A Truly “Fireside” Chat with Steve Jurvetson

nCent Labs Announces $10MM Seed Round at hack.summit() 2018

nCent has announced a $10 million seed round backed by an illustrious group of investors including Naval Ravikant, SV Angel, Sequoia, Winklevoss Capital, Zhen Fund, Floodgate, and MetaStable among others. At Hack.Summit() on July 10, 2017, we launched nCent labs out of stealth mode in a fireside chat with tech visionary, mega-investor and founder of Future Ventures, Steve Jurvetson!

Of course Steve happened to have a working space rocket engine in his living room(!), so we decided to make it a “true” fireside chat. We discussed the inspiration behind nCent, the vision for its growth, and how our community can realize the unrealized potential of the internet by harnessing incentive markets that run on a decentralized blockchain. I’d like to share with you some highlights of our chat.

(NOTE: This interview has been edited for time and space — the full version can be found online here)

Steve Jurvetson (SJ): How would you describe nCent?

KK: We like to describe nCent as a decentralized protocol for incentive markets. Incentive markets are an exciting category of decentralized applications that folks have not focused on. Simply put, an incentive market is an efficient way for people to collaborate and have the value of their collaboration be rewarded. This creates the substrates of incentives that allows markets to form.

The two best ways to organize people are firms and markets. In many cases, firms are all we have, and they can be inefficient. Sometimes the incentives inside a firm can be broken or misaligned, and sometimes a firm’s incentives are misaligned with their users’.

We are trying to create a decentralized platform that allows market forces to enter into the organizational structures that typically are the domain of firms.

There are a lot of ways to break apart the elements of how incentives to collaborate form. We have a laser-like focus on incentives. (Thus the name: “nCent.”)

“We are trying to create a decentralized platform that allows market forces to enter into the organizational structures that typically are the domain of firms.” -KK Jain

SJ: As I listen, it makes me think about not only incentives and motivations, but also economic systems, how markets form, and what makes for liquidity and stability in markets and a thriving economy. We first met at an AI in financial technologies conference that KK helped organize, which was fascinating. I’m curious if you can give us some background on your professional path and what inspired you to transition into starting nCent?

KK: My formative years were spent during the tech revolution. I had dreams of the promise of personal computing and spreading information. I was a Mac geek. That was my first exposure to the development. In fact, one reason I went to Dartmouth was because it was a mac-ified campus. (And they gave me the best financial aid.)

I had some big meetings, including one with Steve Jobs, and eventually became an intern at Microsoft. I started daytrading internet stocks and ended up on a circuitous path to Wall Street. Then I was recruited by DE Shaw, which is a quant hedge fund. I like to say that Wall Street picked me — I didn’t pick Wall Street.

At DE Shaw I learned a lot about statistical arbitrage, what makes markets work, how to model markets, and how to develop safe and secure code at scale to conduct financial transactions. After some time I continued my education by enrolling at Stanford, and then I eventually went back to Wall Street to run a multi-billion-dollar quant fund for Perry Capital and Citgroup’s Principal Strategies hedge fund group.

After a decade on Wall Street, I did some self-reflection. I came up with this analogy in my mind of the “Mouse Olympics.” When you’re successful, you get a better bonus, bigger team, and more capital to manage — these are the gold medals. But even if you win every gold medal at the Mouse Olympics, and at the end of the day, you’re still a fucking mouse.

I wanted to give back by teaching, and soon after talking to my advisor at Stanford I had a faculty appointment in the engineering school running the computational finance program. It gave me a frontrow seat to similar innovative fervor to what I felt in the first dot-com boom.

“On Wall Street all your wins are private, and you monetize your brain and the gifts you’ve been given. I missed the idea that you could contribute to something that was bigger than yourself.” -KK Jain

Being at Stanford in that capacity also gave me the freedom to run a blockchain research group, which was a great opportunity to develop my ideas and my thinking about blockchain.

SJ: Did you tap into that initial pool of talent?

KK: Yes, I think things in the blockchain community are changing super quickly. A week in blockchain feels like a few months in other sectors, and there’s a lot of really good research and creativity at Stanford that we’re certainly tapping into. We’re blessed to have wonderful advisors and contributors, like the chief scientist of Stellar, David Mazieres and others. We’re blessed to be plugged into that cutting-edge research community.

SJ: I noticed you formed a telegram group that went from zero to 9000 in a week, even before you said much about what you were doing. How did you pull that off?

KK: We created it and just asked people that we knew. We didn’t even email all our investors. It grew organically. We wanted to refresh our logo, and we thought it’d be fun to crowdsource it, and one thing led to another. One of our secret weapons is our global community manager — Porter “Popcorn” Cornelius.

Porter Cornelius (aka “Popcorn”) Global Head of Community Management and Cuddles

SJ: Crypto-puppy! Eventually he’s going to be a big ass dog?

KK: Yes, a big crypto-dog! He’s a little bit clickbait-y, but it’s our homage to our colleague Sebastian Thrun, who famously went on TechCrunch Disrupt with his puppy.

SJ: Maybe we can talk a little bit about the timeline and development. Prior to investing, you told me about the DARPA red balloon challenge, and I hadn’t heard of it before you introduced it to me. How did the DARPA red balloon challenge influence your work at nCent?

KK: The story begins in 2009 with the Defense Advanced Research Project Agency (DARPA), which researches challenges of all kinds. They sponsored a challenge that offered $40K as a reward for anyone who found ten moored red balloons, each placed somewhere unknown to the public in the U.S. They didn’t tell people anything about where the balloons were located. The challenge was before quasi-real time satellite images or computer vision. This was a “needle in a haystack” problem. Some people argued that it was an impossible task and that $40K wasn’t enough of an incentive for the resources needed.

SJ: I read that there were a few hundred people that really made a serious effort but that MIT ultimately won.

KK: Yes, MIT won!

SJ: How did the MIT team approach the challenge?

KK: I was going to help, but I was working during the kickoff. Basically, MIT hacked incentives and used online collaboration. They gave a thoughtful incentive structure for people to share information about the system. $40K was the total reward, but they divided it into ten buckets for each of the ten balloons. Each balloon was worth $4K. Their pitch was “Come one and all to join our red balloon hunting team!” If you found the balloon your reward was $2K, but there was an incentive program to propagate information about the campaign. If you told your friend about the campaign and they found a balloon, then you would win $2000. If you recruited someone who recruited the balloon finder, then you would get $1000. This is a recursive structure that converges, so the total cost for each balloon will not be over $4k. MIT was a trusted central actor sponsoring the contest.

So, you got a reward if you found the solution, and you got a reward if the solution was in your subtree, which could be infinitely deep in theory. Everyone found the balloon that was in Chicago, where it was easiest to find, but the postmortem showed that this approach outperformed searchwise where there were deep tails with multiple hops to find the balloons. There were a lot of interesting ideas that came out of it.

SJ: So it turns out that 43% of people believe that the problem could be solved in less than a week. Approximately 30% answered “less than one day,” and another 30% answered less than a year. But how long did it actually take?

KK: It took less than nine hours to find all of the balloons. There’s a lot of subtly and complexity here. One of the dreams of the internet, though, is that it can do more than just connect us. The red balloon challenge inspired me to use the internet in a way that can unlock our common potential. I think the most promising way to do this now is to explore incentive programs through blockchain technology — that’s why we started nCent.

SJ: I really came increasingly to believe that a large part of the value of these infrastructure layers, whether it’s blockchain or other protocols people are trying to promulgate, are largely directory and search: How do I find what I’m looking for and how do I get paid? And when you’re developing nested code, you would ideally want to be able to pay an infinitely deep tree.

It fascinated me when Hotmail and Skype came around. We invested in Skype and my partner Tim and I coined the term “viral marketing.” I’ve been fascinated by the power of what’s endemic to the internet itself — what Kevin Kelly would say is “the inevitable trajectory of this technology towards distributed systems, towards nested and recursive incentive systems, and unlocking their power and potential.” You’ll see that in some areas where there’s no payment, it’s just reputational capital. But if you could layer on an infrastructure layer to address that in some sense, you could unlock the main untapped potential of the internet, at least in my opinion.

My partner Tim and I coined the term “viral marketing.” I’ve been fascinated by the power of what’s endemic to the internet itself — what Kevin Kelly would say is “the inevitable trajectory of this technology towards distributed systems, towards nested and recursive incentive systems, and unlocking their power and potential.” -Steve Jurvetson

KK: Yeah, for sure. One of the things that we like to say is that if aliens came down from Mars, and they saw Bill Gates and Linus Torvalds, they would recognize that Bill Gates is the richest guy in the world. And they would recognize that Linus is a modest, open-source mega hacker, but the value attribution is a little bit strange, because you could probably argue that Linux was at least as impactful as Windows was as an operating system.

We haven’t actually solved incentives for open-source development, and I think that a lot of people that do open-source development aren’t really doing it necessarily for the payday anyway. The point is that if you can somehow have some incentive market, you might be able to broaden the pie of people for whom it’s worthwhile to contribute to something, like open-source development.

So, open-source is something that one may not consider to be a broken market, but it’s an opportunity to expand that market to be able to touch people that don’t have the luxury of contributing to open-source development for “hacker karma points.”

One of the things that really struck me about the red balloon challenge was its exploiting private information that people knew about each other. I might have a hunch, for example, that my “friend” KK is really into this red balloon stuff. I don’t know why, but he just likes puzzles and games of this nature, so this would be right up his alley. You don’t even necessarily need to even know the model. Your brain doesn’t actually even have to grok the model of why, but there’s this huge treasure trove of data — the information in our brains — that we know about each other. I mean, human beings are social creatures, and we have a hunch of each other’s potential: their capabilities, their skills, their ideals, and so on. All of that is essentially going untapped. Only a small fraction of that even appears online.

Even Facebook, if it had a super targeted ad campaign where it literally just took everyone’s profile, sucked it all in, and put it into some machine learning algorithm — I don’t think it would have been able to solve the red balloon challenge in less than nine hours. It was utilizing information that people implicitly had, and there was this awesome aspect of geographic diversity. You wanted to recruit friends from different regions in America because that’s how to optimally solve that kind of problem.

SJ: And you’ve been using this style of search yourself at nCent, no?

KK: Yeah, as a proof of concept, everyone at nCent except for one person came by our internal recursive incentive referral system. Our goal in general is to empower developers to create their own incentive markets to solve problems that they think are relevant and important, and recruiting is no exception!

SJ: It’s like the ultimate capitalistic distribution of value…Someone on the edge can figure out where on the edge to explore this underutilized potential. Everyone could be making a local optimization decision of where they could find the most untapped potential.

“It’s like the ultimate capitalistic distribution of value…Someone on the edge can figure out where on the edge to explore this underutilized potential.” -Steve Jurvetson

KK: And the hardest crypto involves information that’s encrypted in your brain — you actually can’t even explain the model. We have this dream of the internet. It’s connected us for sure, but it hasn’t unlocked our common potential. The biggest companies that have deep enough pockets and enough eyeballs, users, and data to do this are simply not incentivized to do so.

One of the things that people misunderstand about blockchain is that although there’s a trend towards distributed systems and decentralization it’s also a new way to set up an organization. You can set up the initial conditions of the organization so everyone’s in line and incentivized on tokens, on a tokenized network.

You look at something like Facebook or Google, where you have the shareholders as the benevolent caretaker, so to speak, of the user network, but they’re not quite benevolent, because they’re for-profit shareholders. They have a duty and a fiduciary to maximize profits, which is naturally gonna fall out of alignment with its users. And so, at scale, this alignment’s just gonna naturally fall farther and farther apart. So, you know, I think a lot of these controversies that we’re seeing about Facebook and so forth are the result of an initial condition design flaw of how they’re set up from an incentive structure vis a vis their user network. And Facebook’s an interesting example. For example, Steve and I have a natural social connection. And what is more fundamental for human ownership than owning your social graph, owning your social network? It actually exists in real life. It’s just that we all can’t figure out how to create the data representation of that online, so we defer to Facebook to figure out the protocol. And then in exchange, they’re able to extract all this economic rent off of us.

SK: I can tell you, he’s been ranting about this for a while.

KK: Yeah, so, the upshot of all of this is is that these centralized actors are rapidly losing trust, and they’re not motivated to solve the problem of unlocking our common potential. What you’re left with is broken markets everywhere.

You were just talking about ads. The ad market’s been more or less the same for about 20 years. It’s just been the engine. They’re super spammy and inefficient from an information content standpoint. It’s a super expensive way to acquire customers.

SJ: It looks like we might have time for one more topic. Is there something else you want to address?

KK: One area that we wanted to talk about is on the theme of community building. As part of our culture, we have a teacher-learner model. We spend a lot of time talking about blockchain, and we have a reading circle every Tuesday. We are trying to be humble in our approach and really want to engage the community.

Blockchain is an amazing substrate for incentives. I call myself a bitcoin maximalist — I know that can be controversial — but I would argue bitcoin works because of incentives. My first experience with blockchain was bitcoin, and there is a romance to the design of it. Once you grok the aims of the initial bitcoin proposal and how it worked in practice, it activates an interesting set of neurons in the brain. I’m not wading into the Ethereum — altcoin debate. Blockchain is a beautiful piece of technology, a beautiful system. However, even the strongest maximalist would concede that it’s not a particularly elegant distributed system. Proof of work is like a big dirty hammer: It has probabilistic finality, its not guaranteed to come to consensus, and it’s terrible for the environment. The reality is that the system works because of incentives, and a lot of people miss that fact. Incentives makes everything move, so we think more people should be working on the incentive problem.

SJ: What do you make of high-frequency modeling and trading in the crypto space? I am curious about liquidity in particular.

KK: As an ex-statistical arbitrage trader, this is near to my heart. The more things change, the more things stay the same. In my judgment, there is no alpha or outperformance systematically to be gained investing in stat-arb funds in crypto. That might be controversial, but you hit the nail on the head: Liquidity is one issue. If you look at volumes in the space from a technical standpoint, often times volumes are misleading. As the old expression goes: Volume is what you see, liquidity is what you get. What you really care about is exit liquidity. Volume and liquidity go into one statistical model, which gives you essentially your transaction cost when you liquidate. A better model is: what is the probability of having to liquidate the trade times transaction cost that you would incur on the day you need to liquidate? That is an unbounded problem in crypto. I would also say a lot of the fees are kind of absurd, almost usurious given the bid-offer spread. Custody solutions for institutions are getting better to the extent that this needs to happen. There are many reasons why stat-arb is not there yet in crypto.

“There is no alpha or outperformance systematically to be gained investing in stat-arb funds in crypto.” -KK Jain

A lot has been made of exchanges, and I know there is a decentralized exchange panel, which is topical right now. The exchange market is designed to essentially exploit insider information, which is rampant in this space. I think we need to grow up: either this is going to become an institutionalized market that is credible for capital formation or it’s going to be a den of asymmetric insider information. It can’t be both. Anyone with a vested interest in turning this into a real maket should be cognizant of this issue. Be wary of anyone pitching a start-arb fund in crypto. There’s not enough independent bets you can make. Start-arb requires lots of independent bets, and there aren’t enough liquid independent bets you can make in crypto.

SJ: We’re out of time.

KK: Great. Thank you, and thanks to everyone at nCent nation. To the moon!

SJ: To the moon!

We’d like to offer Steve Jurvetson a huge ‘thank you’ for sharing his time with us at Hack.Summit() and for supporting our work here at nCent. It’s always such a pleasure to work in tandem with great minds like Steve’s.

The work of developing incentive markets at scale is still in its infancy. We need your help, whether you’re a designer, developer, artist or teacher. If you want to see a fairer internet, come join us. Email me directly at [email protected]. To stay in the nCent loop, hear me tweet and join our international telegram channel.