may present problems of law enforcement and tax collection. (Some of us like this aspect.) Avoidance of coerced transactions can mean avoidance of taxes, avoidance of laws saying who one can sell to and who one can't, and so forth. It is likely that many will be unhappy that some are using cryptography to avoid laws designed to control behavior.

National borders are becoming more transparent than ever to data. A flood of bits crosses the borders of most developed countries' phone lines, cables, fibers, satellite up/downlinks, and millions of diskettes, tapes, CDs, etc. Stopping data at the borders is less than hopeless.

Finally, the ability to move data around the world at will, the ability to communicate to remote sites at will, means that a kind of "regulatory arbitrage" can be used to avoid legal roadblocks. For example, remailing into the U.S. from a site in the Netherlands...whose laws apply? (If one thinks that U.S. laws should apply to sites in the Netherlands, does Iraqi law apply in the U.S.? And so on.)

This regulatory arbitrage is also useful for avoiding the welter of laws and regulations which operations in one country may face, including the "deep pockets" lawsuits so many in the U.S. face. Moving operations on the Net outside a litigious jurisdiction is one step to reduce this business liability. Like Swiss banks, but different.

True Names and Anonymous Systems

Something needs to be said about the role of anonymity and digital pseudonyms. This is a topic for an essay unto itself, of course.

Are true names really needed? Why are they asked for? Does the nation-state have any valid reason to demand they be used?

People want to know who they are dealing with, for psychological/evolutionary reasons and to better ensure traceability should they need to locate a person to enforce the terms of a transaction. The purely anonymous person is perhaps justifiably viewed with suspicion.

And yet pseudonyms are successful in many cases. And we rarely know whether someone who presents himself by some name is "actually" that person. Authors, artists, performers, etc., often use pseudonyms. What matters is persistence, and nonforgeability. Crypto provides this.

On the Cypherpunks list, well-respected digital pseudonyms have appeared and are thought of no less highly than their "real" colleagues are.

The whole area of digitally authenticated reputations, and the "reputation capital" that accumulates or is affected by the opinions of others, is an area that combines economics, game theory, psychology, and expectations. A lot more study is needed.

It is unclear if governments will move to a system of demanding "Information Highway Driver's Licenses," figuratively speaking, or how systems like this could ever be enforced. (The chemistry of opaque nodes and links, again.)

Examples and Uses

It surprises many people that some of these uses are already being intensively explored. Anonymous remailers are used by tens of thousands of persons - and perhaps abused.  [13] And of course encryption, via RSA, PGP, etc., is very common in some communities. (Hackers, Net users, freedom fighters, white separatists, etc....I make no moral judgments here about those using these methods).

Remailers are a good example to look at in more detail.

There are two current main flavors of remailers:

The Cypherpunks remailers currently number about 20, with more being added every month. There is no reason not to expect hundreds of such remailers in a few years. One experimental "information market" is BlackNet, a system which appeared in 1993 and which allows fully-anonymous, two- way exchanges of information of all sorts. There are reports that U.S. authorities have investigated this because of its presence on networks at Defense Department research labs. Not much they can do about it, of course, and more such entities are expected.

(The implications for espionage are profound, and largely unstoppable. Anyone with a home computer and access to the Net or Web, in various forms, can use these methods to communicate securely, anonymously or pseudonymously, and with little fear of detection. "Digital dead drops" can be used to post information obtained, far more securely than the old physical dead drops...no more messages left in Coke cans at the bases of trees on remote roads.)

Whistleblowing is another growing use of anonymous remailers, with folks fearing retaliation using remailers to publicly post information. (Of course, there's a fine line between whistle blowing, revenge, and espionage.)

Data havens, for the storage and marketing of controversial information is another area of likely future growth. Nearly any kind of information, medical, religious, chemical, etc., is illegal or proscribed in one or more countries, so those seeking this illegal information will turn to anonymous messaging systems to access and perhaps purchase, with anonymous digital cashthis information. This might include credit data bases, deadbeat renter files, organ bank markets, etc. (These are all things which have various restrictions on them in the U.S., for example....one cannot compile credit data bases, or lists of deadbeat renters, without meeting various restrictions. A good reason to move them into cyberspace, or at least outside the U.S., and then sell access through remailers.)

Matching buyers and sellers of organs is another such market. A huge demand (life and death), but various laws tightly controlling such markets.

Digital cash efforts. A lot has been written about digital cash.   [14]  [15] David Chaum's company, DigiCash, has the most interesting technology, and has recently begun market testing. Stefan Brands may or may not have a competing system which gets around some of Chaum's patents. (The attitude crypto anarchists might take about patents is another topic for discussion. Suffice it to say that patents and other intellectual property issues continue to have relevance in the practical world, despite erosion by technological trends.)

Credit card - based systems, such as the First Virtual system, are not exactly digital cash, in the Chaumian sense of blinded notes, but offer some advantages the market may find useful until more advanced systems are available. I expect to see many more such experiments over the next several years, and some of them will likely be market successes.

Commerce and Colonization of Cyberspace

How will these ideas affect the development of cyberspace? "You can't eat cyberspace" is a criticism often levelled at argument about the role of cyberspace in everyday life. The argument made is that money and resources "accumulated" in some future (or nearfuture) cyberspatial system will not be able to be "laundered" into the real world. Even such a prescient thinker as Neal Stephenson, in Snow Crash, had his protagonist a vastly wealthy man in "The Multiverse," but a nearpauper in the physical world."

This is implausible for several reasons. First, we routinely see transfers of wealth from the abstract world of stock tips, arcane consulting knowledge, etc., to the real world. "Consulting" is the operative word. Second, a variety of means of laundering money, via phony invoices, uncollected loans, art objects, etc., are well-known to those who launder money...these methods, and more advanced ones to come, are likely to be used by those who wish their cyberspace profits moved into the real world.

(Doing this anonymously, untraceably, is another complication. There may be methods of doing this. Proposals have looked pretty solid, but more work is needed.)

The World Wide Web is growing at an explosive pace. Combined with cryptographically-protected communication and digital cash of some form (and there are several being tried), this should produce the long-awaited colonization of cyberspace. Most Net and Web users already pay little attention to the putative laws of their local regions or nations, apparently seeing themselves more as members of various virtual communities than as members of locally-governed entities.

This trend is accelerating.

Most importantly, information can be bought and sold (anonymously, too) and then used in the real world. There is no reason to expect that this won't be a major reason to move into cyberspace.

Implications

I've touched on the implications in several places. Many thoughtful people are worried about some of the possibilities made apparent by strong crypto and anonymous communication systems. Some are proposing restrictions on access to crypto tools. The recent debate in the U.S. over "Clipper" and other key escrow systems shows the strength of emotions on this issue.

Abhorrent markets may arise. For example, anonymous systems and untraceable digital cash have some obvious implications for the arranging of contract killings and such. (The greatest risk in arranging such hits is that physical meetings expose the buyers and sellers of such services to stings. Crypto-anarchy lessens, or even eliminates, this risk, thus lowering transaction costs. The risks to the actual triggermen are not lessened, but this is a risk the buyers need not worry about. Think of anonymous escrow services which hold the digital money until the deed is done. Lots of issues here. It is unfortunate that this area is so little-discussed....people seem to have an aversion for exploring the logical consequences in such areas.) The implications for corporate and national espionage have already been touched upon. Combined with liquid markets in information, this may make secrets much harder to keep. (Imagine a "Digital Jane's," after the military weapons handbooks, anonymously compiled and sold for digital money, beyond the reach of various governments which don't want their secrets told.)

New money-laundering approaches are of course another area to explore.

Something that is inevitable is the increased role of individuals, leading to a new kind of elitism. Those who are comfortable with the tools described here can avoid the restrictions and taxes that others cannot. If local laws can be bypassed technologically, the implications are pretty clear.

The implications for personal liberty are of course profound. No longer can nation-states tell their citizen- units what they can have access to, not if these citizens can access the cyberspace world through anonymous systems.

How Likely?

I am making no bold predictions that these changes will sweep the world anytime soon. Most people are ignorant of these methods, and the methods themselves are still under development. A wholesale conversion to "living in cyberspace" is just not in the cards, at least not in the next few decades.

But to an increasingly large group, the Net is reality. It is where friends are made, where business is negotiated, where intellectual stimulation is found. And many of these people are using crypto-anarchy tools. Anonymous remailers, message pools, information markets. Consulting via pseudonyms has begun to appear, and should grow. (As usual, the lack of a robust digital cash system is slowing things down.

Can crypto-anarchy be stopped? Although the future evolution in unclear, as the future almost always is, it seems unlikely that present trends can be reversed:

As John Gilmore has noted, "the Net tends to interpret censorship as damage, and routes around it." This applies as well to attempts to legislate behavior on the Net. (The utter impossibility of regulating the worldwide Net, with entry points in more than a hundred nations, with million of machines, is not yet fully recognized by most national governments. They still speak in terms of "controlling" the Net, when in fact the laws of one nation generally have little use in other countries.)

Digital money in its various forms is probably the weakest link at this point. Most of the other pieces are operational, at least in basic forms, but digital cash is (understandably) harder to deploy. Hobbyist or "toy" experiments have been cumbersome, and the "toy" nature is painfully obvious. It is not easy to use digital cash systems at this time ("To use Magic Money, first create a client..."), especially as compared to the easily understood alternatives.  [12] People are understandably reluctant to entrust actual money to such systems. And it's not yet clear what can be bought with digital cash (a chicken or egg dilemma, likely to be resolved in the next several years).

And digital cash, digital banks, etc., are a likely target for legislative moves to limit the deployment of crypto anarchy and digital economies. Whether through banking regulation or tax laws, it is not likely that digital money will be deployed easily. "Kids, don't try this at home!"

Some of the current schemes may also incorporate methods for reporting transactions to the tax authorities, and may include "software key escrow" features which make transactions fully or partly visible to authorities.

Conclusions

Strong crypto provides new levels of personal privacy, all the more important in an era of increased surveillance, monitoring, and the temptation to demand proofs of identity and permission slips. Some of the "credentials without identity" work of Chaum and others may lessen this move toward a surveillance society.

The implications are, as I see it, that the power of nationstates will be lessened, tax collection policies will have to be changed, and economic interactions will be based more on personal calculations of value than on societal mandates.

Is this a Good Thing? Mostly yes. Crypto-anarchy has some messy aspects, of this there can be little doubt. From relatively unimportant things like price fixing and insider trading to more serious things like economic espionage, the undermining of corporate knowledge ownership, to extremely dark things like anonymous markets for killings.

But let's not forget that nationstates have, under the guise of protecting us from others, killed more than 100 million people in this century alone. Mao, Stalin, Hitler, and Pol Pot, just to name the most extreme examples. It is hard to imagine any level of digital contract killings ever coming close to nationstate barbarism. (But I agree that this is something we cannot accurately speak about; I don't think we have much of a choice in embracing crypto-anarchy or not, so I choose to focus on the bright side.)

It is hard to argue that the risks of anonymous markets and tax evasion are justification for worldwide suppression of communications and encryption tools. People have always killed each other, and governments have not stopped this (arguably, they make the problem much worse, as the wars of this century have shown).

Also, there are various steps that can be taken to lessen the risks of crypto-anarchy impinging on personal safety.  [10]

Strong crypto provides a technological means of ensuring the practical freedom to read and write what one wishes to. (Albeit perhaps not in one's true name, as the nation-state- democracy will likely still try to control behavior through majority votes on what can be said, not said, read, not read, etc.) And of course if speech is free, so are many classes of economic interaction that are essentially tied to free speech.

A phase change is coming. Virtual communities are in their ascendancy, displacing conventional notions of nationhood. Geographic proximity is no longer as important as it once was.

A lot of work remains. Technical cryptography still hasn't solved all problems, the role of reputations (both positive and negative) needs further study, and the practical issues surrounding many of these areas have barely been explored. We will be the colonizers of cyberspace.

Acknowledgments

My thanks to my colleagues in the Cypherpunks group, all 700 of them, past or present. Well over 100 megabytes of list traffic has passed through he Cypherpunks mailing list, so there have been a lot of stimulating ideas. But especially my appreciation goes to Eric Hughes, Sandy Sandfort, Duncan Frissell, Hal Finney, Perry Metzger, Nick Szabo, John Gilmore, Whit Diffie, Carl Ellison, Bill Stewart, and Harry Bartholomew. Thanks as well to Robin Hanson, Ted Kaehler, Keith Henson, Chip Morningstar, Eric Dean Tribble, Mark Miller, Bob Fleming, Cherie Kushner, Michael Korns, George Gottlieb, Jim Bennett, Dave Ross, Gayle Pergamit, andespeciallythe late Phil Salin. Finally, thanks for valuable discussions, sometimes brief, sometimes long, with Vernor Vinge, David Friedman, Rudy Rucker, David Chaum, Kevin Kelly, and Steven Levy.


References and Notes

(1) The Cypherpunks group was mainly formed by Eric Hughes, Tim May, and John Gilmore. It began both physical meetings, in the Bay Area and elsewhere, and virtual meetings on an unmoderated mailing list. The name was provided by Judith Milhon, as a play on the "cyberpunk" genre and the British spelling of cipher. The mailing list can be subscribed to by sending the single message, subscribe cypherpunks, in the body of a message to majordomo@toad.com. Expect at least 50 messages a day. About 600 subscribers in many countries are presently on the list. Some are pseudonyms.

Return to Introduction

(2) RSA Data Security Inc., Redwood Shores, California, is the license administrator. Contact them for details.

Return to white paper

(3) Many crypto texts exist. A good introduction is Bruce Schneier's Applied Cryptography, John Wiley and Sons, 1994. This text includes pointers to many other sources. The "Crypto" Proceedings (Advances in Cryptology, Springer- Verlag, annually) are essential references. The annual Crypto conference in Santa Barbara, and the Eurocrypt and Auscrypt conferences, are where most crypto results are presented.

Return to white paper

(4) David Chaum, "Untraceable Electronic Mail, Return Addresses, and Digital Pseudonyms," Comm. ACM 24, 2, February 1981, pp. 8488. Cypherpunks-style remailers are a form of Chaum's "digital mixes," albeit far from ideal.

Return to white paper

(5) David Chaum, "Security without Identification: Transaction Systems to make Big Brother Obsolete," Comm. ACM 28, 10, October 1985. This is an early paper on digital cash...be sure to consult more recent papers.

Return to white paper

(6) David Friedman, The Machinery of Freedom, 2nd edition. A leading theoretician of anarchocapitalism. (Hayek was another.)

Return to white paper

(7) Tim May, The Crypto-Anarchist Manifesto, July 1988, distributed on the Usenet and on various mailing lists.

Return to white paper

(8) The political opposition in Myan Mar - formerly Burma - is using Pretty Good Privacy running on DOS laptops in the jungles for communications amongst the rebels, according to Phil Zimmermann, author of PGP. This life-and-death usage underscores the role of crypto.

Return to white paper

(9) See Kevin Kelly's Out of Control, 1994, for a discussion of how central control is failing, and how the modern paradigm is one of market mechanisms, personal choice, and technological empowerment.

Return to white paper

(10) Robin Hanson and David Friedman have written extensively about scenarios for dealing with the threats of extortionists, would-be assassins, etc. I am hoping some of their work gets published someday. (Much of the discussion was in 1992-3, on the "Extropians" mailing list.)

Return to white paper

(11) During the "Digital Telephony Bill" debate, an FBI official said that failure to mandate wiretap capabilities within the next 18 months would make it all moot, as the cost would rise beyond any reasonable budget (currently $500 million for retrofit costs).

Return to white paper

(12) "Magic Money" was an experimental implementation of Chaum's digital cash system. It was coded by "Pr0duct Cypher," a pseudonymous member of the Cypherpunks list. None of us knows his real identity, as he used remailers to communicate with the list, and digitally signed his posts. Many of us found it too difficult to use, which is more a measure of the deep issues involved in using digital analogs (no pun intended) to real, physical money.

Return to white paper

(13) Abuse, according to some views, of remailers is already occurring. A Cypherpunkstype remailer was used to post a proprietary hash function of RSA Data Security, Inc. to the Usenet. (Let me hasten to add that it was not a remailer I operate, or have control over, etc.)

Return to white paper

(14) article on digital cash, The Economist, 26 November 1994. pp. 2123.

Return to white paper

(15) article on digital cash, Steven Levy, Wired. December 1994.

Return to white paper