An e-mail exchange with Gnutella developer Gene Kan in The Atlantic Online dated September 2000

One of the founders of a startup currently known as, Gene Kan, who is in his early twenties, formerly worked at the network-security firm Check Point Software and at a company that analyzed network traffic. Last March, when the first version of Gnutella was launched -- and then immediately yanked -- by the AOL subsidiary Nullsoft, Kan was one of the first programmers to see the power of the idea behind it: direct person-to-person sharing of information across the Internet. To him, the idea of using the Internet to help people distribute material far and wide -- rather than get it from centralized sources -- seemed like the essence of what computer networking was about. He soon began working on a new version of Gnutella with a source code (an underlying instruction set) that would be available to all. And with the help of people like Marc Andreesen, the co-founder of Netscape, he launched the new firm to pursue what he believes will be the many commercial applications of person-to-person sharing of information. Over the course of a week or so, he exchanged e-mail with Charles C. Mann of The Atlantic Monthly.


Atlantic Online: Gnutella is often discussed in the media as if it were a piece of software expressly designed for the purpose of bankrupting the record labels by facilitating music piracy. I know from talking with you that this is not how you think about what you're doing. Could you tell me what you're hoping for as you work on Gnutella?

Gene Kan: First, Gnutella was invented by Justin Frankel and Tom Pepper. I became involved after it was bastardized and subsequently embraced by the "open source community."

I believe that Frankel and Pepper's intention was to present a technology which the world hadn't thought of. Distributed technologies haven't been very well studied and aren't very well understood. And most computer scientists still don't have a good understanding of Gnutella.

Gnutella and things like it are clearly what's next for Internet computing. The Internet has always been about connecting people. For decades, Internet technology development leaned toward connecting individuals. Things like e-mail, USENET, IRC, etc. Even the World Wide Web.

Unfortunately, the World Wide Web has worked out to be an experience that mainly enables a large entity to communicate with a large number of individuals. It hasn't shaped up as the conduit for person-to-person interaction.

So-called peer-to-peer technologies such as Gnutella, ICQ, AIM, Napster, etc., make individuals as valuable as large corporations in providing knowledge. They fill the void left by the Web.

Gnutella, particularly, is an important technology in this domain. It is the first truly usable decentralized file-location technology. (Note that there is an important distinction between file sharing and file location.) And in fact Gnutella operates much like real people operate: you connect to a few friends, and you get to network with your friends' friends.

That's incredibly powerful. Computers basically help us do what we normally do, only faster. In general, technologies that don't fit into that model aren't found to be very useful.

And the way Gnutella connects these people is unique to a category of technologies that is populated only by Gnutella and FreeNet. Gnutella and FreeNet are really the only available technologies where tracking an individual's activities is incredibly difficult.

Just yesterday the Chinese government broke up an anti-government Web site. Whoops. Without anonymous information-exchange technologies such as Gnutella and FreeNet, it is incredibly easy to shut down undesirable Web sites. Gnutella and FreeNet could well be the future of free speech in the virtual world.

Gnutella's future is bright. Lots of developers worldwide are working to improve Gnutella software and Gnutella technology.

Atlantic Online: You spent some time just now discussing the free-speech implications of Gnutella. (You mentioned FreeNet, too, but I'll confine my question to Gnutella, because that's what you're working on.) Because Gnutella permits people to locate and exchange information in a truly decentralized way, it's hard for anyone to listen in. As you point out, this would be invaluable for dissidents everywhere. But the same capacities make Gnutella anathema to record companies and movie studios. It sounds like you're implicitly saying: "In a world where enabling free speech may necessarily entail hurting record- and movie-company profits, I'll promote free speech because it is so much more important -- and tell the record companies that they're simply going to have to adapt, as they've adapted before." Is that a reasonable summary of your feelings on this issue?

Gene Kan: Yes and no. I don't smile and say, "Well, you guys are just going to have to adapt." And I particularly don't attribute the tradeoff to freedom of speech, at least not at first blush.

There are more immediate and obvious concerns. Technology is constantly advancing. Any enterprise that predicates its survival on an unchanging technological landscape is doomed. And in this case, it's more than just technology: listeners are demanding downloadable formats.

The music industry is far from doomed, I think. They've come to embrace every technology thrown into their path. Audio-cassette recorders caused a big stink thirty years ago, but over time that technology was brought into the fold.

The Internet has put a lot of pressure on all kinds of industries. Newspapers, magazines, and ... antiques. The antique story is interesting. eBay could well put many antique shops out of business. But those who adapt and sell their wares on eBay -- where there is a big liquid market for antiques and unique items -- will benefit from much higher liquidity than in their brick-and-mortar storefront.

The same will apply to music.

Anyway, whether or not law-enforcement bodies can listen into our peer-to-peer discussions, and whether or not free speech is benefited by peer-to-peer technologies, record companies are going to have to adapt to survive. Every other industry -- banking, stock trading, publishing, oil, steel, and on and on -- has had to adapt.

Atlantic Online: But the record companies now are saying very loudly that with systems like Gnutella they can't survive. What do you say to that?

Gene Kan: It pains me to say it, but I'm going to make an unbiased observation:

Industries are occasionally made obsolete by technological advances. Carrier pigeons, for example, are largely on the dole these days.