📸

grammers outside of Nullsoft began collaborating to reverse-engineer, recode, and improve the program. As they made progress, the twenty-three-year-old Kan offered to help. “I was late to the game,” he said. But when an Associated Press reporter contacted some of the real leaders of the project, they referred the reporter to Kan. A photographer showed up to take his picture, and Kan suddenly became the face of what was being hailed as the most likely Napster successor.

Kan stuck around the open-source Gnutella effort for a few months. In addition to his unofficial-spokesman role, Kan’s personal connections at Nullsoft are extremely likely to have enabled him to help the other programmers. Giving away the source code developed on AOL’s nickel would have violated Frankel and Pepper’s obligations to their employer. It also might have opened them and the company to massive liability if the record industry sued over the program. While Kan insisted in an interview that the open-source Gnutella code was redesigned, not leaked, there is an area in between, filled with helpful hints, suggestions, and carefully worded questions about what methods the new developers were trying. Asked if the Nullsoft team at least gave the open-sourcers direction by means of winks and nods, Kan remained wordless. Then he winked, and he nodded.

Along WITH SOME TECHNICAL DRAWBACKS to a decentralized system like Gnutella, there was a profound economic one: There was no clear way for the creator of a system of ad hoc networks to make any money off of it. So Kan and others who were more business oriented began rethinking the entire issue. Kan spoke with Oliver and another friend about how to overcome that problem and avoid the “bad rap Gnutella was getting because it was associated with piracy and porn and all these kind of muckraking things,” Kan said. During two weeks in June 2000, the three came up with a prototype product and called it Infrasearch.

Based on the Gnutella framework, Infrasearch was a search engine and sharing service that allowed all types of files to be swapped. In a demonstration, Kan’s team linked together a calculator, a database of pictures, a collection of news stories, and a website that tracked stock prices. Typing in “cat” as a search term returned a picture of a cat, a price

📸