The Frontier Pod Episode 1 with Dahlia Malkhi Transcript

This is the transcript for the first episode of The Frontier Podcast with Dahlia Malkhi, hosted by Tom Mitchelhill and SpaceComputer CoFounder Daniel Bar.

---

Welcome to the first episode of the Frontier Podcast, a new show where we interview the pioneers pushing technology and humankind into the future and going deep on the original cypherpunk values of cryptography, digital sovereignty and limits of human ingenuity. I'm your host, Tom Mitchell Hill, and today I'll be joined by Daniel Bar, the co-founder of Space Computer. Hello, Daniel.

Hello, Tom. Cheers for running the Frontier pod. Today we'll both be speaking with Dahlia Malkhi, who carries with her an extremely impressive resume of pushing the boundaries of tech, whether it be cryptography, distributed systems, engineering and more.

Dahlia is an advisor of Space Computer, a professor of computer science at UC Santa Barbara. She led distributed systems research at Chainlink Labs and worked alongside David Marcus on Facebook's Libra / Diem projects among many, many other ventures. Dahlia, thanks so much for coming on the show and welcome to the first episode of the Frontier Podcast.

It's great to have you on. Thank you. I'm honoured to be here.

Talk to me about your background in cryptography and distributed systems before crypto. I understand that you've been a pioneer in this space kind of well before the advent of Bitcoin or before blockchain kind of burst on the scene and occupies the mindshare that it has today. Happy to.

Yeah. So I am an academician. I started my career by gaining a Ph.D. at the Hebrew University of Jerusalem and later joining as faculty there.

And I've been doing research in reliability, distributed systems, foundations and crossing over to security and applied cryptography probably for longer than I care to admit. And what I really what made me passionate about this field and pursuing it over the years was sort of bridging between foundational innovation, foundational research, some of the surprising results that I with my colleagues were able to bring and then bridging them to practice. And I've been fortunate to be involved in a number of projects that sort of bridge that gap.

I'll mention three and then you can you can decide if you want to drill in any one of them or all of them. So starting sort of in chronological order, starting in my chapter at the Hebrew University of Jerusalem, working with my students and some colleagues, we introduced Fairplay, which was the first materialization of this concept of secure multiparty computation from the crypto, but the very, very theoretical crypto world at the time and bridging them that into practice. And we built and shipped Fairplay.

And that started a wave of practical materialization of very deep concepts in cryptography. This was the first chapter, second chapter, moving on to a decade that I spent at Microsoft Research, the research club in the Silicon Valley at the time, working alongside luminaries like Leslie Lamport, who introduced Paxos to the world, an algorithm that everybody in the industry uses to manage replicated services, replicated data and guarantee that you don't lose data when one copy crashes or one machine or one host crashes. Working alongside Leslie Lamport and others, we looked at the way Paxos was actually deployed in the real world, in the real systems and realized nobody actually deployed Paxos as written and that there was a chasm, a huge gap between the theory and the practice.

And this led, after many, many vocal sessions with Leslie himself and others, into a foundational work called Vertical Paxos, which really gave the algorithmic foundation what the engineers had in their mind when they were building systems. That's second chapter. Third chapter is after Microsoft closed down the lab, a core group of us founded a new lab called the VMware Research Group in the office of the CTO of a company called VMware at the time, now Broadcom.

In there, we worked on a technology called Corfu, which introduced the concept of a shared log where a lightweight sequencer assigns a monotonically increasing sequence over a cluster, potentially large cluster, like a data center of distributed storage, and then clients can write directly and in parallel to all of these servers. Corfu has been deployed by VMware as well as adopted widely and broadly in the Silicon Valley by companies and led into a generation of storage systems, distributed storage systems that do not suffer from a leader bottleneck or a sequencer bottleneck. And so these are three examples that gave me personally a lot of satisfaction, not just writing, you know, journal papers and publishing them, but going beyond the pages of journal academic articles and working with engineers and working with companies to understand how to harness innovation and advanced state of the art technology in practice.

And I can tell you one of the most satisfying things that happens to a researcher in such environment is when, you know, one of the engineers in a team that you work with is asked the question and before you even jump to answer, they get up, set up and start drawing your algorithm on the board and explaining it to others and they do it better than you. And you're like, oh, my God, OK, that's it, the transfer, the tech transfer has been completed. You can move on to do other things.

Maybe just like touching on, so you're describing how moving from academia to industry there. I think for a lot of people today, we think about cryptography and crypto and the cryptographers are kind of the rock star. But I think there was quite a long while that cryptography was such a niche thing in the computer science department that like very few people care about.

What was that sentiment like both in the academia and in the industry? Like how how was the experience of the pre cryptographers being rock stars era? I would say it's even worse for people like me who I'm not a cryptographer. I use, I apply cryptography, but I'm really an expert more in distributed systems. And there are at least a decade where I would probably whisper that I'm an expert in consensus, I'm an expert in distributed systems.

And then, yes, Bitcoin and the cryptocurrency world. Now you have to say crypto is cryptocurrency or cryptography. It came to our world.

We didn't quite bring our innovation to the crypto world as much as it came to our world. So it was like quite, I don't know if to say lonely, but it was like quite, I don't know, like a niche that few people care. Like you said, it's like, so like even like the researchers.

Totally. I mean, look, in academia, it was not held to a very high prestige if you were doing distributed systems. Industry, you know, there's a huge demand for distributed systems and systems expertise because at the end of the day, if you use the Internet, you're using at the back end distributed systems all the time.

If you're sending emails, somebody keeps copies of your email and guaranteeing that they are not lost. If you're backing up your files to a cloud, to a service like Dropbox or this or that, there are systems behind that guaranteeing that updates even sent from multiple devices simultaneously or while disconnected and later reconnecting, all of these copies are kept consistent and available for you. And of course, if you're using a blockchain, there's a distributed system behind it keeping track of the history of transactions and the provenance of ownership of digital assets.

All of these are distributed systems. OK, that's really interesting. So I think that's something could be useful for the listeners as well.

It's just like a really, really top line difference between cryptography and then distributed systems, just as two separate fields. What's the main dividing barrier between the two conceptually? I would say cryptography deals with worst case adversaries and distributed systems deals with benign adversaries. So it comes down to the issue of trust in systems, essentially.

To a great extent. And whether it's decentralized or centralized. Let's zoom out a little bit and move back.

I am interested as to why you left the academia side of things at the Hebrew University of Jerusalem and went out to build enterprise software research division. Run me through the process of leaving academia and moving into the more private side. So I have to say very personally, I never left academia in my heart, in my brain.

I took a short sabbatical that was in 2004, planned to spend a year as a sabbatical visitor at Microsoft Research in the Silicon Valley, as I already mentioned, and just fell in love and got sucked into the process of actually being involved in deploying innovation, state of art technology and research in practical systems. This was so captivating that I ended up spending the next decade there, giving up my academic position and then pursuing additional roles in the industry. Spoiler alert, at the end of this, as you already mentioned, two years ago, I found myself sitting down and thinking, OK, what's next for me? And I realized that my heart is really with academia.

I never fully left the academic role. So I just left the academic paycheque, not much of it, but I've always worked with students, interns, other researchers, continued publishing along the years and working from an industrial setting on bringing innovation into technology. So for me, it was an easy choice.

I came to UC Santa Barbara, which is one of the most beautiful places in the world and a wonderful, wonderful computer science department two years ago. Talk to me about HotStuff. OK, so we already talked about, you know, first three chapters of doing research that saw real world impact and then something unique happened while I was a researcher at VMware that you're lucky if it happens even once in your career.

And I'm fortunate to have had it happen more than once. So a bunch of us were looking at the foundations of blockchains and trying to bridge the concepts and the methods that were used in academic papers to the ones that were being built by the blockchain engineers and innovators in the field. And we did that for about two, three years, really grappling with mapping, modeling concepts, algorithms, and trying to understand that there was, you need to understand there was tremendous innovation.

This is back in 2015, 16, 17 days. There's tremendous innovation, you know, grassroots innovation happening by developers and they had insights and they had originality and creativity and created systems, but they never had time to sit back and write the algorithms that they were inventing. And sometimes there were flaws and sometimes they were right.

But there was just no rigorous frameworks to think about it, to reason about it. And even if you start with a correct algorithm, if you don't have that algorithmic foundation, you don't know how to scale it. You will break it when you add features or change something because there's no core that you can always test and rely and fall back with.

So we were doing that kind of when I say we, the experts in the field, my colleagues, my close colleagues at VMware, we were doing this almost full time, just looking at what's happening in the field and trying to understand it. While doing it, I attended a blockchain summer workshop by Tsinghua University, which was held in Shenzhen. So for four days, we were stuck there, you know, constantly talking, chatting, brainstorming about how to bridge this gap.

And on the way and gradually, step by step, we gained clarity and we began building blocks. We didn't quite have the full picture of how do you actually scale academic algorithms? How do you scale them in practice the way blockchains are being built? And so I remember driving to the airport, the Hong Kong airport. Suddenly, all these discussions of, you know, the four preceding days started falling into place in my brain.

And this is this is another embarrassing story, you know, my distributed systems role is one. Here's another one. So it came to me, you know, this insight that you can really have distributed algorithms, the rigorous ones, the ones that, you know, are in academic textbooks, you can really have them built around the concept of a block like a blockchain that repeatedly runs exactly the same uniform, simple unit of exchange between a leader or proposer and a quorum of validators.

The leader sends them the proposals, collect the votes, aggregates them, and that's it. This just repeats again and again and again. And there's no need for these complex algorithms that had stages and states that you're in when you're processing messages and different types of messages.

None of that. It's just a uniform building block that keeps repeating itself. And everybody, anybody that looks at a chain of blocks that have these votes embedded in them just looks for a pattern, looks for a particular pattern and knows whether a block has been committed to the sequence or not.

And this was one of those realizations that as soon as you realize them, you know, you're onto something good. I didn't realize quite how impactful it would be, but, you know, it felt really, really good. And here's the embarrassing part.

I got to the airport, I started typing it and I thought, OK, I'll type it on the flight and immediately email my colleagues and share it. And I get on the flight and there's no Wi-Fi in the freaking flight. There's no Wi-Fi, it's a 60-hour flight.

I'm trapped in the airplane. And all I'm thinking about, literally, I'm not thinking about the people on the plane, I wasn't thinking about my family. All I thought was, what if the plane crashes and I don't get a chance to tell the world? OK, so the rest is, you know, every single line that I wrote on that flight and later shared literally on the tarmac when I landed in San Francisco airport, I immediately uploaded everything and sent all the emails before we even got to the gate.

Every single line was wrong and we had to fix it. And we worked for about a year to actually make an algorithm out of it. But that was the birth of HotStuff.

I later worked with an intern of mine and a co-founder of Avalanche, Ted Yin, and Mike Kreiter, who was an advisor at VMware Research and two other colleagues at VMware Research, Yitai and Guy, and we published the paper. And I didn't quite realize it, but the simplicity of having a simple unit that repeats itself, a regular ingredient or component that just repeats itself was so attractive to developers that forget about, you know, any of the foundational contribution in HotStuff, this is really what I think drove adoption. Developers could really see how to implement it and have confidence that their implementation is correct.

And so HotStuff gained a lot of adoption. The first one to adopt it was the Libra project. Maybe we'll talk about it in a second.

And then other industrial projects took HotStuff, enhanced it, variated on it and continued it. We have Daniel on the line and I'm an advisor to Space Computer. We are also working on one of the enhanced forms of HotStuff and benefiting from it to build resilience and trust among satellites in orbit.

But I just want to mention that from an academic or foundational point of view, the a-ha contribution of HotStuff is a concept or property called linearity. We were able to scale Byzantine fault tolerance by making the protocol have only linear communication complexity, something that nobody did before us. And this is a problem that people have worked on for literally four decades since the formulation of the problem by the pioneers in the early 80s.

People have tried to improve these protocols and deploy them in practice and build systems around them for four decades, but nobody has been able to linearize. And by linearizing the proposer exchange, proposer validators exchange in HotStuff, we were able to settle upper bounds for the foundational problem of the Byzantine generals problem in multiple settings. The one that HotStuff works in is called partially synchronous settings.

But we also unlocked upper bounds, tight solutions in asynchronous settings, in flexible models, and even in settings without cryptographic signatures, information theoretic settings. So it really unlocked a lot of innovation and a lot of follow up work, both in academia and in practice. I want to chime in to say I think interesting that both you mentioned Tsinghua, which is kind of the MIT of China, as well as Shenzhen, which is kind of the San Francisco of China, I would say, where the best nerds do very deep tech.

And it's beautiful that innovation in distributed system happens cross border very much in the permissionless and borderless spirit of blockchain. And I wonder in this context of whether it was the Tsinghua seminar in Shenzhen or otherwise, what was it that kind of led you to this realization and solution of four decades problem? What led to the aha moment? What were the specific insights? Yeah, that's a really good question, because I feel like maybe we should really give credit to the tremendous innovation that came out of the blockchain world. So I can tell you hands down, we would never have thought about HotStuff and probably nobody would have if blockchain innovation by the developers didn't bring a lot of the concepts that we then stitched together.

Some of these concepts were Tendermint, the algorithm that underlies the Cosmos blockchain. Casper, there are two Caspers. Casper is the friendly finality gadget by Vitalik.

That contributed... And the other one by Vlad, was it? Yeah, so that one did not contribute directly into HotStuff. But Casper, the FFG and Tendermint were two of the key ingredients or key blockchain algorithms that led us to invent hot stuff. They were neither linear nor modular, but after brainstorming and Vitalik was there in the workshop and we brainstormed with him for several hours, gradually we saw the mapping between what they did and an academic algorithm and put the pieces together.

But it wouldn't have happened probably without people who probably did not have the tunnel view that experts in the field, including myself, have had and thoughts completely out of the box about these algorithms. And that led back to inventing HotStuff, as well as many other amazing algorithms. Yeah, I think it's beautiful.

And of course, Casper is a cartoon character from a series that HotStuff, the friendly ghost is another cartoon character. And that's the origin of the name. Ah, that's the name.

It's also, I think, in crypto, how there's pretty healthy balance between trying out innovating stuff at the frontier, oftentimes getting wrecked for it, but still keeping the crypto economic as a major frontier for the economy. And then like how it ends up also conceptualizing into like legit science, like legit logic and theory. You know, this is not unusual in other disciplines.

So applied physicists discover a lot of things that theory has to later come and provide the modelling and the explanation for. Chemistry, you know, all of these sciences, a lot of things. I don't know, radiation, radio.

Yeah, Marie Curie came to mind, hanging with the bottle, the glowy bottles. People discover them and they have no explanation for them. And then, you know, experimentally.

And then scientists work afterwards to provide the rigorous foundations. So I think engineering is no different. There's no reason why developers who really understand the systems they're building and have the right intuition.

And I have many other examples, but I don't think we have time for that. You know, academicians go back and then prove when do they provide guarantees, what guarantees they actually provide and give the foundation. That's fascinating.

I had no idea that HotStuff was this foundational in that level of just applications across blockchain and crypto more broadly. That's profound. You also spent time working at the Facebook, Libra, Diem projects, working alongside David Marcus.

You witnessed the birth of the Move language. Talk to me about working there and being involved in that process. And we'll get into a little bit in a second about what what eventuated from the Libra Diem project.

But I'm really interested to hear what it was like around that time. With pleasure. So HotStuff caught the attention of the early founding members of the Libra project, and they invited me to help drive the building of the technology of Libra.

Let's stop for a moment and remind ourselves what what was Libra and what was it trying to solve? Libra was trying to take payments, electronic money, the one that we are used to in everyday life, the one that we have banks support these days, credit cards, all these things, and really implement the notion of singleness of money at the private sector level. So that's a bit of a mouthful. Let me break this down.

When I take a dollar bill out of my pocket and give it to anyone else, they can use this dollar anywhere at any bank and there is a central bank and know it's worth for digital money to be exchangeable this way. You actually need some infrastructure for clearing the deposit accounts of different institutions. If you go to your bank and open a bank account and put $100 there, this is a business contract between you and your bank.

The bank owes you, keeps your money safe and puts an entry in their database that they owe you $100. I go to my bank and I put only $50 in my account. And that's, again, a business contract and I owe you between the bank and me.

How in the hell can you send me, hopefully, all of your $100 to my bank account? Let's say you just want to send $50 from your account to my account. How does that happen? How do you transfer your business contract, your agreement with your bank to my bank, which you don't have any access to and you're not in control of my account? How does that happen? You might imagine that at midnight, some bank clerk goes with a suitcase full of notes from your bank to my bank and gives my bank those notes. That's not what happens with digital money today.

We have a system of central banks in every country that allow these banks, that actually have deposit accounts for these banks, they're called master accounts, and provide clearing services at the end of the day between banks and allow them to operate these transfers. This infrastructure today is open only for banks. You cannot go to the central bank and say, hey, I also want to manage some deposit accounts for my friends.

That doesn't work. And so you have at the periphery of the financial system, you have services like PayPal, Venmo, two primary examples, but there are many, many, many others that are not enjoying this singleness of money. They're not interoperable.

A PayPal person cannot send this dollar, it's the same dollar, they cannot send a dollar to a Venmo client without converting and going through bank accounts and all that. What Libra tried to do is create this common infrastructure for clearing digital payments for anybody that runs a service, a digital payment service. We tried to create a common platform.

We realized that if one company did that, it may not be trusted. It would have too much control. So you have to do it with a consortium of companies.

And that was the Libra Association, which I was a CEO of. And it collected 26 prominent companies, each one of which is capable of running the whole operation by itself. So none of them had control or none of them was indispensable.

This is what Libra tried to do. You asked me about the experience. This was the most honourable and exciting project that I could possibly imagine working on.

This would bring financial inclusion, which would bring access to advanced digital systems to anybody in the world, and would clear payments with single-digit fees instantaneously and transparently. And on top of all of this, it was using distributed systems at the core and was built around an algorithm that I helped contribute. There's no question that I had to drop everything and join that project.

That was the experience. This was the mission. This is what we tried to do.

As you well know, as everybody well knows, we failed to launch. But before the failure, it all sounds very positive. So it's like, I guess, maybe like what? Yeah, so let's stay on the positive.

Yeah, let's stay on the positive. We didn't launch, but I think we completely changed the dialogue in the world. When we started, Washington was asking us, so what's the difference between Libra and Bitcoin? When we ended the project, and today with the Genius Act and Mika from Europe, regulators are asking, do you need 3F plus 1 in Byzantine consensus or why isn't 3F enough? We completely, completely changed.

I think the project had tremendous impact. If you look at all of the companies that spawned out of the project in the end and are now continuing to develop the Move language and some of the technologies behind it, I think the project had tremendous impact, even though we didn't launch under the Libra name. Yeah, and I can talk a little bit about the less glorious side.

Hold on, maybe what was even like, who came in Facebook? So now Meta, but who came in Facebook even with this drive and push? Why start a blockchain distributed system cryptocurrency project in Facebook to begin with? Was it David Marcus that went to Tsak and said, hey, I think it's a field we should look at? Was it from Tsak himself? What was the motivation? Because it's not another industry player. This is one of the most prominent Silicon Valley tech company of the time, right? Yeah, my understanding, I wasn't there. I joined a year into the project.

My understanding is that two people were the co-founders of this initiative, Morgan Beller, who was a product lead at Facebook at the time, and David Marcus, who was the head of Messenger. And they both went to Tsak. It didn't take long, but it wasn't immediate when they were able to convince him that people are sending money anywhere in the world over unsecure, slow and very costly channels.

And at the same time, these people all have this. And remember, this was David Marcus. He was the Messenger guy, Facebook Messenger guy.

And he said, you know, there's PayPal default. And they said, look, everybody has this. When they send money, they go to an agency and take a photo of the receipt confirmation of what they send.

And instantaneously, that receipt pops up at the other end of the world in the recipient's throne. All of this is instantaneous, transparent and reliable and secure. The money itself, which is just the light in somebody's database, incurs sometimes as high as 10 or 15% fees.

Now, people pay $15 fees to $100 remittance, cross-border remittance. Foreign exchange losses, often the money gets lost or stolen on the way, and there's almost nobody you can turn to. And a lot of the people who send money cross-borders cannot really suffer these losses, really.

So for them, paying 15% on every dollar that they send is really, really painful. And again, at the same time, a much, much more complicated system, sending their photos, deciphering them, sending it end-to-end to crypto, all that works flawlessly. They said, you know, this is ridiculous.

We should be able to send money. So I guess just to interject that real clear, the concept of remitting and sending a cross-border payment, you were mostly describing the use case of the developing world, I'd say, that suffers from this completely unsecure and unreliable payment routes. I thought initially you were talking about blockchains, that people are sending money over junky networks, but it sounded more like the use case of transferring money, let's say, via Western Union or the likes or other less known brands.

And then I imagine, from the point of view of someone that runs one of the biggest apps in the world, thinking everyone has a phone and already are a constituent on our network, on our app, that's very, it's a strong allure to think, okay, how do we expand it into also that realm of digital payment? So that makes a lot of sense. Let me, this is a great opportunity for me to actually comment on the relationship between a project like Libra or Libra itself, but also similar projects and blockchains. So Libra did not try to reinvent money.

Libra tried to represent fiat currency at par in a way that is interoperable. It was not trying to introduce a new holder of value. It was not trying to disrupt the unit of accounts that economies and governments really care about.

It was not trying to intervene in any sort of way in foreign exchange or any other kind of monetary policies of government. It's just trying to build a payment trails that takes fiat currency, hold it in reserve in the bank and remits it. The connection to blockchain is how do you run this in a way that does not become controlled by a single entity, which we know historically what happens with them.

They increase rent, they increase fees, they become exclusive and increase the friction of using them. Now, the mission of financial inclusion clearly is the strongest in the developing world. It has nothing to do with blockchains.

It's just the reality. But let me also say our research has shown at the time, five years ago that 20 to 30 million Americans do not have access to financial systems and to any modern rails. And we'll probably never have because they don't have address or the right credentials to ever be served.

It's like almost 10%. Like 10% of Americans, that's like quite a high number for being off grade almost. It is.

And to be honest, you and I are privileged enough to think that all the problems are solved. We can send money to each other and instantaneously you get the money. But we pay a very, very hefty fee under the hood for these services.

And this is when we want to send the same type of currency. You ever travel and you try to send yourself money or you try to take money out of one country and move it to another country? It's ridiculous. It takes forever and it costs so much.

It's impossible. So this was the vision by Morgan Beller and David Marcus. It took them a little while, but not too long to convince Zuck that part of the Facebook mission of connecting people around the world has to also embrace payments.

And it will also be beneficial to the business. The connection to blockchain is collective control. But this was not a new coin.

We did not try to introduce a new coin. And this was not a new value that was mined. It was just the payment trails.

So just representing fiat on the blockchain and using that interchangeably as best as you can on an app with an incredibly wide user base. So just increasing the surface area of usability of currency. So it was a collective, common and universal reserve-backed stablecoin.

And it seems like the timing of today would probably be a bit more fitting for this kind of conversation compared to what it was back then. Because like today with the Genius Act and everything else, it's almost every financial institution and their dog is issuing a stablecoin. So I imagine it would have been a bit simpler.

It might have been simpler. But there are a number of reasons Libra has failed, which could potentially still block something like this today. And also, I think we've learned better.

I think the top three reasons in my list, or probably more, is Libra was too early. So you could say now governments, regulators, everybody's more aware of this technology and more knowledgeable. Regulators have put together some frameworks to regulate it.

Things are a little better. We were definitely too early. But there are other reasons.

There are other ways to achieve this singleness of money without forcing everybody to use the same assets with the same name and the same control. And if you look at the Genius Act in Washington from Congress recently, they actually mandate a plurality of stablecoins and prohibit by law any one of them dominating the market. And the way you achieve singleness among multiple stablecoins is not unsimilar to Libra, but it's not Libra.

The last thing is the elephant in the room. There's no question that Libra itself is a project that scared Washington. And there are people with almost visceral reaction against it because of Facebook.

And I don't know that that would be different today. So if Facebook or one of these large companies try to do something like this today, they might meet opposition in Washington just as much as we did. Do you think it would be... I mean, maybe we don't need to dwell on it too long because there's other questions.

Like, had it been a bit more stealthy and not attracting so much attention about the partnerships, the consortium and stuff, you think it would have been easier to make headways before getting so much of the negative reaction? You know, hypothetical questions are really difficult to answer. But what we do know is that while we were... While Libra invested probably close to a billion, I'm just throwing a number, a billion dollars into the development of the project and four years of some of the best people in the field's time. While we were doing that, you know, Tether, Coinbase, Axios, Gemini, all of these stable coins continued operating without asking permission, without license, putting all of the regulatory frameworks and preparing for the time when they will be regulated in a way that protects both consumers and the stability of the coin in a very admirable way, and asking regulators, asking Washington to come with the frameworks that are needed to supervise them.

But they did not ask permission to operate. They did it on other blockchains, on existing blockchains. We know that that's happened and we know that they grew to be multi-billion dollar, trillion dollar economies themselves.

That was the, particularly when it comes to stable coins, the irony of regulation came through with Circle and Tether, and everyone was quite worried about Tether being the more unstable one. And then when was that? Was that 2023, the Circle DPEG? And it was like, oh my goodness, the one that's pursued the most forthright regulation is the one that has had the most systemic risk. And everyone's like, oh boy, we did not expect this.

The DPEG was temporary though, because it was dependent on the withdrawal from banking over the weekend, if you remember. I actually know quite a few people that were like, okay, this is going to come back. I'm buying USDC at a discount.

It's just free money here. But yeah, I agree. Basically, Tether was always the one that people fought about.

And we came from Circle. Now it's kind of a part of the system almost. I mean, I have to say, I'm going to regret this, but the DPEG in 2023 happened because of a run on the banks, the Federal Reserve member banks, not because of a run on the stable coins.

But obviously, if you have your treasury held in a bank that goes bankrupt or collapses, what do you do? It was a direct correlation with SVB and that entire sector as well. It wasn't necessarily, I do not mean to put the onus on Circle, like, ha ha, those guys tried to get regulated and look what happened. This was a very stressful weekend for a lot of people.

Oh, yeah. I was, I was luckily in quite a hostile, like, hostile assault by the previous administration. I think it's like, yeah, an intense culmination of hostility toward the industry at that point.

That was, yeah. Anyone who thinks that our current traditional financial system is perfect could actually look at that event as well as many others, of course. This was a classical run on the bank.

It has nothing to do with the stable coin issuer. Moving to some more positive slanted topic. Dalia, you and I spoke probably for the first time when the Blue Paper came out.

And this was nearly over a year ago, or very, very close to a year. It's around about that time. Um, I was incredibly interested in space computer from the early days.

It's something that kind of stuck out to me as one of the coolest culminations of, you know, frontier space tech, and then also the original kind of like cypherpunk ethos of crypto, and it's blending the two. So it was taking all the stuff that I used to read in like sci-fi as a kid, and then bringing in the stuff that I've learned working in kind of crypto blockchain as I've gotten older. Talk to me more about the like research and development work that you've been doing at space computer.

I'd love to hear more about what's happened in the time since we've last spoken. Certainly. So space computer, you know, is a new point, is a new kid in the block of trusted computing space.

It takes trusted computing platforms to a level that was never explored before. Um, and the question we're grappling with are what are the security guarantees that you can get by operating workloads in space, and how do you build, you know, distributed services operating in such a manner? So let me talk about the first part. So what is different and what is unique about the platform of, you know, computing on satellites? So a number of things.

First of all, once you securely deploy workload onto a satellite, we call it a box. So you have this box that you deploy software on it, and then it can run your workload. Once that process has been secured and you haven't been hacked along the way, that platform becomes virtually tamper proof.

I mean, except for science fiction, we cannot envision anybody physically tampering with that hardware. So you don't need to rely on any enforcement that is being built. It really is tamper proof.

It is also essentially leak proof. Anything that's not running on satellites, no admin, nobody in control of the system, no state level actor can come with a subpoena and grab, you know, your disks or your systems and scrape bits off of it. It is leak proof also for any type of side channel attacks.

There's nobody who can listen and get physically into proximity of the satellite. It can provide jam proofing. So communicating with a satellite is guaranteed to be jam proof.

So long as you can secure a relatively small area, so a few kilometers, we always joke that Daniel's farm, you know, so long as nobody can penetrate it, we can guarantee communicating with a satellite from that. So long as you can secure a physical, small area on earth, you guarantee that nobody can jam communication to the satellites or receiving transmissions from the satellites. And you can obviously have multiple base stations on earth.

And so long as one of them has a connection to satellites, then you guarantee against any denial of service attacks. Geo positioning certification in both directions is another thing that you can guarantee with satellites that are really difficult or virtually impossible to guarantee on earth. If you want to know that somebody is running your workload on a satellite, this is something that you can actually measure because we know when satellites are in orbit in a particular place and we can communicate with them at predictable times of the day from predictable locations.

So there are a number of these properties that are either impossible or very costly or very difficult to secure on earth. And we get much more easily and perfectly on satellites. There are also downsides.

It's a nascent technology. Clearly, there are challenges. There are control stations on earth.

There are players who are not as trusted yet and don't have the history of today's clouds or anywhere else where people currently deploy their secure workloads. But the vision is that the area will mature and develop very rapidly. And the unique potential properties that I listed are the ones that we're looking to exploit.

And we believe that some of the difficulties and the challenges are things that we'll overcome with time, just like any nascent and any new technology. I would love to know more about the, and this is something I spoke with Daniel about as well at length, but I'd like to know more about what the space economy looks like in future. I know that we're going back to hypothetical scenarios here, but I would love to know what the vision of running trusted execution environments in outer space, having a blockchain that's like permissionlessly running also in orbit, and how that works in with the kind of future economic model of what's going to happen out there once business and vendors start moving off earth.

So I think the key thing is that we envision that for the foreseeable future, a lot of things will still operate off of infrastructure here on earth. There'll be higher capacity, lower latency, maybe even higher bandwidth. And so what you optimize for is really the minimal trustworthy core to run in space and everything else to hinge and anchor with it.

We also laid out a two-tier architecture in the blue paper with some architectural design approaches to do that. And a lot of our research, a lot of our effort goes into designing minimal trusted compute base to run on satellites and allowing everybody to enjoy it. You're also asking specifically on things that will run natively in orbit.

We're not going to be alone there. We see an explosion of activity in the number of launches and the number of companies that are actually operating various services on these satellites. And so we envision that many of the services will not actually have to go back to earth, but will serve each other in orbit, securing each other, attacking each other potentially, and monetizing the services that they provide each other.

And they will need a platform to mediate and manage and orchestrate all of these interactions. Space native settlement layer. Maybe a thought that I have in recent days and I think is somewhat interesting to think from a space computer angle.

So the way that any blockchains are architected today, whether it's Bitcoin, Ethereum, or pretty much any other network, it's a distributed system. It's a distributed network that is stacked on top of the existing internet. So it's actually an overlay network, right? It's a network atop of a network that then has, on the one hand, the native permissionless aspect to it, but then many other choke points, whether it's the internet service providers, whether it's the local governments that can implement firewalls.

Now, with the move to introduce direct to satellite communication from all modern devices, can we look at space computer, the L1 in orbit especially, as something that ultimately can be completely bypassed? You don't need to worry about the existing internet and the overlay network. On top of it, you can actually communicate directly from users on earth onto satellites that then serve as the settlement layer as a completely independent from the current frame of how we build blockchains. Is that something that you reckon is already ready to start thinking of, or is it like several steps in the more visionary future? Yes and yes.

I mean, it's ready in the sense that if we really think about what a system like this, which is not an incremental project. It's really a moonshot. If we think of what it is capable of today, we're thinking about it the wrong way.

One of the stories I sometimes tell when I present the project is the story of how satellite coverage of earth actually started. So let me spend a minute reminiscing about this. Take you back to the 70s when President Eisenhower decided that satellites can fill a gap in military reconnaissance.

You can take photos of places on earth for military purposes that would be too risky or too far to reach with normal surveillance with military airplanes. They launched this project, which had acronym CORONA, funnily, and this was late 60s, early 70s. They managed to get satellites to cover the entire area, the entire surface of earth 24 7, which was really a big deal.

But they had a problem. No digital cryptography and nothing even close to it. So how do you get all this information from satellites back to earth? I mean, radio communication did not have the necessary bandwidth, still barely doesn't, but back in those days, it was very, very far off.

So pause the history lesson and let's think about today. If you sit back today and say, oh, you know, satellite communication is very tough. The devices, the handheld devices with users are very far off from being able to really connect with the satellite to the settlement layer.

Well, if the engineers of CORONA said the same thing, the project wouldn't have existed and maybe we wouldn't have had satellites today. Instead, they devised a process by which they dropped buckets from satellites to earth, carrying undeveloped film, physical film that took photos on the satellite and dropped them in buckets to earth. And these buckets were intercepted midair by airplanes that were circling around the Pacific and hundreds of them were intercepted over a course of about a decade since the launch of the project and until the early days of digital photography started emerging.

During this decade, they captured very valuable photos of very interesting phenomena that are used even today by scientists for research purposes. So compared with what they worked with back in the 70s and succeeded, not a single bucket was missed. What we work with today with a space computer and just generally satellite technology is a luxury.

And like most technologies, there's friction initially. It's difficult. It's slow.

It's clunky. It's unreliable. There's no, the market is not ready for it.

The use cases are not ready for it. But 10 years from now, does anybody have any doubt that communicating with satellites will be just as easy as today? You know, I can send SOS messages from my phone and satellite to satellite communication will be much easier. And there'll be a whole, I don't know, a whole market and a whole action taking place in orbit.

I don't have a doubt. The only question is how quickly between now and 10 years this will happen. And so we're building and we want to be pioneers in this space.

We want to be ready with cool technology and contribute to the field and, you know, have impact in that space. Pun intended. That's incredible.

I had no idea that there were buckets, physical buckets that they were dropping. I can send you some images of that, some simulations from documentary films. And some original photos that were taken over the Pacific.

That's incredible. I might include some B-roll in the footage of this podcast to run the outro. We'll leave the viewers with some, as much imagery as we can as the buckets from space.

That's incredible. So Dahlia, zooming kind of like back out again, moving from space computer and looking at, you know, previous roles that you've occupied across the industry. And this is something that was given to me was that I understand that you rejected co-foundership in what later became Checkpoint, which is now a publicly listed kind of cybersecurity company, and along with like several other projects.

Can you walk me through why you've been so selective with co-foundership on projects? Okay. I don't think I was offered to be a co-founder, but I was probably offered to be one of the first 10 co-founders, you know, party team, just to be precise. But yeah, I like saying no.

Why? I have three principles that maybe I wasn't quite aware of them, but in hindsight guided me throughout my career. Learn, innovate, and have a fun team to work with. And, you know, really, I think our brains like to learn.

I don't like going to Chattopadhyay and just getting the answer. I really want to understand and feel like something clicked in my brain. I think physiologically our brain is built to like that, to enjoy that.

I like to innovate. I like to solve puzzles. I think every preschooler that you go to likes riddles, likes puzzles, likes the challenge.

Somehow, some people, you know, sometimes we lose it in the course of life. But again, this is something that we're designed physiologically to like. And at the end of the day, those of us who are lucky to have a choice of how to earn money, I want to do it with a team that I really enjoy working with and have fun.

This has guided me. And none of these three principles ever led me towards the founding a new company track. So I remain in a place where I can learn, innovate, and really enjoy the fun part of working with the most amazing team.

Dahlia, thank you so much for joining us on the first episode of the Frontier Podcast. This has been fantastic. It's been an honour and a pleasure.

---

Join the community on Telegram.
Subscribe to The Frontier Podcast on Spotify or YouTube.