不先查 Pharos，就别碰稳定币｜DeFi Frontier

Then people started to really see that Faros had done a good job here. I got a DM from a user who told me that he had a balance in pmUSD, and my post helped save him before the depeg happened. I think that was one of the cases that really put Faros on the map as a credibility signal.

Guys, thank you for joining us. How are you doing?

Glad to be here. Thanks for having us. We're doing great, actually building the future of Faros as we speak, and we're excited to talk about it.

This is what we call a DeFi frontier podcast. I think I first heard about Faros through Ike, actually. I saw you doing some sleuthing on X, and you were using or referencing Faros. I didn't even know you were part of Faros at that time. Since then, it's become a core tool for most of my research, and it ends up in our newsletter all the time.

We're going to talk all about Faros, why you built it, and how you're using it to spot certain depegs. One of the more famous ones that you pointed out ahead of time was pmUSD. What signs did you see ahead of time that suggested things weren't looking good? Then we want to give people a simple checklist, so every DeFi user should know a few things when looking at stablecoins. What can they trust, and what can they look at to verify or assign credibility to certain stablecoins? I'm sure we'll lean heavily on the Faros website for that.

1. Origin of Pharos

To get started, I'm curious how you guys even met and what drove you to build this very unique DeFi platform.

I can tell the first step of that story, and then you can continue. How we met is actually quite funny. Faros has essentially been possible thanks to AI. I'm not a developer originally, and AI helped me cross that barrier.

I was really focused on building and refining the formula for liquidity scoring and safety scoring, increasing coverage, and making the infrastructure better. I learned so much about developing and building Faros, but that's another story. I also have another commitment. Faros is not my main commitment, believe it or not. I also have Polaris, which I'm working on, and that's my main commitment.

I quickly realized that I didn't have time to properly promote Faros. I didn't really think about it that way, but I decided to run a contest because I wanted people to use Faros's data. I was frustrated that I had such good data, but people weren't using it on Twitter. They weren't aware of Faros. So I made this contest where I said, "Hey, use Faros's data to tell a story, whichever way you want," and I offered some prizes. I think I put up $3,000 in total prize money with a tiered structure.

What happened is that TokenBrice absolutely crushed it. He made by far the clearest and best research post about Faros. It was very well done. After that, we started talking, and I pretty much offered him a job. I was like, "Bro, do you want to do this for a job? Because you're pretty good at it." He said yes, so I said, "Okay, consider the contest your signing bonus, and you're hired." I started a team, and boom, here we are.

I had been using Faros for about a week or two before the contest itself. At the time, I was still pretty much in the meme community. I'd been a Frax maxi for a while before I got into stablecoins completely. I started using Faros as well, and when I came across his contest post, I thought, "I'm already doing this, so why not just go for it?"

I did some research, and I think the one that made him notice me was the post I wrote about how to use Faros like a pro. I think that was the one that really made him notice me. It had about 20,000 views, and it was pretty good research. He was like, "Hey, what do you think about this?" Then we started talking about more ideas.

That day, we talked about a few ideas that could make Faros better. He said, "I'm thinking about getting someone for this champion role. What do you think?" I said, "Okay, I think we could talk about that." The next day, I presented him with a whole PowerPoint presentation with my plans for Faros. I said, "I think Faros's product could be this good," and he said, "No, I think it's just going to be a small dapp." I said, "No, I think this is a really solid project, and we can make it even more ambitious than you currently envision."

That's pretty much how we got together and started working. There was a lot of synergy because I think we have quite a number of similar interests. Faros is very transparent, and I think I am as well. It wasn't about the money for either of us. It was more about actually making a change and contributing to this. That was pivotal for me and helped increase our bond. We've just been nerding out ever since then, and here we are right now.

Faros started for my personal use. Initially, it was offline. It was just a dashboard I made for myself to track stablecoins because I was tired of having 5 to 7 Dune dashboards open all over the place. A few days in, I was impressed by what I could make with AI. I realized it was already better than things that were out there, so I released it. It took me a while to accept that Faros was a product, and TokenBrice was definitely the decisive factor in that.

It's such a great story of getting involved by understanding the product and leading with enthusiasm. I think that's some inspiration for anyone who wants to find a new role in the DeFi industry.

2. Why DeFi needed Pharos

Before we start to actually walk through the tool, because visually it really comes to life—you have to see all the information and research that you can learn about these different stablecoins through Faros—let's talk about how the tool is designed to look at all these different stablecoin offerings through the lens of risk.

It's not about what's already depegged. It's about trying to get ahead of what might depeg, and I think that tends to lead through the lens of risk. Maybe talk to us about how this tool is designed to look at all these different stablecoin offerings through that lens.

Someone said at a conference that most of the time—I'm paraphrasing him now—what actually causes problems is always in plain sight, but people simply don't know where to look. Most stablecoins that are quite messy are visibly messy, but people just don't know where to look.

The average stablecoin user is looking for the next chart to go to and the next yield to earn. They're thinking, "Okay, it's pegged to $1, it's safe, and it's offering 10% APR or APY, so it's safe." But what really makes a stablecoin worth using is that the underlying layers are safe enough for you to hold your money in.

For us, what we believe makes a stablecoin useful isn't that it can offer you a peg. It's really about it being safe. Before, it was quite hard to do all this research in one place. You'd have to open a Dune dashboard, DeFiLlama, CoinGecko, and various documents. It was just so complicated. You'd have a lot of tabs open because you wanted to know, for example, what BUSD had as collateral or what frxUSD had as collateral.

With Faros, you don't have to open 10 tabs. You can just open one tab and see all your research here. I think that's the summary, and then TokenBrice will go into more of the details as well.

That's good information. One of the key lessons I had from building Faros, which helps explain what the website is doing now, is that you can take a lot of AI compute and crystallize it into a working infrastructure that monitors the situation for you without needing the AI to constantly run.

A question I get a lot is, "Do you have 10 Claudes constantly watching what's happening to update Faros?" That's not how Faros works. It simply reads the chain, reads all the oracles, gets data from all the APIs, tracks the collateral deposited by collateral-backed stablecoins, and tracks the actual collateral to see how much collateral is in every given branch, and so on.

That's really the unlock. You spend a lot of time and resources at some point to create infrastructure that monitors a given point, and then you just have to keep that infrastructure running. But you have constantly refreshed data on the topic. That's the big unlock, and it's pretty unique to the website.

For instance, let's take something really concrete so everybody can relate.

3. Screenshare demo of Pharos platform

If you think of the collateral backing a stablecoin, usually, a stablecoin analytics platform will review the stablecoin at one point in time, then produce a report and tell you, “The collateral is USDC and this and that, and this is good or this is bad.” But this is not what Faros does. Faros monitors in real time, and every 2 hours we adjust the reserve composition of stablecoins. So actually, that’s maybe where we can start to show the website, because it will speak much more. Let me get the tab ready, because we just redesigned the website, and I really like how the new reserve composition is displayed.

Again, a lot of this ultimately goes back to research on the different stablecoins that Faros is monitoring. The goal here is for you to better understand how to use this tool. The next time you’re looking at a stablecoin, thinking about putting money into a yield strategy, or simply holding that stablecoin, we want you to be able to understand how to use this tool to assess the risks associated with it. TokenBrice, lead us forward.

Yeah, a good example of this infrastructure is the reserve composition, which is updated every 2 hours, if I’m not mistaken. I changed it a bit because it was draining a lot of resources, but it’s really cool because it’s close to real time. We also made it very visual, so even people listening to the audio can follow along. Here, we’re looking at the USDT reserve composition, and we can see a very large block, as we would expect.

This large block is essentially T-bills. So far, so good for a centralized stablecoin, right? But we also see smaller blocks, about 10% to 12% each, and those are the other assets that USDT has in reserve, which is more unusual for a centralized stablecoin. So you have some gold—physical gold bars—you have some Bitcoin, and other forms of cash reserves.

For those following along on video, USDC has a simpler reserve composition. It’s all essentially bank deposits; they’re just different forms of bank deposits. You can really see things appearing very quickly with that visualization, and that’s what we like.

Another one to look at is USDe, because this is the one that inspired this feature. I was tired of hearing and reading that USDe was backed by a basis trade when it had been a year since that was no longer the case. I think this reserve composition makes it very clear. For those who aren’t watching, USDe is now 99% stablecoins and 1% basis trade, so you can imagine that on the graph, you can barely see the basis trade. It’s that tiny yellow line on the right.

I think that’s a good demonstration of the whole point: you get real-time data, and you can see those changes happening. The color codes are designed so that if you stumble upon a stablecoin you don’t know and see that the reserve is very orange and red, you immediately know that the collateral is probably not top-grade collateral. I really like that intuitive feeling you can get just by looking at it.

4. What factors into scoring a stablecoin?

TokenBrice, would you mind pulling up Liquity’s BOLD? I feel like that is the highest-graded stablecoin that you track. I’m hoping we can walk through it as a benchmark for why the safety score is so high. This is a great place to start understanding what benchmark other stablecoins should be trying to achieve and what ultimately makes this a safer stablecoin.

We’ve looked at reserve composition, but maybe walk us through some of the other components that make up a good score. I see all this stuff over here, like DEX liquidity, redemption, backstop, TVL depth, decentralization, and dependency risk. You don’t have to give us an in-depth analysis of each one, but a high-level overview of how the system is incorporating these things into the underlying score would be helpful.

Yes, and we won’t go too much into the detail because the new version is coming and will be a bit different. I don’t want to confuse users, but the overall idea will remain similar. You have this infrastructure tracking all these data points, and we’ve taken a deep dive into reserves, but there are other concepts that are very important for stablecoins.

One dimension, for instance, is liquidity: how much liquidity do you have on decentralized exchanges to be able to swap in and out of the stablecoin? But a lot of stablecoins don’t use liquidity on DEXs. They might have a redemption model or a price stability module, or PSM, like DAI. Pharos accounts for all these different types of liquidity, which make up the exit-liquidity dimension.

We’re talking about stablecoins, of course, so one of the big dimensions is whether they can actually keep their peg. Pharos has a very—I’m not afraid to say it—the most comprehensive peg database I’ve ever seen in the history of DeFi, with about 67,000 peg events recorded across all stablecoins. That allows us to give a score for a stablecoin’s ability to maintain its peg.

Other dimensions involve resilience and dependency risk. The collateral we just discussed is a factor in the resilience dimension. Another important point, which we can probably dive into afterward, is minting authority: how the stablecoin can be minted, under what conditions, and whether there are safeguards. That’s probably worth a spotlight because it’s a very frequent attack vector, but it’s also a factor in the safety score.

Dependency risk would involve things like stablecoins using bridging solutions. They might use LayerZero’s Omnichain Fungible Token standard or the Chainlink standard, or they might use another stablecoin as collateral. Then they’re dependent on that other stablecoin, or some part of their mechanism is dependent on other things. It’s the whole network of dependencies around the stablecoin: the stablecoin itself might not fail, but perhaps something upstream fails and contaminates the stablecoin, as we’ve seen many times in DeFi.

All of this is aggregated into one singular metric, which is the safety score. To perform well on the safety score, you need a stablecoin with high-quality collateral, very good liquidity, a minimized dependency network—ideally, no dependencies—and a bridging model or other mechanism that doesn’t create weaknesses. It also needs to be good at maintaining its peg. If you have all of those things together, you can reach a pretty good safety score.

Just to give some context, those are our top performers. With the caveat for those watching right now, the safety score is currently pending an update. I’ve changed my mind on this one because one of the dimensions is decentralization. How decentralized your stablecoin is has been a dimension in the safety score, and through the iterations of the safety score, I kept reducing its weight. Otherwise, you only get maximally decentralized stablecoins at the top, and something like USDC can’t score as well despite being more or less safe.

In the next version, I finally accepted that decentralization will no longer be a scoring criterion; it will be there to inform users. That means you’ll potentially be able to have USDC as a Grade A stablecoin or something like that. So right now, the top stablecoins you see are more on the decentralized side of things.

So, for those still listening to us, the currently A-graded stablecoins are BOLD and LUSD from Liquity, yvDAI, which is the yv wrapper for DAI, and USDS from Sky, as well as Savings DAI. Ike or TokenBrice, is there a threshold of safety score that you don't touch? If it goes below a certain safety score, is there a line that you've drawn in the sand for your own portfolio?

I've sifted through some of the different factors that ultimately add up to that safety score, and that's been the one thing I've been trying to figure out. This is all super helpful, and I'm wanting to act on this information, if anything, just to avoid certain stablecoins. Is there a line in the sand in terms of how low a score could go before you would no longer want to touch it?

For me, it depends on what makes it unsafe. As you can see from the graph that TokenBrice is sharing, some stablecoins are rated poorly for certain reasons. Some have one part that's better than other parts, but generally, without going too in-depth, I don't go below a C.

For some reasons, I would go to a C. For example, frxUSD is a C+, I believe. If I'm keeping money in quantity, I wouldn't go below an A or B. But for general reasons, I can go to a C. Once it gets to a D, then it's very risky. It's a D for different reasons, you know.

I think that's probably the last question about pmUSD, so I'll leave that for when we talk about pmUSD. That was how I was able to spot pmUSD as being very questionable, because it was a D. Every stablecoin that's currently a D is a D for a certain reason that you can really probe into, but I'll wait for the pmUSD story.

5. How Pharos called out pmUSD ahead of its depeg

Hey, Ike, let's get into that pmUSD story. The protocol is called RAAC, or something like RAAC. This was interesting because you pointed out issues with pmUSD, their quote-unquote stablecoin, ahead of the depeg. You mentioned that you used Pharos and it showed up as a D. Could you give us a little more insight into what you were looking at?

You pretty confidently called it out on the timeline on X ahead of time, and then it turned out exactly the way you said it would, unfortunately for the holders of pmUSD. Walk us through that.

TokenBrice and I love to be transparent with everyone. We love to see everything come to us, and we want people to talk to us about everything. We got a DM from RAAC. I think they're pretty anonymous, but they sent us a DM after my video, and they were basically saying that they didn't like how we rated them.

We got talking with them because TokenBrice tried to talk to them about it: their collateral was rated poorly because it was actually poor. There was some back-and-forth there. What I usually do, because I think I'm like TokenBrice's white angel, is try to be the one who goes back and thinks about it in a human way. TokenBrice can be so passionate about these things, but I try to meet everyone as a human being.

I went back to look at it to really see if what TokenBrice was saying was too harsh on them. At the early stages of Pharos, we had a lot of back-and-forth with our ratings: “This is a bit too harsh. This is a bit too lenient.” I was doing the same routine check to be sure we weren't rating pmUSD too harshly, and that was really what got me into pmUSD in the first place.

My first red flag was that it had no redemptions. How can you grow a stablecoin to a $100 million market cap and have no contingency plans—no plans at all for what happens if this peg breaks? We've been here many times with stablecoins in DeFi, and you would expect that everyone is going to learn from this. How would you have a stablecoin with a $100 million market cap at the time and have no redemptions?

Secondly, I went through the documents, and the documents used some very vague terms that the average user would be fooled by. I think the first one for me was the gold in situ. pmUSD was marketing itself as having a gold vault, but the actual term was, I think, the Latin phrase “in situ gold,” which means unmined gold. You're practically telling me that I should trust you with my money for something you haven't even seen—something you think is in the ground.

Users were thinking that there was some gold in a vault somewhere that they could use for redemptions or whatever it was, but it wasn't there. It was nonexistent.

The third thing that was a flag to me was the liquidity. There were very few actual pmUSD stablecoins on the DEXs. They had about $100 million in supply and many bribes on Curve, which was slowly bleeding the liquidity out. At the time of my post, the liquidity was about $375,000 for a $100 million market cap stablecoin. That was a big red flag for me, because you're essentially telling me that if I need to exit a position of any size, I can't do that.

You're telling me that if something happens and everything goes to zero right now, I can't get my money back. Most of these things, like we said at the beginning, were in plain sight, but no one had the time to look at them.

I told TokenBrice what I found, and I think we could push it out together. It was a very heated time for us because I had to spend a lot of time on it. We had some back-and-forth with the data, and we had enough faith to actually push it out. But we couldn't prove everything.

TokenBrice said, “You know what? Push out only what we can prove,” because some things were just stories I had in my head about why this was that way. So we pushed out only what we could actually prove, and the market reacted to it. Some people were saying, “Oh, I can buy this. This was good research.” It was a pretty interesting time for me. I had a lot of racial comments as well during that time.

Three days later, the depeg happened. Then people started to see that Pharos had done some good work. I got a DM from a user who told me that he had $1 in pmUSD and that my post helped save him before the depeg happened. We had a lot of people talking about this, and I think that was one of the cases that really put Pharos on the map as a credibility signal.

What I really looked at there was already visible. I was just using my slow brain to see that these were red flags, and I didn't know why nobody was looking at them. Before that, pmUSD had about $6.3 million in DEX liquidity. Whenever the depeg happened, it was talked about as an arbitrage opportunity.

For me, that was very deceptive, because if a stablecoin is depegging for a reason, you don't need to come out here and say, “Oh, I can arbitrage it every single time.” You need to know that yours is bad. Sorry, forgive my language. You need to fix your mechanism. You need to fix everything.

6. Tensions with stablecoin protocol teams

That was really what got me to pmUSD and how the story began, because Pharos is now becoming such a respected research tool.

You are ultimately the bad guys when you do your job. You're the bad guys for whoever you're calling out, and you're the bad guys for calling out poor mechanism design, which, again, is an opportunity for improvement.

I do think there are lots of builders out there who ultimately have shortcomings in their designs, and some of these depeg events are just a moment in time. It's a wake-up call for them to make an improvement. Some of them obviously are not; some of them are bad actors. It's very difficult to distinguish which of the 2 they are.

If you've grown up with crypto over the last 10 years, part of that philosophy has been to never blindly trust. The best thing you can do is verify, and this is a tool that helps you verify on-chain what's actually backing these stablecoins.

Honestly, with Ethena's USDe, I had totally lost track of how little was backing it in terms of the basis trade. I definitely would have bet more if you had asked me. So here we are, just doing a live interview, and I'm realizing and learning things about pretty notable stablecoins that I thought I understood. In fact, there's something different when you look on-chain.

There are 360 of them or something, so it's very easy to lose track, especially with the pace of change. What you were saying is actually very interesting, because builders' reactions to Pharos can be on both extremes.

With the pmUSD case, we got some pretty nasty comments. I had one person who tried to get the European regulator after me, so we've had some interesting reactions, let's put it this way.

On the other end, builders are using Pharos as a measuring stick to evaluate their efforts. In the early days of Pharos, I got—I’m going to keep it private—just say, a builder reached out to me with a message like, “I saw on Pharos that my liquidity score is only 40. I sent a message to my market maker, and if it doesn't bring it to 50 in the next month…”

I was like, “Wow, that's amazing.” The builder is actually using the tool to measure a weak point in their stablecoin and benchmark against it to see whether there is any genuine, real progress. I must admit, that's not something I considered initially when I built it. It was more about the other angle: the user wants to know how liquid that stablecoin is. But I really love that use case as well.

Yeah, guys, something I was thinking through too, just listening to your answers there: I'm sure it can be hard to know when you should go public with something, because there's potentially a worry of, “Hey, are we going to be the ones who sort of create the cascade?”

But thinking through it, this space needs to be like Darwin in many ways. If the emperor is wearing no clothes, the people need to know, and it’s going to happen eventually.

One other thing I was thinking about is that we had L2BEAT on the podcast, and I feel like they did this thankless job of rating all these L2s and pointing out who was on the right side, who was on the wrong side, and taking heat from everywhere. You guys strike me as the L2BEAT for stablecoins, just shining a light on the good and bad of the industry. It’s much needed because what it ultimately does is, hearing what you’re saying and all the DMs you’re getting, you’re slowly raising the bar, I hope, for the industry. That’s what’s needed. We brought this up, too, with things like Credora, ratings, and all of that. I think it just starts to help things get better. So kudos to you guys.

7. Taking research public, before a depeg

I want to go somewhere else, though. I’m thinking through maybe if there’s a DeFi user out there who’s newer to the space and looking for red flags—we’ve talked through some of them already—but what things should set off alarm bells when they’re looking at a stablecoin? Do either of you want to run through an example of how you like to analyze things? What are the most important things you could run through if there were a top 3 or top 4? Either Ike or TokenBrice.

Before you go to TokenBrice, let’s address a question you asked before. You asked how we know when to go public, and I think that should be answered because it’s quite an ethical part of this as well. You asked when we go public and at what point we actually take our research public.

What I do, which is a key part of things, is try to look at the intent of a player. I will call out a stablecoin that’s doing something wrong in the industry and has been rated, I think, a D on Pharos. I’ve spoken to the founder a couple of times, and I can see that he’s generally trying to make amends. He’s genuinely trying to get his stablecoin to a good standard. I can see that he’s really put in the effort. He’s just a bad dev or a bad founder, but he has a good heart. There’s a difference between being bad at what you do and being a criminal or a grifter.

So, for people who are genuinely trying to do better in the space, I like to give them grace. Obviously, on Pharos, you can see all the detailed data there, but I wouldn’t push a tweet talking about how bad this is. It’s on the Pharos dashboard. It’s public information; you can always see it yourself there. But I wouldn’t push a tweet breaking it down.

When I push a tweet breaking it down, it means that I’ve been able, in some way, either through inside information or through relationships, to see that this particular founder is quite sketchy. That’s the point at which I go public with my research—or with our research.

With the pmUSD matter, these questions were asked on Discord over and over again by different community members, and they were always deflected. Siderum Rock[?] released the news saying that they had an agreement with APX and Kitco. I have sources in these parties who say that it was never in place. Somebody came to Discord and said, “I called someone at APX and Kitco asking about this particular partnership, and they said no.” The guys on Discord dismissed the person over and over again.

Before I made my post, I had someone call someone on the team. You can obviously see someone intentionally trying to deflect or downplay what they’ve done. At that point, that’s a bad player. That’s a Ponzi, a bomb waiting to explode. That’s when I put out my post and pretty much just blow the whistle on you.

Actually, I just want to add to yours because it’s so true. I would say the biggest factor indicating either that there’s a rug waiting to happen or, if not a rug, that you’re going to be played, is exactly what Ike was saying: misdirection and lying.

To give you a concrete example, the whole pmUSD story started because they weren’t happy about our collateral rating. They said, “We rated it too risky.” So I asked them for a bit more detail, and essentially the argument was, “All collateral is fully on-chain, so it should be rated like on-chain collateral.”

I was a bit surprised, so I looked at it, and then I went back and thought, “Those tricksters—they’re telling me the collateral is on-chain, but essentially what I have on-chain is an NFT saying that some entity off-chain has some gold in a mine, and they call that on-chain collateral.” At this point, I knew I wasn’t interacting with an honest-behaving actor, to put it that way.

This kind of thing is what you should have in mind: if you catch a fund actively misdirecting or misleading, it’s game over immediately. You shouldn’t give it a single chance. Get all the funds out, and never touch anything the fund is even invested in again. You need to be this hardcore. If you do that, you’re already saving yourself from 90% of the stress.

8. Case studies of depegs

For the section about what to be aware of with stablecoins, I think we can go back to screen sharing a bit because there are quite a few things I want to highlight here.

First, very quickly, on Pharos we actually have a Learn section, and in there you’ll find case studies of previous depegs. That’s a really great way, if you want to deep-dive into mechanisms and understand how they effectively break and what it means. The case studies are pretty diverse: you’ll have case studies of stablecoins that died, but you also have case studies of stablecoins that survived, each augmented with Pharos data, a timeline, the price chart through the whole event, and so on. I think it’s a great way to learn. You’ll find it in the Learn section on Pharos.

9. Explaining the USDC depeg in March 2023

Could you talk about the USDC Silicon Valley Bank incident just for a minute or so here? That’s fascinating to me. It’s one of the best examples of a major event where there was a momentary depeg. Depending on your preference in terms of stablecoin, USDC remains one of the most used and most liquid stablecoins out there, albeit it’s definitely more centralized if you’re looking for something that’s more uncensorable and decentralized. What else can you tell us about the case study here on the USDC incident?

Yeah, I guess the main story of this one is really—I think the technical term they have for it in finance is a liquidity-duration mismatch. Essentially, what it means is that at no point in time—I mean, people thought so, but effectively at no point in time—was there a problem of missing collateral. The money was there. The problem is: was the money there when people were redeeming and needed it? That’s why it’s a duration mismatch.

One of the underlying banks holding the deposits for USDC, as we saw in the reserve composition just before, was Silicon Valley Bank, which was suffering some stress. But the stress manifested just at the onset of a weekend. What happened is that panic propagated through the markets. People started to offload a massive amount of USDC, but the redemption process—the banks are essentially closed on the weekend—meant that redemptions were not able to be fulfilled. This is why USDC lost its peg, down to 0.87 at the lowest, sustained for a few hours. The whole thing unfolded over a weekend.

So it’s not a major event, but it was still very insightful for that. The duration mismatch—and I guess the second big thing is also something that Pharos shows very well. I’ll show you on screen: it’s the dependency map.

A lot of stablecoins were dependent on USDC, and as USDC depegged that weekend, a lot of stablecoins were caught along in the depeg. FRAX, DAI, and plenty of others. That’s something you can really see if you zoom in. Actually, I’ll show you an even better view.

We have this dependency map that shows you which stablecoin is used where and what it depends on. You can focus it on a given stablecoin if you go into the Context section.

So you can see that USDC is essentially a bit of a center of this spiderweb, right, with a lot of roads going to USDC. Those roads can be USDC being used as collateral, USDC being used in a stability mechanism, or some other form of dependency. Essentially, all those coins that you see now on-screen are reacting to a USDC depeg, as listed here. So you have Aave, for instance, recorded as using it as collateral.

10. What are red flags to avoid in any stablecoin?

If USDC was to depeg again, like it did during the SVB weekend, those stablecoins will likely be affected. So, yeah, 2 big learnings would be that duration mismatch and how USDC-dependent DeFi can be. And then, Bryce, if we're going back to that average DeFi user, what are those top things again that come to mind for you, having looked at this stuff so many times?

Yes. Another thing that might be of interest here is what we call the cemetery. The cemetery is essentially a ledger of all the dead stablecoins, if I can manage to click it with my mouse. One thing that is interesting is that on the cemetery, we're also conducting statistics on the cause of death and the type of stablecoin that died.

If you're worried about this, you can essentially get an idea of which causes are most likely to lead to a stablecoin exploding, without taking my word for it and simply looking at the data. One of the biggest ones, at least this year, has been everything related to the minting permissions of the stablecoin, to the point that we dedicated a whole section of Pharos to it. That's what I was alluding to earlier; that's what we call the minting authority.

I think I just passed it. No, it's a bit below. Yeah, minting authority will be here. I'm showing you a bit of a horror example on purpose, because we're looking here at MIM from Abracadabra, which has a very risky minting setup that has already been abused 3 times. The stablecoin is currently technically dead. We need to pull the plug on it. We haven't put it in the cemetery yet, but it's trading at 12¢, to give you an idea, and I really don't see a recovery path here.

Minting authority is really important because it is the most dramatic failure a stablecoin can have. To put it in layman's terms, with a weak minting authority, you can have uncollateralized minting of very large amounts of the stablecoin. It essentially means the stablecoin is going to turn worthless really quickly. That's what we've seen just this year alone, so you see it's a very common attack vector.

That's why we really wanted to have this on Pharos and make it explicit to people. So I showed you a dirty one. We can do a clean one. When you want a clean one, it's easy: you just go to the BOLD page. On the BOLD page, you can see a very clean minting authority because the only authorized minter is an immutable smart contract. The only way you can mint BOLD is by supplying the appropriate collateral and borrowing your BOLD, so it has a really reduced attack surface on the minting authority.

You can have other setups that are safe, even with other types of stablecoins. Now I'm putting in USDC. It's not necessarily the safest, but it's not MIM-level; it's already a bit cleaner. If I remember correctly, AUSD—Agora Dollar—I think is doing pretty well on this dimension, to show you a more centralized stablecoin that can do well here. Oh, no, it's not doing so much better.

Maybe another one, just to try to have a bit of diversity in the example. I think crvUSD is—yeah, crvUSD is a decent in-between. You see, it's scoring 63 out of 100. You have a situation that is pretty similar to BOLD at face value, meaning that one of the main ways to mint is to supply collateral on Curve and then borrow your crvUSD.

But crvUSD is scoring a bit lower than BOLD because it also has this part where Curve governance—so veCRV voters—can decide to vote to mint crvUSD for certain purposes, creating unbacked crvUSD at the time of minting. This has been done a few times to supply some pools related to yield-bearing assets, and obviously this constitutes a potential attack vector. If governance were to be attacked, it would result in a lower score.

Minting authority is really the all-star killer of stablecoins here, and probably of all time. If you only watch one thing, this is the thing to watch for. All the stablecoins that failed were, of course, scored pretty poorly here.

We have archived data, which I think you should be able to see. UST, for instance, had a 10 out of 100 minting-authority score. But you get it: you will get at least the information here, because it's not just about the score. It will give you exactly the conditions of who can mint and what they can mint, so it's really good data to look at.

One more legacy stablecoin that I think is completely dead is Synthetix's sUSD. Any takeaway on that, or anything we can learn here from what you have in Pharos?

Yeah, there's a lot here as well. This one is a bit of a special case, right? As you were saying, it's been a slow death. It's been hanging in the twilight zone of 50¢ to 80¢ for about 18 months or something like that.

I think we can still see the full chart. You can see the full history here, where it was essentially at peg for its first 3 years of existence. Then, for more than a year afterward, it's been off peg.

Another good lesson from this one is how governance is essentially causing all of that. This was a purposeful decision by governance. One thing that I like about Pharos is that we have these epitaphs for the dead stablecoins. They're written by Fable, but I review them, and I expect Fable to be a bit spacey and so on.

The title he chose here is really perfect: “Save the protocol at the cost of its daughter.” It's a perfect recap of the whole story, right? sUSD was supposedly sacrificed by governance after being completely neglected for 18 months. That is what happened here.

11. What’s next for the Pharos platform?

It's not a case of technical failure. I mean, we can talk about the problem that the very idea of using Synthetix to back a stablecoin wasn't the smartest one to begin with, if we're honest. There are a lot of things like that to discuss, but what I want to stress to you is that there is a timeline where Synthetix makes another decision and sUSD is still alive today.

What's next for Pharos? I've been curious, just as a user. Is this something you're monetizing? Are people paying for access to this in any way? I'm just curious what's generally next.

The general concept, and one of the strongest models that we have at Pharos, is that the general Pharos dashboard should always be free. We want every user to be able to access the general dashboard perpetually and always for free. This is what we're doing to ensure that we have a safer DeFi and safer stablecoin industry.

If you look at most of the competitors we have, there's no one really like Pharos, but the ones doing something close have gated paywalls. What we want is for every regular DeFi user to at least be able to do their due diligence using the Pharos dashboard.

Where it gets a step further, which we started doing last month, is API integrations. At first, we had our API keys, and there were very minimal fees for people to use them to integrate into their own products and build whatever they wanted to build. Right now, we've gotten to a point where we can trust our API keys much more, so we're talking about exploring having our API keys be paid for. Generally, though, the dashboard will always be free.

What we want to do with Pharos, which I believe is the next step for Pharos, is that every major protocol would have Pharos safety scoring integrated. If you go on Pendle, for example, or Curve, you should know how safe every market is. Look at what happened with the Alpha Pink pool[?]. That kind of due diligence should already be on the market itself.

When you want to lend into a pool, borrow, or supply liquidity on any pool or market, you should have a Pharos rating next to it that tells you how safe your money is. If you look at our yield intelligence on Pharos Watch, we basically tell you all the yields—and I mean most of the yields, really. We have a number of places that we pull our yield from, and recently we added Vaults.fyi to it as well.

We tell you all of that and how safe the yields are. We have our yields grouped into 4 categories: “Why bother,” “Playing it safe,” “The sweet spot,” and “Danger.” Essentially, what we want to do with Pharos in the next step is that, whatever you do with your stablecoins—borrow, lend, stake, or anything else—the protocol should have a safety integration.

Right now, we're talking through a number of protocols and maybe having a couple of meetings with protocols to fine-tune these discussions. This quarter, we're expecting to have many more integrations with safety scoring, because every DeFi user deserves to know what they're putting their money into.

It’s beyond just the APIs. It’s beyond the API, rather. It’s beyond the trenches and all of these things. DeFi users deserve to know how safe the strategy is, how safe the collateral is, and how safe the liquidity is.

Yes. On the monetization of Pharos, we do have a funding page where you can see the current cost of Pharos, and we’re open to receiving donations. But that’s something I’m very firm about: I would rather take Pharos down than gate the website behind a paywall, because that defeats the whole purpose of what I’m trying to do here and what we’re trying to do with Ike.

We’re exploring several paths. Community donations are one, and we conducted a pretty successful first Gitcoin round that gave us about 2 months of budget. To be honest, the other good news is that Pharos is a very lean ship. The current burn rate is about $1,800 a month, and most of it is compensation.

Compared to similar projects—which, first, nothing is similar, but second, they have teams of 10 or 15 and are raising money—I’d rather take the other approach of keeping a lean team of people who really believe in it and then growing from here. We’re looking at a few things. Most notably, it would be high-frequency API keys.

I also want to keep the API openly accessible, but we did have to restrict the number of requests per minute that people can make because we were getting hammered in the early days of Pharos. There was a bit of panic, I can remember. I’m considering reintroducing higher-capacity keys now that the infrastructure can stomach it, with some form of subscription or something like that.

I really think the endgame is what Ike was saying: essentially, a form of service provider for DAOs, centered around what Pharos already does but with an additional layer of work and customization adapted to the DAO-specific need. Our project doesn’t necessarily have to be a DAO, so we’re looking into that now. That would be the main path toward monetization.

So far, I’ve been sponsoring the losses that are happening every month. We kind of give ourselves until the end of the year at this burn rate to see if we can bring the ship to balance. Honestly, I’m not too worried that we’ll get there because, accounting for the donations and everything, out of the 4 months or so that Pharos has existed, we were able to cover about 2 months of expenses. We’re already about 50% sustainable, if you can put it that way.

Let’s see how it goes. The target is really not insane or far in the future. You just need 1 whale that really likes Pharos to send us $15,000, and boom, Pharos is funded for a year, essentially. It’s as simple as that, if someone’s listening.

For the next steps, the real big one is this new safety score. We’ve been adding a lot of features to Pharos over the last few months, but the safety score kind of remained just an incremental upgrade. I’ve reached a stage where I realized it was due for a complete rework from A to Z to factor in all the nuances.

The new safety score will bring some really interesting things. Maybe the main one is a change in how the logic operates overall. Right now, the safety score grades you in different dimensions. What that means is that you could have, for instance, very risky collateral, so you’ll have a poor resilience score. But if you have excellent liquidity and minimal dependency, your overall score will be lifted up, and you can end up being, let’s say, a B grade with very unsafe collateral.

There’s no logic of, “This is disqualifying,” or, “We should really hammer the score of this project because this specific dimension represents a broad risk.” That doesn’t match how we assess risk in real life. In real life, there are some items where a threshold is crossed and you’re like, “No, I’m not touching that. I don’t care about the rest. You lost me here.”

This new safety score will have that logic of thresholds being crossed on certain dimensions that will hammer the score down, so that you cannot have the compensation effect that you have in ratings right now. I’m really excited about this one.

Another feature is that Pharos will become mechanism-aware of the stablecoin being reviewed. To bring it into more practical terms, if Pharos understands your stablecoin as a centralized stablecoin dependent on real-world assets, then the questions of regulation, transparency, and whether you have the appropriate MiCA authorizations will become part of your score.

But that same question doesn’t apply to a maximally decentralized, immutable stablecoin. Who cares if BOLD is MiCA-compliant? I don’t know if it is or not, but what can the EU do about it? Nothing. It shouldn’t be factored into the rating of BOLD.

There’s this branching logic. At the same time, Pharos detects that BOLD is a decentralized stablecoin reliant on oracles, so the oracles will become a factor in the rating. Of course, there’s no oracle to review for USDC, for instance.

Instead of trying to have 1 logic fit all, it has more of a branching logic based on the characteristics of the stablecoin, which leads to much more accurate, fine-grained ratings. I’m really excited about this one, along with the overall simplification of the dimensions and a few other things.

So, safety score V9. I’m really feeling like it’s going to be my magnum opus. I have a good feeling about this one because it’s really taking the Pharos infrastructure and translating all that accumulated data into scores that I think are going to be very hard to challenge.

I’m willing to accept that the current safety score has some bias. We were talking earlier about the fact that decentralization is a factor leading to USDC being only a B, for instance, in the current Pharos safety score. It’s totally challengeable. You can tell me, “No, that doesn’t make sense.”

On the new one, I’m rolling it out and calibrating it right now, so it’s still not perfect. But when it rolls out, I’m hoping it will be fair enough to become the standard for the whole space. That was kind of the goal of the safety score since V1, but of course, it took a few steps to get there.

Well, it’s very admirable work that the two of you do. It’s difficult to be the adult in every room. I do recall when we talked with L2Beat, I was struck by how thankless the work they were doing was behind the scenes. I recognize a parallel here.

The difference is that, as important as it is to understand what goes into Ethereum L2s, what more folks are interacting with—at least at the forefront of trying to earn yield—is stablecoins. We’ve been covering on the podcast that there’s this huge next wave of liquidity coming on-chain, and it’s coming in the form of stablecoins above all else, but it’s also coming in the form of other RWAs.

No matter what, I think we’re always going to see stablecoin liquidity continue to rise exponentially as we bring more and more assets on-chain. It’s very important that we understand what goes into those stablecoins. I can’t recommend or think of a better tool right now that folks should be using to research those stablecoins, especially if you’re holding any part of your portfolio denominated in a stablecoin, setting aside earning actual yield, which takes on even more risk.

One thing I do want to call out really quickly in regard to their funding: if you’re as much of a fan of Pharos as we are, I highly recommend going to pharos.watch/funding. That’s where they track their monthly running cost. It’s currently around $1,700.

Just a quick one, because I realize we didn’t touch on it at all: the most popular Pharos feature of all time is our Telegram bot, which enables users to set up custom alerts. They can pick which stablecoins they want, and they can pick what they want to be alerted about—like, warn me if there’s a depeg of more than 100 bps on only those stablecoins, or if the safety score is changing, or, with the new features that just got released, if the reserve composition is changing.

Pretty much every piece of information you have on Pharos can be monitored through the bot without having to constantly open Pharos. I was talking about how I made Pharos, so I don’t have 50 tabs open, but then you’re always on. The bot allows you to switch the logic the other way: “Warn me when something happens on the coin I care about.”

You can find it on pharos.watch. If you’re interested in everything we discussed today, you’ll probably like this bot, so I wanted to mention it.

12. Closing

Very helpful. On that note, I think this is a great place for us to start to wrap up. Guys, thank you so much for your time. Thanks for coming on.

Thanks for all the hard work you do. We would love to have you back in the future, and I would love to give you the final word here before we go.

Well, thanks for having us. We’re always happy to talk about Pharos. We did talk about the funding, but another way to support Pharos is to spread the love and knowledge of it. Many people are still tracking the stablecoin on CoinGecko when they need to check a peg or something like that. Don’t stay in the dark. You have Pharos now, and we intend to keep it there for as long as we can keep it free. So make use of it. Thanks, everyone, for tuning in.