http://www.vanityfair.com: Even in Silicon Valley, the home of the ever-present origin story, Chamath Palihapitiya has a pretty incredible narrative. The Sri Lankan immigrant escaped a civil war in his homeland before eventually landing a job at a new company called Facebook. Then he started a venture-capital firm that is now among the most formidable in the Valley. He is also an owner of the Golden State Warriors, the defending NBA champions, who are on pace to achieve the greatest season in league history.

Palihapitiya’s firm, Social Capital, has backed numerous tech companies with valuations in the billions, such as Slack, Box, and SurveyMonkey. But that doesn’t mean that he is bullish on unicorn culture. Here, Palihapitiya speaks about Mark Zuckerberg’s secret sauce, which start-ups are going to make it, and the saga between Apple and the FBI, among other topics. Following are excerpts of an interview with Chamath Palihapitiya:

Q: Vanity Fair: You were at Facebook early on in the company’s history. What’s the best advice you learned from Mark Zuckerberg?

A: The most amazing thing about working with him is how even-keeled he is. He’s very happy to just be the last person to talk in a room. So he’ll just sit there and you’ll tell he’s actually actively listening and thinking about everything people are saying. He had this amazing ability to unpack what was anecdote and what was fact.

Back then there was a very turbulent phase in our company. We had a lot of ups and downs. We had a lot of scrutiny. One day we’d be getting sued by the FTC, the other day we were celebrated for being the next great company. In all of that tumult, it was really important to stay as calm as possible and think as far out as possible. I think that’s what allowed him to ultimately turn down the acquisition offer from Yahoo. It’s what allowed us to negotiate really important financing from Microsoft. It’s what allowed us to invest and take a bunch of shots on some random things, like our online advertising, that has turned out to be a huge business.

Q: Funding is slowing down, both in seed rounds and mega-rounds. There have been fewer tech IPOs recently, more companies are raising down rounds. Are we in a downturn?

A: I think we’re in a phase where we’re realising that the people who have been allocating capital thus far have done a horrendous job. Most people’s inherent reaction is to make sure they never lose their job, and so they become risk-averse. I think what we’ve had is a handful of investors who have extreme vision who make great investments in things that are amazing businesses: Facebook, Google, Uber. And then everybody else reacts to that success by trying to do the thing that most approximates the thing that’s working. 

As a result, most of those businesses are fundamentally not good, they’re poorly run, and they never should have been invested in in the first place. But the capital came in because the person who had control of the capital was able to justify it intellectually to themselves versus something else that could have become the next Facebook or Google.

The reality is great companies can go public in any market. When we talk about the IPO slowdowns what we’re really saying is that there really just aren’t that many good companies being built. We need to divorce ourselves from venture capital as an occupation and focus on using capital as a way to take really big bets on things that just seem totally audacious. Right now we haven’t done enough of that, and the result is that most of the things we’ve funded are mostly crap and largely worthless.

Q: What advice are you giving Social Capital’s portfolio companies in the event of a tech bubble burst or correction?

A: There’s a tension between investing in visionary companies versus investing in beauty pageants, and we have always erred on the side of the former. When nobody else is funding cancer research, we were funding cancer research. When nobody else was funding diabetes, we funded diabetes. When nobody else was touching asthma, we were spending enormous amounts of money on asthma – all of these things that frankly take lots of time to build. Around the early part of 2015 we came together as a partnership and said that the most important thing that we can do is get every single one of our companies well financed, and make sure they really understand that they need to make this money last for two to three years.

We’re trying to coach our CEOs that the window dressing is both expensive from a cash perspective and tremendously expensive from a culture perspective. It distracts the team from building what they need to build. Don’t waste money on things that get away from your mission, which confuse employees about why they’re actually there. Meaning, the quality of the office and the quality of the food are all part and parcel of a lack of discipline, which speaks to the fact that the mission isn’t compelling enough. Because I can tell you what it was like at early Facebook: the food was terrible; we’d ship in lunch and probably two to three times a week the lunch had maggots in it. But we were there because we believed, and it didn’t matter.

Q: What are the characteristics of the companies you see surviving a downturn?

A: If you look back, historically, companies like Microsoft or Apple have taken 25 to 35 years to get to the kinds of valuations that Facebook got to in 10 years. And the reaction to this in Silicon Valley is to try to find things that work really quickly. Over the last seven or eight years, Silicon Valley has really fallen in love with fast growth. Part of this is the risk aversion of the financiers, who themselves want to have short-term wins and want to fund things that look like they’re working in the short term.

But the problem with things that work really quickly is that they can stop working equally as quickly. Valuable companies take decades to build. We try to find businesses that are technologically ambitious, that are difficult, that will require tremendous intellectual horsepower, but can basically solve these huge human needs in ways that advance humanity forward. Those things don’t necessarily take lots of money, but they generally do take lots of time. And they require really mission-driven people. But when you build those companies, my gosh, they are so enormously valuable.

The businesses that, in my opinion, are worthless are many of these negative-gross-margin businesses that grow really quickly because the growth doesn’t indicate any long-term business value.

Q: What kinds of companies are you talking about, specifically?

A: Most companies in e-commerce right now are negative-gross-margin businesses. Amazon has such enormous scale, and they’re at about 13 or 14% gross margins, but on a huge number. In order to compete with Amazon, these businesses have to sell goods for less than what they cost. These companies are in the delivery businesses (Postmates, DoorDash, Instacart) and in the food business (SpoonRocket, Munchery). Basically, a lot of these new-generation, remote-control-type businesses – where the phone acts like a remote control to replace an offline experience – are generally, to date, highly, highly, highly unprofitable.

There’s a lot of what I call ‘venture philanthropy’ to prop these businesses up. Time will tell whether any of those can become a real business. If a shoe costs $ 20, Nike doesn’t sell it for $ 14. They sell it for $ 400. We have to get back to this world of having pretty reasonable discipline on business models and understanding that many of these gross-margin businesses will never, never break even or become profitable.

Q: A number of VCs have been calling on mature, late-stage companies to go public. There’s even been somewhat of a quiet rally in the public tech stocks recently. Is now the time for big, late-stage companies to go public, or does it make sense for companies to stay private longer?

A: Any company that is making its decision based on external timing is probably not in control of their own destiny and should probably not go public. Facebook could have gone public whenever it wanted. We decided the right time was 2012. It could have easily been 2010 or 2014. When you hear the call for these companies to go public and there’s pushback and they don’t, what’s really happening is the realisation that the structural strength of their business is not yet in place. So they’re worried about how the public market will react once they have to transparently demonstrate what their business will look like. The great companies can always go public whenever they want; every other company is trying for some window of time where there’s essentially some combination of intellectual laziness and greed in the public markets that will allow them to exploit a window.

Q: You’ve voiced your support for immigration reform. Can you talk about why that’s important to you and why it’s important for the country?

A: I’m a living testament to the value of immigration. I escaped a civil war and I came to Canada as a refugee, and they gave my family protection. I did my best to pay that country back and I think I did that. I emigrated as a knowledge worker to the US. The US gave me a shot and I think I’ve repaid the US in enormous ways.

I’m also fortunate enough to be a co-owner of a basketball team. What you see is that when you try to recruit the best people, you can win. In the case of the Warriors, they recruited based on skill and culture. It’s another way of reinforcing to me that America is this shining beacon where the best of the best want to be there.

Q: A number of tech companies have aligned themselves with Apple in its legal entanglement with law enforcement over the San Bernardino shooter’s iPhone. A few others, including President Obama and Bill Gates, have taken a middle-of-the-road approach. Where do you stand?

A: I’m more on the side of the government on this one. I think we’re almost lying to ourselves in pretending that Apple can’t do something that they do every day. Instead, what I think has happened is we’ve started to fall on our sword on some weird marketing thing. They want to seem like they’re on the side of the consumer. But there’s nothing stopping an Apple engineer today from rooting your phone and taking whatever they want. It’s a bit of a red-herring conversation right now.

A would-be Wi-Fi paradise

Bloomberg: Sri Lanka has enjoyed an era of strong economic growth since its bloody, 26-year civil war ended in 2009. To keep it going, the Government is trying to make the island nation a technology hub. It’s investing in new undersea internet cables, putting money behind startups, and working with Microsoft to embrace cloud computing. It’s also been wooing Google and Facebook to host tests for some of their most ambitious experiments, from self-driving cars to drones. First up: the balloons.

Google’s Project Loon is an effort to develop high-altitude balloons that can bring internet connectivity to remote areas. The technology has been tested over the past couple of years, but not at scale. Rama, a quasi-public company controlled by venture capitalist Chamath Palihapitiya and the Sri Lankan Government, aims to do just that. Google sent the first Loon balloon above Sri Lanka in February, and the Government says it’s working with the company to blanket the country with coverage from another dozen.

Within a year, Palihapitiya says, Loon balloons will turn the Indian Ocean country, which is about the size of West Virginia, into one big Wi-Fi zone, giving Google the first real sense of whether Loon can be commercially viable. “This is really a profound thing the government has sponsored and is really pushing,” says Palihapitiya, an early Facebook employee who runs Silicon Valley venture firm Social Capital. “If we can do this in Sri Lanka, that sets the tone for the rest of the world.”

Under a deal Palihapitiya negotiated, Google is leading the Loon launches, development, and maintenance, and Rama will run the software to control access to the balloons’ Internet connections and handle billing. People will still pay local operators for data, and carriers will pay Rama an as-yet-undisclosed fee to help ferry traffic.

The Loon fleet could offer a cheaper alternative to undersea Internet cables, which pass through the regional choke points of Singapore and Hong Kong. Carriers in developing Asian and South American nations pay more than 10 times the bandwidth prices their European and US counterparts do, researcher TeleGeography estimates. (The median wholesale price for a monthly 10-gigabit-per-second connection in Los Angeles or Frankfurt is $ 1; in Mumbai, it’s $ 15.) Sri Lanka’s data use is growing 45 percent a year and will likely do so for the next decade, the United Nations estimates.

Increasing the country’s bandwidth is crucial to achieving President Maithripala Sirisena’s goal of propelling the nation into the ranks of developed economies through technology. The average Internet speed available to the country’s broadband customers is about 5.1 megabits per second, Akamai Technologies estimates – better than in India (2.5 Mbps) or China (3.7 Mbps), but a fraction of the average speed in the US (11.7 Mbps), Singapore (12.5 Mbps), or Hong Kong (15.8 Mbps). Rama “is bringing universal connectivity to the whole country,” says Muhunthan Canagey, the head of Sri Lanka’s Information and Communication Technology Agency.

The Rama project began a little more than a year ago, when the president met with Palihapitiya. “I said to them that if they created a mandate for Internet access and making it a right, they would be the first country to do it,” recalls the 39-year-old venture capitalist, who emigrated with his family from Sri Lanka to Canada in 1983, at the height of the civil war. “I had very close relationships with the team at Google, and so we were able to get that team interested in providing Loon technology to the country.”

If the Wi-Fi-beaming model works, Rama could also partner with Facebook’s Wi-Fi drone program or with satellite providers, Palihapitiya says. “If we can solve this in Sri Lanka, then we can solve it in the Philippines, Vietnam.”

Following the first Loon launch, local news outlets flashed photos of what looked like the wreckage of a balloon near the capital, Colombo. A spokeswoman for Google parent company Alphabet said that was a planned landing following a successful test. Google has disclosed few details about Loon, except to say it’s working with carriers around the world. (It’s also testing the balloons in Indonesia and working to bring them to India.)

No matter how many balloons are in the air, Sri Lanka will still likely need fibre-optic connections to handle growing internet demand, says Rohan Samarajiva, chair of technology infrastructure think tank Lirne Asia and a former director of Sri Lanka’s telecom regulator. Loon is “a rather unusual solution and has got potential,” he says, but it’s “an unproven technology we should look at with reasonable and realistic expectations.”

Canagey, the government official, is less equivocal. “This is the way the world is going,” he says. “We have to evolve in the way the technology is evolving.”

The bottom line: Rama’s partnership with Google could give Sri Lanka another way to boost its ever more precious Internet bandwidth.