Monday, September 8, 1997
"THIS IS A COOL HOLIDAY," says Sameer Parekh over a July 4 breakfast in a cafe near the University of California at Berkeley. "It's the day we celebrate overthrowing the government."
A disheveled 22-year-old, 135 pounds, shirttails down to the knees of his jeans, with a 4-inch black goatee hanging from a cherubic face, Parekh is no violent revolutionary out to establish a dictatorship of the proletariat. Parekh is a libertarian of a new sort. His weapon: software.
Parekh traffics in a substance known among his peers as "strong crypto," cryptographic software massively stronger than the stuff American companies are allowed to export. Cryptography is the science of scrambling messages so they cannot be read by prying eyes. It is the lifeblood of telephone commerce;credit card verifications, bank teller machine transactions, wire transfers. It is useful to crooks. And it is magnificently antiauthoritarian.
Encrypted with a sufficiently powerful code, a cellular phone conversation becomes untappable, a written message or computer file indecipherable. Federal authorities are attempting to limit the spread of this technology abroad. But they are no match for Parekh and other rebels with his programming skills.
For the last three years Parekh has been mixing sophisticated computer science with libertarian philosophy, selling a cryptographic product made in an undisclosed foreign country through an Anguillan subsidiary. His company, C2Net, thereby skirts U.S. export restrictions.
Looking further out, cryptography's challenge to Washington's authority;indeed, to that of all governments;is daunting. Even if the federal government can somehow keep strong crypto out of the hands of Muammar Qaddafi; extremely doubtful at this point;it would still have all manner of domestic users to worry about.
"I realized that protection of privacy on the Internet couldn't be viable without a viable business behind it."-Sameer Parekh
Cryptography is very useful to anyone who can't afford to leave behind a paper trail. That could be someone running an illegal gambling business or doing insider trading or distributing child pornography or arranging the details of a cocaine shipment. It could also be someone who is a perfectly legitimate business operator except that he doesn't want to pay income taxes or otherwise submit his transaction to the prying eyes of increasingly intrusive governments.
Parekh envisions a revolution in which federal buildings don't burn to the ground but rather just run out of money. There would still be a government, but it would not be the expansive welfare state we have today. It would be a minimalist version of the sort seen in a place like Hong Kong;strong on law and order, sanctity of contract and minimal social security but that's about it.
Walter Wriston, former chairman of Citibank, devoted a chapter of his brilliant 1992 book The Twilight of Sovereignty to the history of cryptography. Wriston foresaw the weakening of national governments through the power of technology and recognized that cryptography would play a key role. He knew something about it from personal experience. During World War II he was responsible for the electromechanical devices used by the Allies to encrypt their messages. Wriston sees encryption technology as a key ingredient in the transfer of social and economic power from the governments of nation-states to the PC-packing populace. Since the success of Internet commerce depends on strong cryptography, its proliferation is inevitable. "The government can't do much about it," says Wriston. "It's another thing slipping through their fingers."
Rejoice, libertarians. Lament, Hillary Clinton and partisans of the nanny state. If you want to participate in the cryptographic revolution against Big Government, you don't have to traffic with an arms merchant in a dark alley. Go to the Internet. Pretty Good Privacy (PGP), from $99 to $249, is a popular program. Another one is SynCrypt, by SynData Technologies Inc., just out.
Using this off-the-shelf stuff you can transact business in total privacy. Don't worry about spies. With what is presently known about code cracking, it would take a supercomputer a billion years to divine your message.
There is another dimension to the spread of crypto. The same mathematical tricks used to encode a message can be run in reverse, to generate a so-called digital signature. This is a computer stamp of authenticity. It can be used to prove that an electronic document originated with a particular sender, such as a bank depositor or a bank officer. Assemble a few digital signatures in a clever fashion and you have created a mechanism for digital cash;a system of electronic payments akin to Visa or MasterCard but with the added feature that it can be made anonymous.
Think about that. Money transfers that are genuine but untraceable. Anonymous, secure E-cash could give rise to a blossoming of commerce on the Internet and a reduction in the billions of dollars spent annually processing paper checks and paper credit-card chits.
Bad news, of course, for the Internal Revenue Service and its 3,570-page maze of a tax code. And what happens to the Fed's control of the money supply when more and more money takes the form of digital blips on a satellite in the sky? How do you stop money laundering once cash is invisible and leaves no paper trail?How do you catch tax dodgers?
"It's easily the most important privacy issue of the decade, and perhaps the most important policy issue."-David Friedman
The IRS figures that it is already losing $120 billion a year on income that goes unreported. When E-cash becomes commonplace, that number is going to get larger. The underground economy, after all, does surface at times. Dogs can sniff the traces of cocaine in a satchel of bills. In the ionosphere economy there is no odor for dogs to sniff.
For the libertarian set, today's encryption technology is the best thing to come along since the right to bear arms. After all, why risk getting arrested for dumping tea into the harbor when you can just order the tea from a tax-free jurisdiction over the Net, encrypt the purchase order and pay with anonymous digital currency? Libertarians see encryption technology as the weaponry for a bloodless grassroots revolution in which revenue streams replace street barricades as the fields of battle.
"We are looking at kidnappers, we are looking at terrorists, we are looking at banking integrity, we are looking at propriety interests and economic secrets," Louis Freeh, director of the Federal Bureau of Investigation, told members of the International Cryptography Institute two years ago. Freeh has been stumping for tougher government controls on encryption technology. He wants a "key escrow" bureaucracy that would hold cryptographic keys that could, if law enforcement deemed it necessary, be used to unscramble any encrypted message.
But do we want to put that much power in the hands of bureaucrats? "Back in England, when the king wanted to smoke out people we'd call terrorists today;the people we see in retrospect as patriots;he wanted to steam open envelopes," scoffs Wriston. "Nothing has changed;now governments want to steam open your E-mail. If I were the national drug czar, I'd want to, too. The problem is, none of us trust the government to limit that interception to those particular messages."
If the FBI is threatened, that doesn't entirely dismay the libertarian crowd that seems to be overrepresented in the hacker community. "With the Internet being ubiquitous and crypto being cheap and easy to get, it's going to be more and more difficult for governments to control transactions between people," declares Adam Shostack, a 24-year-old cryptography consultant whose clients include Fidelity Investments. In February he instructed attendees at a financial cryptography conference in the finer points of using encryption to protect large networks from attackers and con men. Site of the conference: the Caribbean tax haven of Anguilla.
It's too early to give it a name, but computer technology and modern communications are at the threshold of creating a new kind of political movement. Talk to David Friedman, a professor of economics at the University of California at Santa Clara. Friedman espouses an anti-big-government philosophy a little stronger even than that of his famous father, Milton Friedman. He calls the set of questions raised by encryption "easily the most important privacy issue of the decade, and perhaps the most important policy issue." He concedes the downside to a technology that will be useful to lawbreakers. But he says the advantages of pervasive privacy outweigh the disadvantages: "On the whole, it'll be a change to a freer and more interesting society."
Cryptography, the craft of secret writing, has been around almost as long as writing itself. Bad guys have always used it. So have rebels. Its better-known applications through the ages have been in making secure military plans and espionage communiques. Today's biggest user is, if not the government, the bank industry. Encryption safeguards the more than $1 trillion a day that flows over the Fedwire and the Chips systems.
Modern cryptography was born two decades ago at Stanford University with the invention of so-called public key encryption by Whitfield Diffie and Martin Hellman. In classic cryptography, keys were kept private. The sender would use a key, or formula, to encode a document; the receiver would use a closely related formula to decode. To communicate, the sender and receiver would have to share a key. This was usually the weak spot. A messenger sent to transfer the key could be intercepted or compromised.
With public-key encryption, this problem is finessed. The receiver of confidential messages simultaneously creates an encoding and a decoding key. The peculiar arithmetic of these keys, perfected by a trio of MIT mathematicians, is such that the one cannot be divined from the other: Knowing the encrypting formula tells you absolutely nothing about how to unscramble a message. So the receiver need not be particular about his choice of messenger to deliver the encrypting key. Indeed, he can afford to publish the key for all the world to see. Modern-day practice is to dump the key onto an Internet home page or server.
What makes encryption a killer application just now? The MIT algorithm requires that both sender and receiver do several billion calculations on each message, a practical impossibility not too long ago. Moore's Law to the rescue. The doubling of computing power every 18 months has placed the ability to process virtually unbreakable cryptographic algorithms within reach of anyone with a 166-megahertz Pentium.
Legitimate users? Any company planning on doing business on-line. When you send an order over the Internet, the contents of your message pass through a series of network routers and servers before reaching their final destination. Anyone who gains control of one of the machines along the way could intercept your credit card information. You're not going to send in the order unless you know it is secure.
Illegitimate ones? This is an imaginary scenario. You work at Apple and know, two days before it is to be announced, that Microsoft is going to pump in some cash and probably give a kick to the stock price. You're going to tip off your brother-in-law, who is going to feed the tip to a third party, an active trader in technology stocks. Do this with phone calls and you stand a fairly high risk of being caught, even though you have never met the trader.
So you encrypt the stock tip with your brother-in-law's public key and publish it on the Internet, perhaps in the middle of a chat room that lots of people visit. Your brother-in-law does the same, using the trader's public key. Both of these messages look like meaningless garbles to an outsider. They betray nothing about whose key was used to encrypt them.
Convicted insider trader Dennis Levine used a secret account in the Caribbean. But how do you use an offshore account without going through customs or making tappable phone calls? Given the power of encryption and digital signatures, a modern-day Levine could do anonymous E-trading from the comfort of his home PC, without making any suspicious phone calls or getting on a plane. Might the government have to throw up its hands someday, accepting the libertarian view that laws against insider trading just impede the efficiency of the marketplace?
"With crypto being easy to get, it's going to be difficult for governments to control transactions between people."-Adam Shostack
Someday it might. In the meantime, the government is trying to put the crypto genie back into the bottle. The current export controls permit the sale of weak crypto (the sort that could be cracked by the National Security Agency) but not crypto that would take the NSA a trillion years to crack.
The problem is that the basic tricks are widely known; indeed, the equation that drives these public key systems was published by a Swiss mathematician in the 1760s. Any reasonably competent Russian programmer can reinvent the software from scratch, and that is just what is happening (see box, p. 174). At this point, trying to regulate cryptography is like trying to cut the murder rate by regulating the sale of kitchen knives.
The next battleground will be fought over digital cash. One system, Mondex, has been adopted by an international consortium of banks led by National Westminster. The system, however, includes a digital trail that could be subpoenaed.
Not good enough, says cryptographer David Chaum, whose rival product is anonymous and untraceable;except to the spender. Customers want anonymity, he says;that's why 2.6 billion $100 bills are in circulation.
Chaum distances himself from the libertarian crowd, but his preaching about getting snoops off our backs is music to their ears. His invention may very well lead to the society espoused by Friedman and Parekh, in which widespread encryption forces the government to accept a less intrusive role in information flow.
Now the government polices what can be claimed about a prescription drug, what can be said in a real estate ad, who can talk about a publicly traded stock and when, and who can finance a political ad. Is all this necessary to preserve the union and insure domestic tranquility? The answer depends on your politics. If you are a Big Government liberal or Big Government right-winger, the answer is yes. But not everyone thinks we need as much government as we currently have.
When he was a 16-year-old high-school student in Libertyville, Ill., Sameer Parekh typed Henry David Thoreau's 9,000-word essay "Civil Disobedience" into an Apple II GS computer and posted it to an electronic bulletin board. The treatise begins: "I heartily accept the motto 'That government is best which governs least,' and should like to see it acted up to more rapidly and systematically. Carried out, it finally amounts to this, which also I believe,;'That government is best which governs not at all.'"
Look Thoreau's "Civil Disobedience" up on the World Wide Web and chances are you'll see "typed by: Sameer Parekh....1/12/1991" at the end of the text. Over the years, scores of people have made copies of the document and posted them, complete with Parekh's name and typos, on their own Web sites.
Thoreau's essay paraphrases another hero of Parekh's, Thomas Jefferson. It happens that one of Jefferson's many passions, along with fighting to keep the fledgling U.S. government as small as possible, was cryptography. In the 1790s he invented an elegant, handheld rotary cipher machine. What could he have done with a laptop!
There is something about the Internet that brings out resistance to authority. "Libertarianism is much more important in cyberspace than in real space," says David Friedman. "Nearly all political discussion on-line is pro- or antilibertarian. Libertarianism is the central axis."
Libertarianism as a central axis? Perhaps. The cyberheads have always been defiant of authority, going back to the Phone Phreaks of the 1970s, who used their knowledge of electronics to beat Ma Bell out of long-distance charges. Some of the phone acrobats evolved into today's self-styled "cypherpunks," a term combining the science fiction genre cyberpunk with the British spelling of "cipher." The cypherpunk clan (check out the Internet newsgroup "alt.cypherpunks") includes John Gilmore, one of the first employees at Sun Microsystems and an early member of the Electronic Frontier Foundation. The libertarian axis is particularly strong in the EFF, which defends hackers and cryptographers against their federal adversaries.
At Berkeley, Parekh programmed E-mail servers to allow subscribers to send and receive E-mails anonymously or under pseudonyms, and to surf the Web through a specially programmed "anonymizer" server without leaving the electronic trail that could let the Web site operators;or an enforcer for the Securities & Exchange Commission; know who visited and when. After dropping out of Berkeley in 1995, he went full time into the business of protecting Web surfers' identities.
Realizing that his subscribers' privacy was only as secure as the servers their accounts sat in, Parekh invited his friends in the hacker community to try to break into Web servers sold by Microsoft and Netscape. Successful break-ins were rewarded with T shirts. The results ("I had to stop giving T shirts out") convinced him that the real money to be made was in selling crackproof, Web-server software.
With Microsoft and Netscape restrained by U.S. law from exporting server software with strong encryption, Parekh saw an opportunity. He took a copy of Apache, a popular server software package available free on the Internet, and set about the arduous task of weaving heavy-duty encryption programs into the server software. Once he had figured out how to do that, Parekh contracted with programmers in a country he won't name (lest the U.S. lean on the country to tighten up its cryptography export laws) to write the software and formed a sister corporation in the Caribbean tax haven of Anguilla to sell it to the rest of the world.
If you don't like the tax rate or export laws in your native country, set up a Web server in a Caribbean tax haven, on the Isle of Man or on Vanuatu, incorporate there and run your business from anywhere over the Internet. Countries like those use low taxes and secrecy protection to compete for corporate custom.
As the world economy becomes less land- and factory-based and increasingly server-based, expect more nations to welcome boundary-jumping business. Encryption provides two essential functions in this sort of economy: It keeps transactions secure as they course along the world's networks, and it makes the nature of the transactions invisible to the prying eyes of border guards and tax collectors.
After less than a year, C2Net's encryption-studded software is running on about 30,000 domains;more than any other commercial software outside of the stuff sold by Netscape and Microsoft. In the process, the U.S. lost a few high-tech jobs, not to mention the taxes on the sales of the software.
As more and more business transactions are hidden from the IRS by encryption, Parekh predicts, tax revenues will decrease. Declining tax revenues will lead to the privatization of many of the government's present functions, and people will be more free to choose what services to spend their money on. They will even be free to choose what kind of money they will use.
Maybe it won't be the Federal Reserve's notes. Curiously enough, Fed Chairman Alan Greenspan is not entirely unsympathetic to the libertarian aim of taking money out of the exclusive control of the federal government. Greenspan was a libertarian in his youth;a regular in the salon of Ayn Rand (1905-1982). Greenspan predicts that electronic commerce will give rise to private currencies.
"As the international financial system becomes ever more complex," Greenspan said at a Treasury conference last year, "we, in our regulatory roles, are being driven increasingly toward reliance on private market self-regulation similar to what emerged in more primitive forms in the 1850s in the United States."
Greenspan made it clear that he did not think the government should seek to stem the tide, even though it will undermine the authority of organizations like his Federal Reserve Board. "I am especially concerned," he went on, "that we not attempt to impede unduly our newest innovation, electronic money, or, more generally, our increasingly broad electronic payments system."
Ian Goldberg is working on that very innovation, folding encryption schemes together to create a universal digital currency that will incorporate all the disparate forms of digital cash in the electronic marketplace. Goldberg, 24, a Canadian graduate student at Berkeley, spends most of his time tinkering with and poking into cryptosystems.
In September 1995 Goldberg sent a chill down Netscape's spine (and a ripple through its stock price) when he announced that he and a colleague, David Wagner, had found a major vulnerability in the security layer of the Navigator Web browser. Sixteen months later, in response to a challenge by security software company RSA, Inc., Goldberg devised a program that harnessed the spare computing cycles of about 250 assorted workstations in Berkeley's computer science department to attack a message encrypted with a 40-bit key, the government's limit on unregistered encryption software for export. Trying 100 billion keys an hour, Goldberg's gang broke the cipher in 31/2 hours. The experiment got noticed.
Where does all this end? Goldberg predicts that tax laws and commercial regulations will need to change to adapt to the world of encrypted on-line business. The FBI will have to go after bombers by keeping an eye on ammonium nitrate rather than its ear on phone lines. "Taxes will have to be based more on physical things like land; assuming one believes in taxation at all," he says. "With encryption, not only can you hide your transactions, but your assets as well. Intellectual property can be hidden easily."
As for governmental restrictions on encryption, Goldberg finds them more ridiculous than pernicious. "I don't think terrorists will say, 'Since there's a law against strong cryptography, we won't use it.'"
Against all this governments are fighting a battle they have no prayer of winning, says Walter Wriston. Fortified with strong cryptography and growing exponentially, Wriston says, the Internet will irrevocably weaken governments as we know them. "They haven't got a chance in hell with that thing," he chuckles. "There's no way anybody can control it."
We do not know where all this will end, and neither does anyone else, but for better or worse, the implications for politics, for economics and for human freedom are enormous. The 20th century was the century of Big Wars and Big Governments; fascist, communist, welfare state. The 21st century is going to be something quite different.
This was the cover story for the September 1997 issue of Forbes.
This was the cover story for the September 1997 issue of Forbes.