Because of the low stakes, these companies are not likely to draw Google-size payouts. A more realistic model of success is Buzztracker, a news-aggregation site sold to Yahoo in September for $5 million. "The outcomes are lower," Siminoff says. "The internet has gone from minting a few billionaires and a few hundred multimillionaires" to creating 10,000 success stories at $10 million apiece.

The new low-risk, low-payout strategy has another side effect: Failures need not be the cataclysmic events they once were. If Kim's company comes crashing down, he won't deem it a major setback. He has sunk a few months into the project but only about $10,000 of his own money. Serial entrepreneurs have long been a fixture in San Francisco and Silicon Valley, and in the wake of failures, they used to spend years between startups. In today's environment, Kim could be working on another idea within weeks.

In fact, he has failed before—several times in the past three years. He grew up in California but attended high school in Shenzhen, an industrial region in China, and it was there that he created his first company, a website for online gamers. His father, who now resides in China, found him some office space. There was nothing lean about Kim's approach back then. Cheap Chinese wages meant that he could scale up quickly. Before he knew it, he had several dozen employees and multiple projects under way. Big mistake. "Our burn rate was too high, and it became a management nightmare," Kim recalls.

After high school, he was accepted at the University of California at Berkeley. Even before classes began, he was thinking of new startups. He tried Stalkerati, a search engine for social networks; then, Hawtppl, a version of the popular website Hot or Not. Early in the fall semester, he started Yaqqer, a mobile instant messenger. (None of these took off.)

This past April, halfway through the spring semester of his freshman year, Kim dropped out to devote himself full-time to starting WeGame. He cleared his meager belongings from his dorm room and set up a live-work loft in San Francisco. It's not the most attractive location, squeezed against a freeway exit ramp, but it offers a ton of bandwidth. The building sits within a three-by-five-block area served by fixed-point wireless service that blasts data at 45 megabits per second, about 10 times the typical broadband speed. WeGame's corporate headquarters is a single desk topped with two 24-inch L.C.D. monitors, where Kim spends his days writing code.

The main draw for the website he's building will be a software "camera" that can record short segments of videogames. Those highlights can then be posted on the Web, YouTube-style, for other gamers to see. Similar software already exists, but it costs about $40 and compiles uncompressed video that takes up an enormous amount of space on a user's hard drive. Kim's version, which can be downloaded free from WeGame, will compress recordings with, he claims, no significant loss of quality.

The idea faces challenges. Most significant, Kim's recorder will work only with PC-based games, not with popular consoles like Xbox, PlayStation, or Wii. Of the $33 billion worth of videogames sold last year, only 22 percent are PC-based. Kim says his plan is to gain traction with PC gamers and then negotiate deals with console makers. And he insists that the market for PC games is still significant enough for him to be able to use it to bootstrap himself into the big leagues. "All kinds of virtual worlds are strictly PC at this point and will be for the next half-decade," he says, referring to role-playing games like World of Warcraft, EverQuest, and Club Penguin.

Another concern for Kim's website is that gamers might simply continue to post their highlights on YouTube. You can find clips from Halo or Doom on the video-sharing site today, and some are big hits. A World of Warcraft how-to guide has been viewed more than 6 million times. Kim says his version will be better. "If they post a highlight on YouTube, it will be lost in the millions of videos, most of which are nongaming," he says. "YouTube doesn't allow them to categorize their content. They can only post into its Gadgets and Games section, which is very generic." (YouTube includes a search function, but it's limited.) There's a technical drawback as well: YouTube videos are converted to rates of about 300 kilobits per second. Kim says that WeGame files will convert much more information at the same speed, producing clearer, richer images.