Unchained - Not Sure a Cryptoeconomic System Will Work? Gauntlet Can Help - Ep.124

Starting point is 00:00:02 Hi everyone. Welcome to Unchained, your no-hype resource for all things crypto. I'm your host, Laura Shin. You may have heard Unchained is doing a survey. We want to know. How do you think we can make the show better? How would you like to see Unchained Expand? If you could just take a moment and go to SurveyMonkey.com slash R-changed Survey 2019. Again, that's surveymonkey.com slash R-chained Survey 2019. Your answers will be a huge. huge help to me and my team here at Unchained. Also, those who answer the survey can enter to win one of five free Kasa Bitcoin Lightning nodes, plus a free year of Kasa's gold membership, including a multi-six security app for iPhone and Android, a Treasurer hardware wallet, a Kasa Faraday bag, and 24-7 support. Those of you interested in learning more about Kasa or about protecting your Bitcoin investment generally should check on my interview with CEO Jeremy Welch. Thank you to Kasa for donating. the URL is surveymonkey.com slash R slash unchained survey 2019. Go there now to give us your thoughts in the future

Starting point is 00:01:10 direction of unchained and enter the giveaway. CipherTrace makes it easy for exchanges and crypto businesses to comply with cryptocurrency anti-money laundering laws. Avoid illegal sources of funds and maintain healthy banking relationships. CipherTrace is helping you grow the crypto economy by keeping it safe and secure. Cracken is the best exchange in the world for buying and selling digital assets. It has the tightest security, deep liquidity, and a great fee structure with no minimum or hidden fees. Whether you're looking for a simple fiat on-ramp or leveraged options trading, Cracken is the

Starting point is 00:01:50 place for you. My guest today is Tarun Chitra, co-founder and CEO of Gauntlet. Welcome to Roon. Hi. So tell us what. Gauntlet does? Yeah, so Gauntlet is a blockchain simulation platform. So we basically try to model all the different types of users who interact with both layer one chains and consensus protocols as all smart contracts. And one of the reasons for really doing this is to have

Starting point is 00:02:19 a strong statistical understanding in addition to a security understanding of how these systems perform when different types of users are interacting with these systems. A simple kind of example is, you know, in a lot of crypto protocols, people tend to compare Byzantine and really kind of irrational actors versus honest actors who perfectly follow a protocol. But in reality, most users of these protocols are traders or interacting with other systems and modeling how their incentives work and how their rationality works gives you a better idea of how the system will behave under sort of more realistic conditions. Yeah, I remember when you first told me about this or reminded me of Monte Carlo simulations that sometimes retirement

Starting point is 00:03:14 professionals will do for you or there's software that does this to kind of model out with different inputs, whether or not you'll have enough money to. the last retirement and, you know, et cetera. So it will kind of show the different scenarios under different marking conditions. But you have a really interesting backstory to how you came into the space. Let's just start with the first time you ever heard about Bitcoin. How did that happen? Yeah, definitely.

Starting point is 00:03:43 So I was working at this kind of odd research institute called D.E. Shaw Research, So it was a private research institute. And this billionaire who was formerly a CS professor was actually spending his fortune on building A6, which are applicated specific integrated circuits, which are the same chips you use in minors or in GPUs. We were building A6 for doing physics research. And at that time in 2011, there weren't really many people building A6. There were telecom companies who built them for routers. There were research groups who built them for new chip architectures,

Starting point is 00:04:27 but most of those didn't make into production. And then there was Apple, Samsung. And so the interesting thing about these types of orders is that if you don't have an order that's of a large enough size, fabs, which are kind of these fabrication facilities, mainly in Asia right now, they won't really talk to you. So you have less than $100 million of chips that you want to get produced. They won't really speak to you, and you have to go to these aggregators who take many small orders. And they do a lot of the technical work to make sure that your chips don't interact with someone else's chips and the formal verification.

Starting point is 00:05:08 And the behavior of these chips is as it would be if you have the whole order yourself. And so we went to an aggregator. we had this roughly $25 million chip order, and they said, okay, great, we'll be back in a few months with kind of the first samples. And then they more or less ghosted us for a while, and we were like, hey, what's up? You took $25 million from us.

Starting point is 00:05:36 And they were like, well, you know, actually, how about a 10% discount? And we were like, you can't just give us a 10% discount without telling us why you just didn't talk to us for three months after taking our money. And that was in 2011. And so that was really one of the first Bitcoin ASIC miners was coming online. And that was when I was like, wow, I really should be trying to understand more about this. I thought it was, you know, not quite a joke, but I thought the paper was a little bit

Starting point is 00:06:05 grandiose, you know, for a distributed system style paper. So I didn't really truly believe it until I started seeing people building hardware for it. And that sort of To be clear, so the FAB basically got a big order from somebody who wanted to create a Bitcoin ASIC, and so that's why they postponed your order and like pretended

Starting point is 00:06:27 like they hadn't taken your money. Is that what you're saying? Yes. I'm sure the Bitcoin ASIC order paid them some, you know, increased cost over what we paid for the same amount of space. But yeah, we sort of got front run

Starting point is 00:06:43 by a Bitcoin ASIC manufacturer. sure. And so then how did you go from, you know, at that point, just learning about Bitcoin to eventually here. I mean, so now that was 2011, why don't you fill us in on the last eight years and how you came to launch a company in this space? Yeah, definitely. So basically, I'd mined quite a sizable amount of Bitcoin because at that time, even then, you were still reasonably profitable with GPU mining. And I just, my parents lived in a place that had relatively cheap electricity. And I just kind of had a computer running in their basement. And I had a lot of Bitcoin. I saw the crash of 2013. And I really was scared. And I just sold all the Bitcoin I had.

Starting point is 00:07:33 And I was like, I'm never getting back into this. It's stressful. I have to worry about, you know, security and all this stuff. But I kept paying attention to the academic literature. And I'd really been taken by a couple key papers that had come out in 2013 and 2014 and 15. And they were the ghost paper. So ghost stands for greatest, greedy, heaviest observed subtree. This algorithm is actually one of the key components of the first version of Ethereum. It's what kind of let Ethereum have a faster block production rate. but it was the first academic paper that really studied the probabilistic aspects of how blocks travel through the network,

Starting point is 00:08:19 as well as how different types of adversaries would potentially try to interfere with the growth of the chain in a very formal way, a much more formal way than the original Satoshi paper, which over time has been found to have a lot of, at least on the mathematical side, has had a bunch of mistakes. And the second paper was a selfish mining paper by Amundsir and Itaiel. And that was also another paper where people really used more rigorous probabilistic tools to try to find a bug. And that in selfish mining, what happens is a miner kind of holds out a bunch of blocks that they produce. And then as long as they have an advantage over the rest of the chain, they keep growing on their private chain. and if the rest of the network comes close to them,

Starting point is 00:09:09 they publish all of their blocks. And the idea with this attack is it kind of reduces the efficiency of the network and gives all the rewards to kind of the selfish, a larger percentage of the rewards than they are due to the selfish miner. I tried getting people I worked with excited about this, but I think there were more traditional distributed systems and hardware people, and they just really were like, well, this is just a novelty, and this is crazy people making A6 in Taiwan.

Starting point is 00:09:39 You know, can this thing be real? And then, you know, after I worked there for five years, I end up working in high-frequency trading. And that was, you know, that was kind of when I saw the invention of a lot more sophistication, a lot of financial sophistication in the space. I think the algorithm really made me, for the first time, realized that there was sort of a melding of tech and finance in a way that hadn't existed, because people were finding novel ways to make structured products that,

Starting point is 00:10:16 you know, securitize the security of a network in terms of proof of stake. And I'd kind of, you know, in trading, when you make a trading strategy, you basically design your strategy, you take kind of the statistical features of it that you think are important, that represent what other people are. during in the market or what you want your ideal strategy to be doing. And then you kind of run these back tests, which are like the Monte Carlo tests you discuss with regards to your retirement account, where you basically can run the strategy on historical data and say, okay, I expect to make X amount of dollars. The standard deviation of the return is this. This is the worst case loss. This is the

Starting point is 00:11:01 best case return. And you can analyze how this kind of algorithmic strategy that might just say, 0.2% of the time send an order here, 50% of the time, send an order here, something that's very not super human intuitive. And you run these simulations and you can get the economic intuition for what it's doing. This is how it makes money. This is how it loses money. This is how statistically broad the distribution of outcomes is under different stress test scenarios. And that was when I started kind of seeing a lot of similarities to what people were doing in crypto, in that cryptographers have a method of proof where they basically make an idealized simulation. So they say, hey, let's pretend there is an Oracle, so someone who has full control of the whole network.

Starting point is 00:11:55 And the Oracle picks an adversary every once in a while and lets the adversary read someone else's state. So it means they hack their computer and read their private keys and they can sign signatures as them. Or they hack their computer and don't forward blocks. And basically, the cryptography version of simulation, most of the proofs say, okay, let's pretend we have this Oracle. The Oracle can call the adversary. It can also call honest users. And then it kind of interlaces actions between them. And cryptographic proofs like the ones in Algarand prove that with a certain probability the adversary succeeds and that probability is really low under certain choices of parameters, like how big your block is, your block production rate, assumption of how delayed

Starting point is 00:12:42 the network is. And that was when I started seeing that this kind of gap between modern finance where people really assume a lot more about the user. You assume that they're rational and they have very different ways of measuring what rational means to them, versus, is kind of the cryptographic proof version where it's everyone is trying to destroy everything or everyone is perfectly good all the time. And in reality, everything is somewhere in the middle. And picking these kind of parameters in your protocol becomes increasingly important in proof of stake because now instead of collateralizing kind of these assets with proof that you spent energy or proof that you expended a certain resource like space, you do a

Starting point is 00:13:31 it with this proof of I locked up this digital asset.

Unchained - Not Sure a Cryptoeconomic System Will Work? Gauntlet Can Help - Ep.124

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.