Fireside Chat I Kartik Talwar and Vitalik Buterin I Pragma Cannes 2025

[Music] All right, we are ready for our last session for today. And for this one, I'd like to bring on Vitalik, who is the chief scientist of the Ethereum Foundation and founder of Bitcoin Magazine on stage. [Applause]

Great. So uh no thank you very much for uh you know continuing to do all of these uh lovely yeah lovely panels at Global Cardic. Um so uh so Cardik, what's your favorite animal?

I'm going to have to say a sloth.

Okay. Um what do you think was uh the worst thing about the Robin Hood launch? Uh, good thing I control if the video goes up or not. Um, I think the

the worst thing about the launch was not having clarity on what is actually going to go live. It it's

from everything I was able to catch up on, it felt like it was the announcement of a proof of concept or even functional intent to let you get access to private stocks,

but no proof that it's actually going to happen for the companies you would like.

So, it's tech, but without knowing if it can be applied to the actual goals you would like to. Okay. Yeah. I think uh

we kind of saw that by OpenAI also tweeting about the clarification.

Yes. Yes. It's a a derivative of uh OpenAI stock where uh OpenAI stock itself as pointed out.

Yes.

Yes. It's a proxy of uh intent to get access rights to.

Yeah. Okay. So, one of the things that uh fascinated uh me just because I was in uh Prague, you know, a month and a half ago and then I was in Berlin for a while and now I'm here is like it's interesting to see just like the sheer level of diversity and culture that uh that Ethereum has. Um so, okay, one question for you guys. raise your hand if you've used a privacy protocol in the last 30 days.

Okay,

keep your hands up if you're lying.

Now, so what's interesting is uh when I asked this question in Prague, the uh the the portion of hands raised was even higher. And uh when I asked the same question at a different event that was also at HCC, the number of hands was like five times lower. Um so it's um and then you know you can also ask things that go in the uh in the other direction right like uh you know like who here has uh done a a DeFi thing more complicated than a token swap in the last 30 days.

Okay. Pretty good. Yeah. Um is that not expected behavior though?

Yeah.

This is this will be a polarized audience,

right? It will. Yes. So I guess I mean I wonder from your perspective, right? because you run these uh ETH globals and and ETH pragmas everywhere like what are things that you personally notice that are the differences between like say what you see here what you saw in Prague what you saw in Taipei um elsewhere

yeah very similar trends so the the context behind why we created Pragma was that we want to keep the hackathons for developers only um and that's just because we know how to serve that audience best and the second you kind of figure out how to accommodate everybody it becomes kind of lowest common denominator is the the right outcome. Um so Pragma was the result of saying hey like we know there's a lot more people who want to understand what's happening in the space and we know that group is going to be a much bigger audience than the developers just by sheer probability

but instead of figuring how to shove everybody into the weekend. Let's get people to come in the day before and we'll do a dedicated day where there are no checks for like whether you're a developer or not to be in the audience and we'll talk about ideas and and conversations that way and we'll still keep it slightly smaller so it's it's cozy and and um and interesting but you get a proxy to what people are excited about. um anybody who's applying for uh Pragma as in or buying a ticket uh we get to kind of ask them what do you want to kind of hear as topics or any speakers you would like and we basically fulfill that directly by reaching out to the speakers that are most commonly requested and uh you kind of notice trends and it's never a new thing. It's always the ratios that are off um in terms of like difference. So the biggest thing here which should be no surprise to anybody is there's a lot of AI related talks and it's because this is what people requested.

This is what people are excited about and when they kind of say I would like to know about this topic that's the matchup they look at and this if HCC itself or just being here you have the highest ratio of just people from all countries coming in. So that's the normal distribution of that uh trend. We went to Prague it was very much local people. um it was a very heavy local audience and they cared about privacy a lot more

and we saw that in the hackathon project we saw that in pragma talks you saw that in your other ETH proc talks Taiwan on the other hand was back to AI but for a different reason because all the majority of the participants were students in university and that's all they cared about because that's all they see and that's the immediate thing that they can apply their existing CS education to

um that's trends we see the thing that it sort of answers the question, but the part that it doesn't answer is to me the the part that I think is helpful for you to know is it's the inverse that's more important, which is what I don't see is DeFi being asked for a pragma. You see DeFi becoming more practical for hackathon projects. Uh, and this is a very obvious trend for 2025, but I've not seen demand for DeFi beyond just knowing what's happening for stable coins or how does that kind of overlap for some kind of speaker working on a company that touches stable coins. Um so so to me the question becomes are people not excited or do we think things are done and there's nothing new that you can do so you're just kind of doing some 10% gain related outcome and that's not exciting to people so they move on to the other thing that has a slightly bigger shot of becoming big um but overall it seems like the one pattern that is extremely clear as of not only this year but I would say probably here on Bangkok last year at DevCon is that people do not care about where they're deploying things. Everything has become table stakes.

It is almost equivalent on features and properties.

So they go for the closest connectivity to the other things they're trying to leverage whether it's a composable app or another contract and that's the decision for the destination not any other feature set. Mhm. What are the uh trends that you've noticed and like what kinds of things defiers in the in the hackathons focus on?

They are largely trying to just gify existing consumer behavior. So a lot of them just end up saying hey we'll add the notion of yield as part of an incentive.

Um or we'll kind of touch on making bridging and crosschain application slightly more um

abstracted so you have maybe a less step or you can kind of reduce the latency. Uh that's kind of the biggest pattern. I've not seen any new ideas.

So this is more going after existing crypto users.

This is integration versus innovation of

D5 primitives.

I see it. Um what what do you think can be done to get more innovation? And maybe second question is like do you think we need more innovation?

I think yes on the second question. Um largely because you have a unknown unknown scenario there. Um, so you'll only find out what could be possible unless if you kind of just continue to experiment. You have almost infinite possibilities. Maybe one of these things would work.

And we've kind of seen that because

I mean it was a time where everybody thought AMMs were impractical and extremely poorly executed. And look where we are now. Um and if anything, we've seen probably the most amount of innovation with AMMs getting more and more efficient and tight in terms of uh the spread than other things at all, including in comparison to traditional finance um kind of exchange rates. Um, but I I say that without too much research on this, but I feel like the place we're in, especially with DeFi as a builder, is um at the edge of the limits of like what people would just know without being too deep into finance to realize how you can structure more exot exotic instruments to kind of offer the the cool things that we may not have access to. Uh so it's like a lack of knowledge knowledge limitation but anything that you have access to in the traditional sense traditional finance sense you kind of have all the tools to do that exact same experience in in DeFi.

It may not be as price efficient but it's possible.

Mhm. One thing I um I think is sometimes worth keeping in mind is that uh an idea that failed five years ago may well actually be an excellent idea to do now.

Yeah. And we've seen that cycle happen and it becomes possible and it's one of those things where there's no such thing as a bad idea. It's just bad timing.

Um

what

we're seeing that a lot.

What's the failed idea from 2015 or 20 that you're most excited about?

Oh, that's a good one. Um I think well I'll say I've seen the most progress on this one, but I would not say I'm excited about it. And this should be fun for you. I the answer is microp payments. M

um it's finally kind of working

and we've seen a bunch of attempts on it.

Um

what what's your favorite microp payment use case?

Uh it's the practicality of it, not necessarily the the implementation. It's the fact that uh it can be seamless and almost um zero cost and equally fast. And we're kind of seeing that on uh some Farcaster mini apps

um which I'm sure you've kind of also probably seen on your feed there. And they're they're working in terms of pure volume and number of transactions. Uh but are they mass adopted? No. But are they now physically and technically possible than they were as an idea from 10 years ago or even 5 years ago?

Absolutely. In the best version of that form.

Um we I think we're also oh actually you know what uh well tokenized stocks.

There we go. Let's bring back to Robin Hood. Um,

yeah,

we've seen that happen every year, whether it's an ICO or some bank or some uh country saying we'll do this

to some uh project with no website and team page to um actually seeing a public company try it out.

Yeah. Um or for that matter um like people doing synthetic assets of of every stock. That's uh

Yeah.

Yeah. you know, it's like the thing that uh I remember continually telling people to do and then I keep getting the reply that like, well, there's not enough interest, but you know, eventually there is interest. So,

h what's uh

was there anything on your list other than other than that?

Um h

maybe maybe privacy.

Yeah, I mean like putting privacy into everything, but uh I think okay, maybe the way that I think about it is like like if I had to draw it out as you know like the the galaxy brain meme, right? It's like you know Neanderthal brain is like privacy as an application and then you know human brain is privacy as a feature and then galaxy brain is data leaks as a bug that you're trying to remove. So it's uh yeah I think it's worth you know figuring out how to integrate into everything. Um what what's something that you were excited about five years ago that you feel disappointed now that it hasn't happened more or better? I really wanted VPN and bandwidth sharing.

I think that was like the most obvious use case of idle compute and bandwidth.

And I just feel like any attempt that I saw never got executed well

or made it to hands of more than a few people.

And I feel like that's still a great cool concept.

Mhm. Okay. Yeah, that's uh yeah, it's a good one. Um

also uh paying for uh uh microtasks

Mhm.

in general.

H well what is uh on the other side? What's something that uh you were kind of grumpily dismissive of but turned out to be super successful?

Oh that's a big list. Um, uh, Solana.

Mhm.

Uhhuh.

Um, there were definitely a lot of, uh,

well, I would say Compound is in that category.

I don't know if I I don't know how I classify that now. I think it's in a gray zone because it's it is successful, but not as successful as another DeFi lending protocol in comparison. Um, I think overall I feel like most of the times when I've done that, it's been a reason where I just think the the developers care more about something than the users. So, anytime I look at why I had the belief, it's been because I looked at it from the lens of a of an engineer and not as a user.

And uh it that's probably like a eight out of 10 hit rate on that's why I was like, "Oh, I dismissed it because I thought

a developer would dismiss it, not because a consumer would like it."

Makes sense. Hm. I want to um get back a bit to this uh question of these uh you know like different uh Ethereum cultures that we see, right? Because uh like I you know I feel like you know I see one thing at uh at Pragma than uh I was at the the Robin Hood event two days ago and uh it was a very different crowd and you know the Robin Hood people were there. Justin Sun was there.

Um yeah.

Oh I did not know that. That's awesome. And uh like you know it it felt you know like very different from this room right um and then um you know there was also that presentation yesterday about you know the funds that will be mission focused on making the price of ETH go up. Uh so we uh

that's an idea I'd like to see come back.

Yeah that's uh like I feel like it's it's interesting to see how in like these uh these different corners of uh the world are maturing. Um is uh is this uh something that uh you think is uh is is healthy? Are there are there aspects of it that excite you? Are there aspects of it that on the other hands even worry you?

Um I think it's a mix of all the above

because the stage it's at to me is that one like you actually kind of want this culture. Um because at the end of the day, you're coming to an equilibrium of like what we collectively believe.

If there's not going to be like a top- down way to think about this thing and communicate it that like a public company would or any other startup would, um you need these types of things to actually help you gauge what's going on and how people think about things. So the the kernel of any kind of subculture is important to know as a perspective. and you get to measure the weights of how important it is based off of like how big and how repetitive it is in my head here. Um, but the part that I don't like about this is that it's not often that these things get executed on. It's only about ideas and broadcasting I'm going to do something or we will do something but then there's no followup or actual work put in to do it.

So if somebody wants to promise it goes to 10k lovely. If you're going to actually make that happen and say that this is demonstr demonstratively how you got there,

that's the part I'm skeptical of. Right.

If ETH goes to 10K within a year, will you bow at the altar of Zach?

Hey, I mean, sure.

A win is a win.

Okay.

Uh but but I mean to be fair, I don't think I think price is the least important piece here. M uh but you see all these types of initiatives around just general progress and growth like we see this type of stuff for research grants, we see this type of stuff for promising mainet or a test net or this tech being ready or a performance improvement. Um and I feel like these cultures end up optimizing for being top of the news cycle versus

actually showing and talking about things when they're available. U and that's kind of like maybe one discourse that I that I feel like is uh um the thing that I I kind of hold harder on.

What do you think we can uh do better to try to turn Ethereum from being a petri dish where the the selection function is the new cycle to being one where the selection function is building cool things.

I mean I think it is what it is that right now. I don't disagree with that. it actually is the biggest kind of set of experiments going on in parallel. Um maybe one thing to do would be if somebody is giving somebody an audience

uh probably a decent rule to have is only give them that distribution if they whatever they're talking about exists which is like not a hard check.

Mhm. So it should not just be promises. It should be actual demos and things that people can try out which means that the the the line to get there is well I need to actually ship something something. It could be not in a best shape but it exists and not a consistent promise of it will happen soon and then you only talk about what's going to happen soon. Um, and if you kind of even reduce things by 40%, I think you end up creating a bigger impact because it sets the the standard that we should be talking about things we've done, not things you would like to do.

Mhm.

And if you want to be a builder culture, then you have to enforce actually showing off builder culture.

Yeah. No, I think uh that's I mean it's it's a good prompt for everyone. You know, bidd your culture. Um h what are some uh things that Okay. What what do you expect uh ETH Pragma 2030 to be like 2030?

Um, by the way, if you haven't figured this out, I have no idea what we're talking about. There was there's no prep for this thing than you asking me questions.

So, so I'm going to have to think a little bit every question I get asked. Um, I think the

Well, I don't even know if we're going to have a pragmma in 2030. M

the goal was to be the best curator of upcoming ideas and themes

and that's kind of why we started it

and um I feel like if still 5 years from now we're at a point where people are unsure what's happening in our space

um it feels like not only we failed as as ETH global but also the rest of the ecosystem

and that's a pretty big window

for

this type of miscommunication to be the default.

Um, okay. Maybe here's one interesting way to phrase the question. Um, imagine that there was Gen Xcardic and Gen Xcardic is an idealistic young hacker who wanted to create a frag pragma events to surface the best and brightest ideas of uh the internet. And this started in uh 1995 and then eventually 2005 happened and then eventually they just kept on going. uh what would uh what what should um internet pragma have turned into by now?

Okay, I like this phrasing. Um I think in that world the thing that would matter is uh kind of focusing on at least three of these different categories and and this this is like a uh at any given time in the f in point in time you can actually stay consistent to this one is showcase what's

recently been done that's cool

which is anywhere from new launches to things that are like live that people could try out and that's consistently like hey this is the cutting edge of this thing exists now and it's new and you should give it a shot. The second piece becomes what is upcoming from a research standpoint.

I I I think no matter how advanced we are as an industry, there's always something technical impressive you can do and kind of break your own record of performance or efficiency or just in general like a full shift of how something is done

technically. Um and those are multi-yearlong ideas, multi-yearong executions. you can consistently talk about what's happening there and and demonstrate this is the outcome that you'll see if this thing kind of ends up being figured out or or implemented. Uh and that's kind of like what I bucket of what a lot of core devs do um and researches. Uh the third thing will be generally talking about um economic ideas.

I think the the part that I feel like doesn't get discussed a lot more even now is economic implications of what we're building. M

um and it's not like the audience should be the economist. It's purely that you can still understand effects of your actions and

and consequences of basically decisions. Um and giving people a better way to think about hey if I set up this incentive this is what I can expect. If I set up uh this whole mechanism whether it's gamification to hey if this whole ecosystem exists this way in DeFi this is what the secondary effects could be. I'd love to see general thoughts around how to think about decision theory there.

You kind of need that as DeFi matures.

When you do a hackathons, is there a uh like is there a rating uh section for economic impact of an application?

There is not, but uh that's something we should definitely try. the I think the difference is that it only applies when you're trying to make this ready for somebody to use from a financial standpoint.

And if if this becomes a metric, then it becomes a wrong thing to optimize for

um for a hackathon.

Um so we don't want to

be prescriptive on it should lean towards this way because not everything should be a product that you make money from

as a hacker project. Which Ethereum application are you the most excited about the economic impact of right now?

That's a really good question. Um, I'm trying to see if I have a better answer than unis swap. [Music] Um I mean I'll start the bidding with just uh global universal access to permissionless uh you know like dollars including in uh including interestbearing fiat and uh investments but that's just a category.

I mean it's a it's an application. It's a it's a thing that you do in your wallet.

I see. Okay. I see. Um, I I feel like ultimately

it's going to be harder to beat a permissionless market maker.

Mhm.

If you care about economic impact because to me that is the kernel and like the the least common

the lowest layer of a flow of funds.

Um, and any economic activity is exchanged is basically a function of exchanging money from one party to another for a transaction. Yeah.

Um and this feels like the ultimate way to price anything and you can represent anything as now a token.

Yeah.

So that's kind of my my thinking on it. But

um

potentially being able to actually leverage distributed compute that's idle across the globe.

That's maybe big enough to make a dent, but I don't think it's been uh executed well so far. H

if the if that existed, would you rent your GPU out?

Uh yeah, but the the way I think about that one running is you wouldn't have to make that decision. You kind of just say, "Hey, I'm going to max allocate

20% of my GPU CPU time on this thing. And if I'm not using it, feel free to uh keep it busy."

So speaking of uh you not making decisions, uh one of the hot topics uh for the past year has been uh AI. Um, so, uh, what what I I think, uh, we've already seen like at least 18 months of like people seriously trying to make crypto AI applications. What what to you has succeeded and what has failed so far?

I don't think anything has succeeded.

Why?

Uh, because it feels forced. Um, you're trying to figure out where can I inject AI into things

when uh, you don't have a good answer for what you're replacing. Um, so when you're piling on things on top of something that's already mostly feature complete in terms of what it should serve, uh, you end up with more Frankenstein type of outcomes and, um,

that becomes slightly harder to convince others to care about. Um,

so there is one crypto AI application that's been super successful for the past decade. Can you guess what it is?

Crypto and we're talking just general machine learning AI. If it's 10 or LLM

AI is defined artificial intelligence

Bitcoin

um arbitrage bots on dexes

you I wouldn't call that AI

why it's it's a bot it's it runs algorithms they're often pretty sophisticated

but they're that's cheating that's cheating that doesn't count

okay I I think um

it's like well if I have a 10,000 if else statement

okay so I think for me yeah The reason why it kind of counts like for me is that I think it it actually has

it's not even inference, is it?

So I mean I'll say the the reason why like I think of it as being one is that I think it's like actually stylistically similar to a a very big category of applications that's going to be successful. Uh because uh like for example like the next natural thing you can do is you can have AI playing on prediction markets. Um then if you have something like Claros like any of these AI based judgment things or or blockchain based judgment things you can have AI playing on that. Um then uh but these are future forward scenarios which I agree with

indeed. Um so basically I think uh if you combine AIS with uh the the right kind of economic incentive structure then uh like it actually does become a very good way of outsourcing.

I I I don't disagree on that. I was trying to answer what has happened so far and my answer is no and I still disagree heavily with

you calling a trading bot AI.

Okay. Even if you only look at statistical statistical inference definition um because it's

it's very deterministic case by case scenario only and there's no

predictive power in there

right okay I

in most cases

right I I I'm pretty there there I I know people have tried at least a bit with like me bots that uh

fair

um but but

I don't think we've seen anything successful there yet. Yeah,

I do think the most obvious piece which to me was um I think it is a bit of a bit of a time thing which is we are seeing agents themselves become more um

stable

and now you have the ability to kind of throw one of them in in the wild and equip them with the ability to take and receive money um or or send money and um if they're really good at a task which can be nothing related to crypto at all

um that's a potential candidate right like if if I just have a bot that just summarizes an essay, a news news article for me and charges me two cents,

I I think that that counts. Uh, and that's more likely to happen when you can trust the quality of the outcome more, which I you couldn't a year ago,

right? No, I think uh that makes uh a lot of sense. Have you seen wallet AI use cases? Um I not directly but one of the speakers today did talk about this uh which is they have uh on the flow wallet.

Mhm. there is a mini agent that can um try to expect um if you're going to run low on your balance and uh try to make a connection with anywhere from fiat or other credit cards on file just to prefill your your wallet so that you never run out of funds or miss the the transaction which I thought was like a cool wallet agent. Um, but I'm sure this will be a very cool thing to put as a default in the future where you actually have a mini uh fraud detection bot somewhere and it just flags and pauses a transaction if it thinks it's weird and that's a great use case. You could be false positive heavy. You'll only get better, but it actually prevents risks.

Okay, cool. No, so uh still uh ve very little on AI so far, but uh a lot of interesting directions to go. Um

I have a question. What uh what was your talk yesterday about? I couldn't see it.

Um it was uh on uh um urging a shift from thinking about decentralization in terms of what technologies you use to decentralization in terms of what properties you give. Um

so is the property just having a choice?

The the property is uh guarantee is that you won't be rugged. Uh guarantee uh guarantee is that the operator can't suddenly take all your money. um guarantees that your data is private, guarantees that if the operator's uh operator disappears, you can still take your assets out.

And you feel like people have forgotten that?

Um often, yes. I mean uh actually there's a there's a lot of examples of uh like projects where theoretically they use crazy uh crazy tech but you know there is a hidden uh like developer instant grab button or there aren't like any alternate um alternative UI interface flows. It's uh like it it is actually something that I think our ecosystem needs to be pretty on the ball on. Um h

did anybody come to you after and ask any interesting questions about the the topic? Um in that case uh no because I mean I though I mean I did see I mean like some of the feedback from uh you know like both kind of chats and the internet and I think you know it was the whole range from like basically like yes these are valuable things to think about um to wait is uh is Vitalik saying crypto is bad now? Uh so you know

that's not what you were saying. Um.

H. Okay.

And that's why we want to eat that 10k. Um.

H. Cool.

What uh

so bad. Um,

how do you think about a a normal person and and I say that as you and I are not in that category and neither is most of this audience.

A normal person

thinking about proactively having a choice. Um,

we've seen that in the world of AI where it's been definitively proven that people are very much okay with their entire life and personal problems being uh on a server server to uh to uh get some therapist agent to act and they have no regrets at all putting that information away. Um, how do we make that? H

yeah I think I mean one important point there right is that individuals often don't care but like institutions do really care um and uh actually like we have been uh talking to financial institutions more and more of an as the ethereum foundation recently and uh one of the interesting things that I've seen right is that I think uh there's this misconception that sort of only the idealists care about decentralization censorship resistance and privacy and like serious people just want you know, IBRL and the reality like we've like we've had banks like actual, you know, like licensed major nation level banks tell us uh that like actually, you know, they appreciate Ethereum for its uh for its censorship resistance, you know, like for its global neutrality and and the thing they're exploring is figuring out how to get privacy. Um

um does this answer change when a bank has I mean so my my default again I I have no context in any of this. My default would have been a bank's best interest is to care about this thing if

they're working with another bank

especially outside of that country.

Yes,

great guarantee.

Mhm. not always the case if you want to do this for your own users because every time we've seen things go wrong,

it's your bank freezing your assets first so that there's no bank run.

Yeah.

Um and more moreover, I guess the part that I'm I'm kind of curious to hear your perspective on is

even if a bank is good with this, which is that's a progressive bank. I'd love that.

I wonder how many governments want this.

Yeah. Um I mean I think more and more. Um so uh I think I uh I published uh on Twitter I screenshotted this article a few weeks ago. It was uh some of the Nordic uh governments uh and how they uh discovered that uh the pushed cashless society that they had made over the last 10 years was actually making their society like very fragile and it had privacy risks and like potentially like vulnerable to like hackers from like Russia or wherever else. Um and uh they uh are like like some of them are actually thinking about going backwards and mandating cash acceptance.

Um so on and on on the philosophy like you can see more and more of these instances of uh like governments actually yeah actually I think even a little bit uh like even thing like parts of the NSA um accepting that uh you know like maybe this whole you know like we're at the top and we see everything approach privacy and security doesn't actually work in the context of all of these scary governments have hacking each other and like increasing level of global competition and uh like maybe a uh uh like an approach to security that's kind of philosophically like more sort of like local first um is actually and like minimizing central data collection is something that they want. So I actually am optimistic that the the philosophical backing is uh there but I think it's something where like we have to step up and like actually offer a concrete solution.

What is a concrete solution like just uh it has to be efficient enough or we're missing something technically that doesn't exist?

This I mean I think this depends on the context. And I think in the context of like say like privacy for banks for example like just an SDK existing that makes privacy preserving payments very easy for them to integrate.

Cool. Um were there any other ideas discussed in that talk or was it was this the main message? Yeah, I think another one was uh I used Android as an example and I think Android's a really good example to mine uh because uh Android is Linux and Linux is known for being this idealistic ecosystem that constantly wants mass adoption but keeps having a hard time breaking above 1%. Although recently

technically Android is

exactly Android is you know the big crowning exception increasingly enough there are starting to be other exceptions I've been told that uh because of gaming and steam actually Linux on the desktop is going up and up a little bit more now um but

with the second you bundle in any web server you uh you're extremely

diluting every other opportunity and uh

yeah with Android I mean like it brings mass adoption but then the question is uh like does it achieve the goal of making the user free right and uh I think like

right after you sign in with your Google account yes

yes so I mean the answer is like in interesting ways sometimes yes sometimes no you know it's like you know yes you can uh sideloadad your own APKs if you want but

but you have that choice you don't own an iPhone

exactly you don't have an iPhone but also on the other hand the Android that gets put into on into consumer faces by default is one that has all kinds of uh like spyw wear and crapware on it that's put by various manufacturers, right? And so it's uh I mean it's like an interesting one to kind of explore both ways, right? And to kind of analyze the successes and uh and and also to see where we can do better. Here's um

I don't really answer like I actually don't have an opinion on this either, but do you think only people who are sort of techsavvy are the ones who hate the default random apps that I that you see on your phone versus like the billion other people who have them? They're like, "Oh, this is great. This is my new photos app and this is my new contacts app and I'll just this is their default." Uh do do you feel like because you know what

the other side looks like you prefer that choice instead of uh it just being indifferent to

sorry what's what's the

like like kind of building on top of your uh comment on like hey if I buy a new Android phone it comes with all this uh manufacturer software for default apps that I may not need or you think are not the best versions of what you can download from the app store

but do you think a billion other people care notice? Yeah.

Yeah. I mean, I think I mean like in general a lot of people don't care because they're like not educated to like know a lot of the a lot of the nuances, but in if we ask the question kind of is it good for them, then I mean often it's a mix, right? Like there's definitely a lot of uh default choices that like actually make sense for people, right? And uh you know the average person is more likely to worry about something like uh you know like losing their account than to worry about some kind of like very advanced um like h like hacking attacker um against them right but uh on the other hand you know there are things that like actually are abusive crapware right and there are things where it's like uh actually I mean like these are companies that are taking advantage of unsophisticated users your battery life uh drains faster ads that exploit you and uh so

default analytics.

Yeah. Yeah. So it's I think a mix of both and uh I mean I think uh like actually uh thinking about like I think there are values that crypto is good at that are very beneficial um even to normal users um even if it it takes work to make them intelligible and like I think there is even like value in us thinking about and focusing.

Yeah, there's nothing wrong with that. That's a good point to uh to have. You want to try anyway to see

if there is a

a in line where after that it just becomes okay for everybody to consider.

Um

I'm going to change the topics a little bit and we'll

do a couple more questions then we'll uh we'll end.

Um okay so Robin Hood now has a chain. I'm sure 10 other institutions similar will do the same. Mhm.

And then everybody else will have the same idea and now you have 100 more institutional chains.

Uh and then probably the same amount of them wanting their own stable coin too.

Mhm.

Um is that a good world or what is definitely happening. So I wonder like what does that mean for Ethereum from the engineering side but also like more fragmentation in general?

Yeah. Okay. So I think one world that would be a better world is a world where assets are issued on L1 by default even if a lot of operations happen on L2. And uh I think the reason why this is good is because then even if your L2 has some kind of like some you know like governance properties that are not very uh decentralized if people want to use the assets outside of your L2 then like either on L1 or on a different L2 then they can fully escape those governance properties right on the other hand if your asset is natively issued on L2 then like your users are vulnerable to those uh governance properties even if they bridge over to L1 and bridge over to other L2s. So I think a world where L1 is the center of issuance is one that is uh more composable that has better trust properties that has uh better properties in terms of people being able to make applications that like combine things together better.

I mean that's a great world to have but that's exactly why we're seeing them do their own thing because they don't want to. Right.

Right. I mean it's uh it depends. Right. I think sometimes you know it's easy to overstate the extents to which uh some things like other things happen because of perverse incentives when often the thing happens because there just like aren't good technical options and we have to actually build them there.

Um we talked about this thing a few months ago but that would be a great world to have if if the assets are on L1 first and we kind of bridge them to L2 and that's the the in a way the proper layer 2 version of this. Uh but you have to kind of make it a lot more easier for the L2 to understand what's happening on the L1

on the the just the EVM side.

Um

is that a priority or is that in progress?

Uh I mean I think yes. I think uh so like I've been uh a fan of the L1 SO op code. I mean I think uh either L1S load or remote static call which one of them reads state the other does a call and you know like optack I think supports remote static call natively in principle and then there's efforts to do all one z um as well I think like doing more to standardize those components into the stack uh would be good and like that lets you do even more things like that lets you even have an account on L1 where uh then like you can do different operations on L2 but then the logic of your account still still lives on L1 which makes uh

oh cool

yeah a lot of multi2 stuff much easier um and then the other thing is minimizing deposit withdrawal windows right because uh like this is a big reason why uh like we're not seeing more adoption of like more trustless versions of bridging which is that optimistic bridging takes a week and uh like we're actually getting to the point where like we have real time zk VM proving um and So the one the main challenge is security and so uh to try to kind of mitigate that there's like two approaches like one of them is the yeah two of three opzk te that I recommended a few months ago and then the other thing is like formal verification of the zk and like actually there's been surprising amounts of progress on that. So like some of them they've managed to prove a property that's called determinism. And what that property means is that uh the uh the zkar accepts exactly one state route. It could be the wrong state but it accepts exactly one. And that guarantees to you already that uh like it's like if there's a bug it's not some kind of like soundness bug that lets someone like put in whatever fake data they want.

But this is only testing for the ZK circuit. Yes.

But it doesn't have any knowledge of what is actually being executed.

Right. It's well

you can still have a bug in the smart contract but

Right. Exactly. So the the three components are the ZK circuit. Then there is the ZK sonark algorithm itself. Um and then the third is like the solidity implementation of both that algorithm and like things that are dependent on it.

And so I think for the third like some combination of standardization and formal verification a lot of that's already started to happen. Uh for the second that's the hardest piece. It's it's a hard piece because uh just because of all of the the need for constant optimization the ZK algorithms just keep on constantly being in flux. So I think realistically we'll have to wait a bit but you know I agree it's super important. Um and then uh yeah and then for the first one like it's still ongoing progress happening already and like hopefully the the L1 proving race will will keep kind of galvanizing some of that attention.

Amazing. Um I want to end off by asking a question you asked me a little bit ago

and um but I'll kind of reduce the time frame. So, let's not talk about 2030, but 2027.

H,

what do you think we should Well, what would you like the culture to be

in 2027? And what do you think we should start doing now

so that we can actually start tracking things? I

mean, 2027 supposed to be the year AI takes over the world, right?

Depends on if I can buy the OpenAI stock on Robin Hood. Yeah, I know. I think I mean I I mean I think it's uh like the best thing for Ethereum is to have a combination of you know the culture that's able to build out these more mainstream things and like and really make them available to more people and uh like really like make the highest quality version of uh the thing whether it's DeFi, whether it's uh tokenized stocks or even whether it's the non-financial applications that are still important or prediction markets. Um and then the second and then also continuing to have the frontier culture that's uh trying to be more aggressive on privacy properties, security properties, uh decentralization properties, constantly looking at the next thing. So uh it's uh you know 10 years ago it was smart contracts.

Um you know today it's it's ZK in five years it might be like offiscation. um than potentially like something else and something crazier. Um so I think uh continuing to have very strong versions of both of those uh like that will make me very happy.

An excellent note to end on. Thank you so much for this amazing conversation.

Thank you Curt. [Music]