Wen Escape Velocity, Wen Ossify by Venkatesh Rao & Danny Ryan

Transcript

That was really great. Could not agree more. Next up, we've got Danny and Venkat talking about when escape velocity, when ossify. Welcome. All right. When escape velocity, when ossification. Uh I'll tell you what, let's uh start with a quick introduction. So, Danny Yeah. I'm Danny Ryan. I work at the Ethereum Foundation on the research team. I work very closely with the protocol support team as well. So, I do some research. I do a lot of specification writing. I do a lot of coordination around major infrastructure upgrades like uh the merge and uh coming coming upgrades

from beyond. And I'm Venkatesh Rao. I'm a blogger at Ribbon Farm. And I've been working with the EF for the last few months on some stuff we'll actually talk about in um few minutes. Uh so, to kick us off, Danny, um at Devcon a few months back, uh right after the merge, you showed a bunch of really boring graphs, like the most boring graphs I've ever seen in a talk. And you were excited about the graphs being so boring. All of the audience was. And it was great that it was a non-event. So, now what, 6 months later, what's in delayed assessment?

Yeah, so the merge was a non-event. Um I mean, it was a spectacular event, but if you look at those graphs, very flatline because things just worked. Um I I believe if you look at those graphs, things have still worked. You know, we're finalizing at a regular clip. Uh we have like 99.9% of attestations and blocks all the time. So, things have gone well. I mean, the staking ecosystem, the node ecosystem, and everything kind of around it continues to evolve um as they talked about in the last panel. But, uh steady state. Things are good. Things are primed to keep keep moving.

All right, so it continues to be a non-event and that I think that's what we all want to hear. So, what's on deck now? Like we just talked about like withdrawals and 4844 is on the horizon. So, what are the main things people should be looking at and what should they be paying attention to in the short term? Yeah, so often when I think about protocol upgrades, I think about you know, we need to make things sustainable, secure, scalable while still being decentralized. The merge helped knock out a couple of those, made the protocol more sustainable. It made it more secure. So, we're certainly keeping our eyes towards withdrawals, which are going to kind of complete the loop and complete this mechanism. Like you not only can you jump in, but you can jump out. Uh but the big thing is

really scalability. Um it's been the big thing for years, so it's that third S. Uh 4844, which adds blob data transactions, which is just a fancy word for bringing more scalable data availability to the blockchain in kind of like an initial step. Um that's that's huge. We're certainly looking like semi-production type testnets and and getting feature complete and really ironing out specifications. Um it's been in parallel to the the upgrades coming, so very excited to to bring some scale.

Beyond that, there's a lot of like iterative improvements, a lot of um continued adjustments and security improvements to proof of stake, but a lot of good stuff, but initial scalability is the big one. Okay, and um I guess the short short-term stuff that we're getting into now, it's the start of um the big like roadmap which Vitalik laid out over the long term with his like, you know, purge, surge. I can never keep track of like how many rhyming words there are. Uh but you just published a very interesting reflections essay on so, where you go from these short-term priorities to the longer-term roadmaps.

Uh can you speak to the themes a little bit on that? Yeah, so there's certainly some technical stuff in there. So, you know, how do we go from 4844 to full sharding to a much more scalable landscape? There's some technical work to do. There's a lot of iterative improvements to security that I think are are very important on the, you know, 3-5 year time horizon. But a lot of a lot of what the essay ends up being about is my concerns about all the things that we have to do, but also in this in this very dynamic environment where there's a lot of different interests, there's a lot of different pressures, and just how we can continue to uh get the protocol to where it needs to

be to be sufficient for its purpose uh without kind of crumbling under the weight of all the different uh I don't know, pressures. So. And you came up with I think a very interesting categorization of protocols as uh being sufficient in a number of different dimensions, and connecting that to your idea of like the set of concerns you have about the roadmap. So, what are the main kind of like sufficiency things that you have your eye on for the protocol? Like, what does it have to be sufficient on in the next Right. I mean, certainly sufficiently secure.

We need enough crypto-economic security to support this thing. Sufficiently scalable. That's obvious. We need to be able to have enough capacity in this or subsidiary environments like layer twos to be able to support global demand. And sufficient expressivity, you know, in in the applications that you can design here. So, one of the big questions, especially as we continue to have more and more EVM improvement proposals, like people want to tweak the EVM, extend the EVM, modify the EVM. One of the questions I'm always asking myself is, but is it sufficient? Right?

Does the EVM do what it needs to do to be the solid base layer of you know the Ethereum blockchain to then build the applications and the layer twos that have the expressivity on top. So, um that's definitely a a big question to me is like what is sufficiently expressive? Security and scalability I think are a bit more obvious to me, whereas expressivity is um something to be debated for sure. So, I guess the sufficiency criteria some of them are like more hard-edged and maybe hard or easy to actually build and there's some things that are fundamentally more fuzzy even to conceptualize whether or not they're hard to execute and sounds like expressivity is one of those considerations that's like kind of

fundamentally fuzzy. Yeah, it's a bit fuzzy. I mean, obviously you make something Turing complete and you say, "Okay, it's expressive enough cuz it can express everything." But can it express things optimally? Can it express things in a way that is probably safe? Can it express things where we can use formal verification and auditing and and other things, you know, like that is less clear to me, right? Like right now you can't do a lot of like pre-stack validation on the EVM. And so, you might say, "Okay, well, if it's going to be a smart contract platform where we're going to have billions of dollars behind this these code, we need uh more static analysis and stack validation." So, it's like it's not sufficient. And but you could also argue, "Uh but it works."

And I guess we're going to be having some of these debates for like years rather than months and some of these things it'll take like a decade maybe to figure out what sufficient actually means on some criteria, right? So, Certainly. And the criteria keep changing, right? Like if you asked everyone 3 4 years ago, uh their definition of sufficiency might not have had, you know, that it was sufficiently rich to be able to run the zero-knowledge proofs to run ZK-EVMs and and other and other VMs, whereas now it's like that seems like very critical and you know, we should be evaluating if it's a sufficient as such.

And I think part of this the changing landscape I guess is backstage when we were talking you mentioned that you were a little naive and felt a little surprised when the merge was done and suddenly now you have this whole new ecosystem of players doing like you know, withdrawals and staking that whole corner of this ecosystem has emerged and that's part of why I guess the criteria keep changing because now more people have been activated to be part of the conversation in a sense.

Yeah, I mean I I I'm continually surprised by what gets built upon this platform that we continue to work on down here and you know, sometimes those surprises are like, oh, okay, cool and the platform is sufficient. Sometimes those surprises are more fundamental, you know, like MEV. MEV not that it was a surprise to everyone, you know, some people have been thinking about this for years, but you know, it turned a lot of our thoughts about security and protocol design on their heads. So, surprises, we got to got to stay on our toes. So, speaking of surprises, there are I guess some good surprises, but what are the things that you think could go wrong in the near future?

Yeah, good question. to watch out for. I mean, I if you back to the essay, like if I if you look at my recent reflections, I'm deeply concerned about capture. You know, the the Ethereum protocol is not done. I don't know if the Ethereum protocol can ever be done, but in the meantime, at least until we get to such a state, there's increasingly huge interest involved. And so I I worry about this messy process of governance that upgrade this incredibly valuable thing is is right for capture. You know, like if if you have enough money on the line, you know, this is a thing to manipulate to your interests. So, that's something I'm deeply concerned about.

I don't know what else can go wrong. So, I guess the flip side of what can go wrong is what can people do to make it go right? And I think you're one of the people in the ecosystem who's earned the right to like deliver sermons. So, if you look around at the L2 world and the world of application developers, what sort of the if you could do like a 2-minute mini sermon on what they should be doing. we spend a lot of time and this is this is part of what I was talking about at Devcon, but we spend a lot of time at the base layer like making sure this thing works and making sure this thing is decentralized and making sure this

thing is is open and accessible. Um and quite frankly, if you look at a lot of applications and you look at it most of the L2s, it's a a landscape of cutting corners to go to market. You know, it's a landscape of throwing uh poorly thought token models on top of things to like get the liquidity and get your exit. Um and unfortunately, like we can have the strongest foundation as we want, but if we have like little brittle Legos that we're stacking on top of each other on top, then it it's all for naught. So, you know, I challenge the the application developers and I I challenge uh layer two developers for sure to uh to get serious. You know, we're we're serious. I get it. I get it. You know,

it's it's always a balance between, you know, the runway. It's a balance between user acquisition and all sorts of stuff, but it, you know, if we're if we're architecting re-architecting for the future, then um you know, we have to take application design uh as as seriously as we take in the layer one design. And I guess one of the things that will happen if they do follow your advice is yeah, blockchains can start fulfilling their potential in the broader world. And we had a quick announcement to make. Danny and I and a bunch of others, we've been doing this little project which we call Summer of Protocols. So if you go to summerofprotocols.com, you'll see a program that we're booting

up and as a pilot study. And the intent there is to start a broader technical conversation that's beyond crypto and maybe even beyond the tech world of Silicon Valley and get like I don't know the whole all of society to start thinking about protocols as one of the things in the world to work with, right? So Danny, you've been heads down with the merge and like deep technical core Ethereum stuff and you've been able to come up for air after a long time and like look at the rest of the world, right? So what strikes you about the rest of the world today that crypto should be paying attention to it and Ethereum in particular.

Right. So I don't know. When I when I go to an event like this, I look around and I think this doesn't look like the rest of the world. Like fundamentally, you know, the way that these hackathons work, the way that companies are built, the way that protocols are built on top of this protocol looks very fundamentally different than big tech, than traditional companies, you know, Silicon Valley, etc. And so one of the things that in our discussion of protocols and looking at this program Summer of Protocols is uh you know, it's it's kind of like the protocol first nature of of Ethereum.

You know, if you look at Silicon Valley, protocols are usually like incidental. Like they're they're a means to an end. They're like, "Okay, I'm constructing this way to connect my device, the TV and other things. Okay, we should probably have a protocol so that people can conform to using my product." Instead of there are like let's architect like fundamental protocols to be able to like interact, coordinate, and solve problems and then architect on top of that. And I I I I think if you look at things like ETHDenver, if you look at a lot of like the Ethereum ethos, uh a lot of that comes from being protocol first. And so, you know, I I challenge the rest of the world to treat protocols as first-class citizens um and and layer from there.

is there a particular example that um you think really illustrates the difference between like traditional tech in its sort of like marginal attention to protocols as like an afterthought almost as a means to an end as you put it versus a protocol first. Like what's a good example here? I might flip that to you. You've been thinking about protocols. What's What's the example? Uh I have a very I'm not like a tech guy in this world, but so but for me it's a very retail experience of Last night in the hotel room I used Chromecast for the first time. I was like being able to take my Netflix account and just have it show up on my hotel screen. And it struck me that it's really impressive that a big platform company like Google

is able to like put this protocol out there and get all streaming services and hotels to sign up to actually use this way to watch TV, right? And I don't know if it were a crypto world, how would that look? Like what would Chromecast look like in a protocol first world where protocols rule the earth? Right. We We would architect the protocol such that devices agnostically communicate with each other beautifully, you know, at a at a click of a button. And And, you know, Google becomes peripheral. Google becomes like one of the many products that's layered onto it. It becomes like a much more like reasonable landscape of competition of ideas um because there's a common protocol to build upon. Whereas, you

know, your Chromecast, like you're now in like Google's little walled garden moat to do whatever they want and like what they want is for people to adopt their proprietary protocol in many places as possible so that people are kind of locked in to continue to use the suite. Whereas if you start protocol first, you know, everything that chooses to opt into that protocol now gets to live within that protocol, gets to architect within that protocol, gets to compete within that protocol. And you as a user get to be like very fluid in using in what you use and the decisions you make instead of being locked in. So So the way you're describing it sounds like you would be very sympathetic to the people who think of like protocols as public goods and you know,

provisioning them like public goods, streets, national park. Is that how we should think of them or Yeah, I think so. I mean, I think if you start with protocols they're naturally open and public goods. If you architect a product and then shoehorn a protocol into it, usually it's just a means to an end that, you know, helps facilitate their walled garden. It reminds me of like the like a national park versus like a nice big open-air mall with lots of landscaping and gardening but fundamentally a mall, right? We have about a minute left so as a closing question like do you think there's some fundamental values differences in the conversation around technology as it currently exists in the

mainstream? So Silicon Valley, venture capital, the technology conversation that happens there versus the technology conversation that we are trying to like catalyze here with the summer of protocols thing we're starting. But yeah, in general, what are the axiomatic differences you you would point out? Quite frankly, I'm not in those Silicon Valley conversations very much so my my view is very biased. I mean, but the um kind of like collaborative first seems like the native crypto way and seems like the native protocol way. Kind of like creating new positive sum games Um um through the construction of protocols and layers of these protocols, um whereas

my intuition of the other game, the other side of that coin is it's it's much more winner take all at the expense of anything. Right. Um we're just at time, so thanks very much, Danny, and thank you all for listening. Thanks, guys.