The Third Web #5 - Brian Behlendorf

On this episode of The Third Web, we hear from Brian Behlendorf, the head of the Hyperledger project at the Linux Foundation, and leading figure in the open-source software movement.

Brian began developing open-source code during the mid-1990's and was also the lead developer of Apache Web Server — one of the worlds most used pieces of software.

Today, in the midst of a blockchain boom, Brian’s focus is still on free software development. I spoke with Brian in October. We were both in Cancun for Devcon 3.

Arthur Falls: Welcome to The Third Web. I'm Arthur Falls. 

On this episode of The Third Web, we hear from Brian Behlendorf, who leads the Hyperledger Project at the Linux Foundation. Brian first began publishing open source software back in the late '90s. Brian was the lead developer for the Apache Web Server. And it was the vision of an open internet that inspired him to open source the code, leading to Apache becoming the most broadly used piece of software in the world. 

Today, in the midst of the Blockchain boom, Brian's focus hasn't changed. I spoke with Brian in October. We were both in Cancun for Defcon 3. He told me that in 2015, he joined the Linux Foundation, a position that would lead to his role in the Hyperledger Project.

Brian: Cool. All right. So, I'm Brian Behlendorf. I am Executive Director of Hyperledger, an umbrella effort to develop Blockchain technology, products, and services that are all open source in nature. To try to build different building blocks, particularly around the permissioned ledger side, but not exclusively limited to that. So, focusing on things like: What are the consensus mechanisms that we might wanna use? What are smart contract languages that we might wanna use. 

So, our premise is to be somewhat like how the Linux Foundation is operated, somewhat how the Apache Software Foundation is operated, which is really focused on governance around these projects, and try to bring the businesses in this space, building projects and services on top of these underlying technologies. Bring them into the picture as well. 

Arthur Falls: So, when will there be the Hyperledger ICO?

Brian: There will never be an Hyperledger ICO. We're actually pretty determined to say, "These are underlying building blocks. And the way that they get built, the way that they get funded is not actually up to us." 

Our business model, like the rest of the Linux Foundation, is anybody can use this code. It's all open source. And anybody, without paying us a dime, can participate in development. And we certainly hope that, when they find a bug they report it. When they have a fix for it, they bring that in. If they want to extend in some useful way, they think about contributing that upstream. 

But there are companies who want to get closer to the project, understand how it works. They wanna see what our priorities are, what are roadmap is. And, they also want help in explaining what they do to the broader market. 

So, that's the value that we provide to the 170 different organizations that have joined as members of Hyperledger. Many of those are also contributing code back into the project. 

Arthur Falls: So why is that approach, in your mind, coming from your background, preferable to the ICO operation?

Brian: Well, first off, there's nothing wrong with tokenization. Right? 

Let me back up a bit. So, the use cases that drove me to say, "I need to spend more time paying attention to this is", I think as technologists we got really good at building centralized systems, at being able to build something at the scale of Twitter, at the scale of Facebook. Something that could handle millions of transactions per second, but required centralizing on one company in the middle, which is ironically very different from how the internet came about. The internet was not just one central company providing a central single network. The internet was actually lots of disparate networks with different technologies coming together with a common routing protocol and a common domain name system.

So, I've gotten really worried about the over centralization in the technology space and the internet space, where everybody's email was going through Gmail, or everybody's social was going through Facebook, et cetera. And felt like there had to be an answer. There has to be someway to decentralize this usefully. Right?

I read Satoshi's paper in 2008, or at least the abstract to it, and said, "This is a good idea." But, I really did not like the energy consumption implied in proof of work. Burning all the CPU power in the world to run a lottery? To decide who gets to put the next block in a chain seemed like an incredible waste of resources, just morally to me. Right? And I think for a lot of other people. 

And the second thing was, it seemed like it was something that would be very vulnerable to inadvertently taking the enthusiasm that developers normally have for their own products and services, and turning that into pump and dump schemes. And, I was concerned then, about could you inadvertently cause a lot sincere developers, building something brand new, into hocking a financial investment product? And if they do it slightly wrong, ending up in jail. 

I was concerned about that in 2008 when I saw that thing. And so, that's why, much to my bank balance's chagrin, I ignored that space. And I still feel like currency is they underlying basis for these things, is the same kind of criticism people have for how unrestricted capitalism has led to a lot of the structural inequalities that we have in this world. 

So, anyways, long story short, this is why I've been attracted to the idea of other alternative approaches to doing consensus, to doing smart contract languages that don't require a financial instrument at its underlying core. Most of the time, when we do commerce, we actually know who we're doing commerce with. And, if we're doing commerce in a marketplace, the marketplace has rules about how to enter and what you agree to the participate them at. And, that's much more the common case, and that's totally in line with the concept of permission ledgers.

It's really about this idea of how do we get the best of these ideas that are out there, whether from the Ethereum community or elsewhere, kind of explored and built in a way that ... and most importantly, that these technologies become reference standards and they become built by multiple organizations, so you never have to worry what happens if a certain developer, or a certain company is hit by a bus? Like, today, with Linux, you don't have to worry as much. I mean, it would be a horrible tragedy. But, if Linus Torvalds got hit by a bus, there's now a broad community that would take it up, and a governance model around how that would work. 

That's really what the Linux Foundation has provided for the Linux Operating System, a way to bring the developers together with a governance model, as well as to bring the corporate world along, as well. So, there was a time when all of these companies were thinking about, "How do we get involved with Linux?" And many of them were thinking, "How do we become the Linux company?" Right? "How do we own that technology landscape?" 

And, you need a strong organization at the center of that ecosystem, to help keep those interests balanced, and say, "No, here's the rules for how you use the brand, how you talk about the platform, that sort of thing." 

And, the Linux Foundation went and replicated that to other domains, cloud computing, software defined networking, et cetera. And, when the topic came up of doing that for the Blockchain space, instead of saying, "Well we'll become the Ethereum project," or anything like that, 'cause that already existed, the idea was, let's look at one layer lower, at these enabling technologies that make those kinds of things work. And let's try to be an answer in the same way that Linux now runs on the enterprise. A lot of these technology projects are really geared towards how do we get companies to adopt them and incorporate them into our products and services, so that it becomes the new default. So that, cars can run on Linux, and phones can run on Linux, and all of these other things. Why don't we see about doing that for these underlying concepts around distributed ledgers and smart contracts? Around which, currencies are one application, but not the only potential application. 

Arthur Falls: So, there are two different trust graphs here, right? You have an individual or an ecosystem participant trusting this decentralized structure that is managing assets they might be storing on one of these crypto ledgers. But with Hyperledger, or with Consortium ledgers, you have a different set of assumptions. You don't have distributed consensus. You have a kind of a decentralized consensus where you choose the actors who are going to come to consensus. And, it's about whether or not you trust those actors to maintain that ledger. 

Brian: Sure. But ... There's a big difference between trusting that a bank holds your ... When you have an account at a bank, it's technically possible you could wake up tomorrow and discover that bank forgot you had an account there, right? In essence, they play God. And there's all sorts of regulations and rules that they have to follow. But, they play God in that kind of system. Right? Or, if you have somebody like an eBay or an Uber, or a PayPal, who sit at the center of their markets, and everybody really has to route through them, you kind of have to trust them to be God. Right? 

But, think about, instead, a referee on a football field. If the referee is the one who has a lot of power, they can kick players out. But, they're not playing the game themselves. The game is still between the players. And frankly, if a ref shows themselves to be corrupt, they can get kicked out by the team. Right? 

So, I think a lot of these permission ledgers are not about presenting themselves as a preferable alternative to the public cryptocurrency applications, but instead, to say there's a lot of existing applications today where that are much more centralized, that are much more involved these hubs, or these central market actors, who have to much power and you can decentralize that by changing the role of these central actors. Not saying do away with all intermediaries, 'cause frankly, there's a lot of intermediaries out there people like. But instead, changing the role from being one where they have a lot of leverage and can extract rent, and set the terms and play God, destroy our assets if they so choose, into ones where, they're there to help enforce the rules. They're there to guide the community around evolving those rules, should there be a new set of rules. 

But there will always be rules that you cannot script. Things like, if we're using a network to share health data, we're not gonna write health data directly on a ledger. We're gonna write it off chain. But, we're gonna have keys and signatures and hashes on chain so that I can share records with you. And there's always the risk that there's a breakdown in the crypto and we inadvertently reveal somebody's HIV test results to somebody else in the network. And we need a legal agreement between the parties on the network, that ensures that in that case, when there's a failure like that, the response to that is something that no smart contract can do, which is, compel the other parties not to share that data onward.

There's no magic DRM, even with a Blockchain tech, that keeps somebody in possession of data from being able to make a copy of it. There's other things that you can't enforce through smart contracts. So, having a governance mechanism around that, that is a mix of algorithmic governance, like prevention of double spend, with human governance, which is, we all collectively agree to abide by these rules, and if somebody violates that, they're fined. If somebody does this, they're required not to share data, on penalty of fines or whatever, or they get kicked out of that network, those aren't rules you can enforce on a public network.

So, the world will be a hybrid. They'll be the public chains for much the same reason we have public newspapers and public currencies. But most of the interesting transactions will happen on private ledgers. 

Arthur Falls: And that is because of the actual ... the relationship between the parties involved in this. That's because of those trust [crosstalk 00:10:47]

Brian: And the governance mechanism around that. And, who is that governance organization? It'll generally be a nonprofit, a consortium, somebody accountable to their interest. And the final thing I'll say about trust is, if I'm forced to trust you because I have no other option, that's not really a choice. But if I have to vest a lot of trust in you, and I know I have other options, you're gonna be much more accountable to me. So the reason why banks don't just steal money from people willy-nilly, is not just that they're regulated, but because they know if they have a bad reputation like that, there's other banks.

And so, hopefully what we'll see is not just a permission ledger for this sector, for that sector, et cetera. But, competition between these ledgers within any given sector. So that, if the governance organization, the consortium around that, starts to really go down hill, and make unreasonable rules and start kicking people out for arbitrary reasons, the actors in those networks can move. And they can bring with them their transaction histories, which is not something you get to today when you wanna quit Uber and move to Lyft. 

Arthur Falls: No. How do you foresee these changes expressing in the structure of the internet and the nature of the services that are available to us through the internet?

Brian: You might remember that famous comic, "Nobody knows you're a dog in the internet." And it had like a picture of a dog sitting in front of a computer. Right? We've had a really challenging time with the concept of truth on the internet. The concept of, the same kind of longterm histories that we develop as people in the real world. It's actually rather hard for people to change their identities. They do it. And the fact that they can do it is a really good and useful thing. I'm thinking about people who ... transgender, et cetera. Right? 

But, generally speaking, society works because people develop histories. Because we get to know each other as individuals, and there are objective truths about the world that we can attest to. You know? A country being founded on a certain date. Or, a certain distance from one city to another. Everything is much more malleable online. And that's terrific for a number of good things. It's great that you can pretend to be three different people online and explore what it's like to be three different people. It's great that you can whistle blow anonymously on companies that are violating the public trust. It's great that you can publish, on a blog, anything that you wanna publish. 

But we're having ... especially the recent election in the US, showed that we are really suffering from almost the extremes of that, which is, people don't know what they can trust online. They don't know what verifiable truth is. They don't know who people are. 

And, this is finally an answer to some of that. And beyond programmable money, I like the idea of smart contracts. I like the idea of tokenizing a lot of different things out there. We will see nations issuing fiat currencies as digital tokens, which will eliminate a lot of the practical advantages of Bitcoin or Ethereum, 'cause you'll be able to have programmable money with US dollars, which people would ... I think most of the world, most of the time, prefer to deal with. 

The bigger impact, though, will come from being able to interrogate the veracity of what you see online. Not just as consumers, but also as businesses, right? To be able to do business with somebody and know that you're actually dealing with a company that is what they claim to be. Or, even if you've never met them before, find a verifiable history of their transactions with other companies so you know that this is not a fly-by-night operation. This is somebody real. And if I send them money, they actually will send me the goods that they promised me. Right? 

And that's going to matter. I mean, we still live in the real world. This bridge between the digital world and the real world needs to be really solid. And I think these are technologies that help with that. And I think the underlying protocols of the internet, I don't see this replacing TCPIP, or HTTP, or even the domain name system. I see this complementing, as all new technologies really have. 

Arthur Falls: What do you feel is characterizes people who have gotten to the Blockchain space? There seem to be different camps. 

Brian: I don't see this space as all that different than the way that other technology revolutions have emerged. I mean, in the early days of the web, you had ... you have the pioneer phase. The people who were working on this for the sake of the technology, see what's possible. Right? I think that same kind of [technophylic 00:14:42] kind of interest drives a lot of early interest. 

And then you kind of have the settler phase, where you have kind of like the gold rush in California? People come out 'cause they're attracted by these promises of riches. Right? And I think that is characterizing a lot of the people in this industry, now. And I would be cautious about building a vision of a future society based off of that kind of level of interest.

And then, I think there's a third way, which is almost more normalization. If that second phase is characterized by the 1999 or 2000 bubble, the phase after, where everyone ran away from the technology for awhile, and it was hard to get funding for a startup. It was hard to get funding to try new ideas, well, many of the great companies we know today, Google, and Amazon, and to some degree, even like the resurgence of Apple, and the rebuild of Microsoft, happened at these times when the amount of chaotic energy kind of subsided a bit, and a lot of other startup companies came about at that time, or in the aftermath of that, and a lot of good technologies, too. Cloud computing came out of this aftermath where people didn't have the funds to buy their own servers anymore. They wanted to rent capacities somewhere else. Right?

So for me, I'm really more driven by what happens in these lean moments. Right? And by this third wave of normalization. I think that's when the most economic value at the end of the day, ends up getting created. And where the most substantial reinvention of society is possible. 

And, all I can do, really, I'm not really a demographer of people who show up at conferences like this, or others. Although, this is more my crowd than, say, where I was last week, which was Sibos, which is the banking software conference. I identify more as a developer. I think there's [devs 00:16:16] who sincerely wanna make a positive impact on that. And maybe cryptocurrencies give us a path to that. My hope is that we don't also, inadvertently, create a way to widen the inequalities out there in the world. And I think there's [devs 00:16:28] that are worried about that. 

Arthur Falls: Because, this is supposed to be a ... the idea is that we wanna minimize the wealth gap that is fast becoming the biggest problem in the world. 

Brian: I think that's important to some, but I don't think the majority of people here are focused on that. It'd be interesting to find out how many people in this space are as concerned about positive social impact and addressing what people see as issues out there in society these days. Is this a path to being able to address the hollowing out of the middle class. Right? Of the extreme wealth discrepancies that we have in society. Right? And this is a root of instability. It's a root of injustice out there. 

I don't see that cryptocurrency technology necessarily affects things one way or another, inherently, as part of the technology. I think we'll still see a scenario where, those with wealth will accumulate more wealth. I'd like to believe that these technologies offer a path to avoid rand seeking behaviors, avoid monopolization of markets. But, they don't guarantee that by any stretch. And, just as we found with the internet, you equalize and liberate access to information at one layer, you create opportunities for new, dominant players the next layer up. So, it's something I think developers here are concerned about and should be thinking about is, how do we ensure that these technologies remain empowering enabling technologies, and don't end up just feeding into a board at the next level up.  


arthur falls