The ruthless Rose

April 12, 2011, 9:00 AM UTC

FORTUNE — A month ago, Kevin Rose, the founder and former CEO of Digg, built something creepy. He had been nursing an idea that would test the limits not just of how much we wanted to share online, but also how much we wanted to know what others were sharing. So Rose and a friend coded a quick program called Peeka. It featured a “bomb button” that would do more than just share an article with a friend. It would take over her browser window, forcing her to read whatever you bombed. Rose invited a handful of friends to try Peeka, and the reaction from all, Rose included, came clear and came quick. “It’s not a good product. It doesn’t feel right. It doesn’t feel fun. It feels scary,” Rose says. It was an idea that simply wasn’t working. So, after a few weeks of development, Rose shot it in the head.

“Shoot it in the head” is Rose’s favorite way to murder an idea. Last week, when Techcrunch’s Sarah Lacy first reported the details of Rose’s next startup, Milk Inc., he told her that at Milk, “if something isn’t working after four months, we’ll just shoot it in the head and start again.” A week later, on the phone with Fortune as his dog whined in the background, Rose picked up the revolver again. “We just wanted the flexibility going into this to be able to shoot something in the head if it’s not working.”

The similar verbiage isn’t surprising, and not only because Rose has just finished making the rounds to venture capitalists to raise money for the startup. (The amount of which he says he’ll announce soon.) It also suggests that Rose, a veteran and one-time mascot of Web 2.0, thinks startups have become timid in the past few years. Innovation calcifies as companies attempt to salvage well-funded ideas that aren’t working. The solution, according to Rose, is to not be afraid to acknowledge failure, cut ties, and start something new.

Not coincidentally, Rose is in the midst of doing the same for his own career. After founding Digg in 2004, Rose spent years building the site into a cornerstone of the meme-junkie’s Web, nearly sold the site to Google (GOOG), and tried to counteract the way social networks were eroding Digg’s audience. Those efforts didn’t work. In October the company laid off 37 % of its employees. Earlier this year, Rose left the company. It was a tacit admission that there was only so much reinvention that could be done.

But Rose’s willingness to reinvent has not dimmed. Fulfilling his duty as a dotcom entrepreneur, he says his new company is “trying to build something big and that changes the world.” And when he starts describing his ideas, the words cascade, kinetic in their movement from one detail to the next. After he described Peeka, he took a moment to reflect on the slapdash process with reverence. “It’s awesome.”

Rose wants Milk to have a similar flavor of awesome. For now, Rose is only talking about the kind of place it will be, not any actual plans of what it’s going to build. He says it will be a “little lab company” that develops mobile apps and the websites that accompany them. But, he says, these won’t be the kind of mobile apps we’re used to.

“I’m not just going to add badges for the sake of adding badges,” Rose says about the foursquare-led craze of rewarding people for using an app. But he acknowledges that on some level these kinds of gimmicks — which he refers to as “candy” — are successful. “A badge is pretty much useless other than just a few pixels on my little portfolio. So what can we do that’s different there?” Maybe actual, tangible rewards, similar to a frequent flier or credit card program? He pauses at the thought, says, “I’m very excited to show you some of this stuff in a few months,” and chuckles.

Most significantly, though, Milk won’t be a company that’s afraid to look at an unpopular product and ask, “Is this something we want to continue to iterate on top of, or should we just can it altogether?”

Pivot transformation

What Rose is trying to build, then, is something that’s immune to something we’ll call the “cult of the pivot.” In techspeak, “pivot” is really a euphemism for “Our idea failed, but we have way too much venture and emotional capital invested in this thing to bail in a heap of embarrassment, so instead let’s totally reinvent what we thought would be our life’s work.” Every now and then the pivot works. Groupon started as a site that was far more charitable than the consumerist haven it is now. Instagram was once a location-sharing app. And even Twitter used to be a service far more rooted to text messaging than its current iteration.

But those are the success stories. The ones we don’t hear about are the thousands that pivoted even further into oblivion. Rose, who has both pivoted (with Digg) and bailed (with his old startup, Pownce, which he and a cofounder sold to SixApart in 2008), prefers the latter. “There are a lot of companies that will sit there with a bad idea and burn through millions of dollars trying to make it work. And we’re not fooling anyone, we’re going to look at it and if it’s not working we’ll kill it.”

There’s that murderous language again. But he says he won’t pull the plug unless he has to. “It’s not like I commonly come into the office and frame it as, ‘Hey what are we going to kill today?’ If we’re having a conversation a year from now and we say, ‘Well, what did you end up killing?’ And I say, ‘Actually, nothing,’ I’m going to be stoked.” So maybe it’s best to think of Rose as a Kevorkian type. He’s one of the few willing to acknowledge that some ideas don’t need to stick around forever. The trick, as ever, is knowing which ones are which.

–This is the second installment of Fortune’s new column on startups and the tech bubble that may or may not exist. Email me your ideas, rants, and love letters, or just lurk on my Twitter feed.

More from Fortune: