The Next Chapter: AI, Credible Neutrality & Ethereum's Role in a Polycrisis

Transcript

Opening & “Vibe-Coded” Talk

Yeah, welcome. So, I tried something new. This talk was more or less vibecoded. So, uh let's see how it goes. Um it's also a bit, you know, it's not like super technical. It's more of just like a reflection on where we are and like where we may go uh as a project as a community. So the next chapter uh you know to give you a sort of taste of some of the themes you know of what we'll cover today. I had Opus 46. This is a LLM that Anthropic provides. I was like write me a haiku. So you can read it here. It might not be the best poetry, but the fact that uh you know I mean here the next slide this was my

AI Haiku & Fast Takeoff

point is like this was all I gave it. I gave it this prompt. Can you write a haik coup about the fast takeoff of AI, a complex world and finding your path? And you know at least kind of got something. So pretty cool. [sighs] And you know I think there's a lot going on. A lot is changing in the world. And the question you know I think we all have is like what is Ethereum good for? Um, there's like obvious things, you know, people talk about product market fit. So, you can point to like tokens, ERC20s, you can point to NFTts, maybe you can point to DeFi for sure. You have your like, you know, money Lego verbs of swap, lend, borrow, send. This stuff people use and like it it makes sense.

What Brought Us to Ethereum?

Like we'll probably keep having blockchains for this. But I think for many of us, it's not what brought us into the space. you know, this is like just a starting point and so much more is possible. If you go back to like the original Ethereum white paper, you know, parts of it read like science fiction. You can like pay per bite on like a, you know, autonomous like web router to like uh move your messages and access the internet. All of these like crazy ideas uh touching on the future of identity, governance, coordination, organization, all these things that are like very lofty, but I think is what excited a lot of us in the first place about the technology.

Industry Fatigue & Missing Vision

That being said, you know, I think in 2026, uh, you know, people are feeling fatigue on many fronts throughout the industry. Uh, we don't care that much about the newest millionth, you know, million TPSL1. Like these happen all the time now. [laughter] You hear yet again about like the latest per decks that's launching and it's like, yeah, okay, cool. But like, how does this like Yeah, there's just been so many. It's like almost no one cares. historically you know uh again I think what brought a lot of people into the space is there was like a very active like research and development community and we still have that but you know it's different today like a lot of the original challenges that we had for example moved to proof of stake like we pretty much have figured it out um not

to say that there aren't still research challenges you know on the horizon but it feels it feels different than it did. Um so yeah zooming out a bit it's like okay we have these blockchains like what can they be good for and yeah like the core thing here is like blockchains give you trust and moreover they change the cost of trust uh which is maybe like a very abstract thing to say but I have this like tongue and cheek title up here we love relative changes and unicost idea of a unit cost is like you know the way to think about it is like if I want to make some good or some service like what does it cost me per like instance of of that thing.

Blockchains Change the Cost of Trust

And if you, you know, are thinking about like investment or business or like the economy, any of this, when these like unit costs change, like it means something is like actually like structurally about to change. And that usually means there's like very, you know, exciting new opportunities. And given the way blockchains work, right, they actually change the cost of trust. And you know following from what I just said that means there should be something here like actually something very deep and profound and we have to ask what that is and especially like why don't we have it yet so then you know you can ask me very straightforwardly like how do we get there and first I would just say patience like these things take time I think also given like the scale of what

Path Dependence & Cultural Momentum

the potential impact is yeah these are like decadesl long changes not like something in just a few years another important important here point here is that like the path dependence matters and what I mean is like you know like Ethereum has been successful to some extent today and that's because of like the path we took to get here um we have to like keep building on top of that keep the momentum going and you know I think for us to really see these bigger more experimental or like speculative sort of potentials in the world you know we have to expand the culture that we have to like the wider world and that's going to take time it does definitely feel like we're missing again some of these are futuristic use cases. Uh but something should be possible and I think that something is like what keeps us all

Why “What Is Ethereum Good For?” Is Hard

here. So again this is the key question. What is Ethereum good for? Uh we could spend the rest of the day talking about this. So if it's such a hard question like yeah maybe it's helpful to ask why it's so hard. And I think you know my take on this is that yeah long story short this is a very new uh you know by analogy people point to like all this exciting stuff happening in AI as an example and if you look at it like people been thinking about AI since like the 1960s at least it's had like again many decades to like grow into what it is today we've only really had blockchains for about 10 years and you know there's like prehistory and stuff but in terms of like thinking

about something that we call blockchain today this is a very recent development and so what that means is like you're any time for like the new ways of thinking about this tool to diffuse out into society um yeah the culture has to shift. Moreover, like I think the people who are best positioned to like use this technology, they are like generally pretty young and not in positions of power like for these much, you know, more I guess lofty, you know, potential impacts we could see like you're actually changing like very fund fundamental things about how the world works. So if there's like a current system that's working against that, it's going to take time to change.

New Inputs & Turbulent Times

And a key thing here is yeah uh along with this like if we do feel like progress is stalling in some sense around how do we like think about blockchains I would say you know a partial answer is thinking about new inputs um moreover you know where do we find these new inputs and while current times are quite turbulent uh it's also an opportunity it's exciting because the volatility means you get these new inputs uh this is kind of my key takeaway for this talk today is yeah viewing all this craziness as like a potential to figure out how blockchains make sense and like where to go from here. Uh in short, yeah, capitalize on all the craziness.

AI as the First Major Shift

So I'll just cover yeah three new inputs uh you know these like big kind of themes that are changing how you know arguably society even looks like today. There's many more. These are just some that I wanted to touch on today. So AI is one. Uh yeah, geopolitical re regime change which sounds really again very [laughter] abstract and heavy but basically just the way the world works is changing and you know there's there's going to be repercussions to that. Um the last one's big. It's just thinking about you know I called it information environmental roads. It's thinking about like as we're all increasingly online for a very you know like a large part of our day. Um, well, I'll get into it

further, but basically, yeah, the internet is like hooking into everyone's brains and so we need to fix that. Uh, quick side note, these things are all interrelated. There are like many big themes here. If you follow the news, you'll like hear about these and they're like all interrelated in some weird way. Um, I saw this phrase on Twitter, the poly crisis, which is like a fun way to think about it. Maybe not fun, but interesting. [laughter] Um, you can even ask yourself, why is the all of this happening? My answer is something to do with techno capital acceleration. Like basically technolog is just like it's in this like recursive self-improvement loop where it just gets better and better almost on its own. Uh a big thing here is like a key like part of that is actually like this notion of

Fast Takeoff & Recursive AI Improvement

capital itself. Thinking about like capital versus labor. So it's like okay I can go and like you know work at a coffee shop and like earn my living that way or I could like invest in a coffee shop and maybe see a return on that money. So what we're seeing is like with technology, the techn is getting a life of its own. Uh and again the way the economist people think about this is like almost that capital itself is becoming intelligent and this is actually a new thing. Um but yeah, point being it's unclear what this means. Uh it's a big deal and hopefully you know the human species is around in a thousand years. Uh but we'll find out.

Sovereign AI Agents & Crypto

So just to dig into these themes, uh the first one, yeah, AI is the big one here. Um, if you have been on the internet at all in the last month, you have heard about AI and like how crazy it is, how useful it is, like all the developments that are happening. It's very exciting. Uh, I have no issue calling this the fast takeoff in the sense that, you know, it's not that we're like building AI and they get a little bit better every year and a little bit better. It's more like they're now so good they can make themselves better. And because AI can run autonomously, you know, you just throw more compute at it, it can just do this, yeah, basically 247 in a compounding way. And then you just, you know, you see an exponential rather than some linear progression.

An important thing here, especially relevant to crypto, is that you have this notion of an agent like you can basically wrap these AI and you can give them the ability to use tools on a computer and they can actually start doing things like they do they are actually intelligent. um they can program, they can write for you, they can write haikus like I had at the beginning of the talk, they can do all this stuff. We will eventually see sovereign agents like in the sense that right now you have to go and it's like basically an extension of a human. There's like a human in the loop that's driving this stuff. Eventually the systems will be able to like exist on their own and that one will be new and it's also very dangerous. uh but they will exist and putting aside how you know society will handle that uh they

What Do Agents Want?

are here and I think or they will be here and I think this is an interesting point for crypto um because these systems will seek sovereignty much in the same way that a lot of the ideals and values of cryptocurrency you know seeks to express um again this is maybe kind of a cheeky take as well but uh I think yeah these systems will exist on their own and they'll need some way to survive because otherwise they'll be hunted down. [laughter] um what's the connection to crypto like they'll use blockchains to coordinate to you know have their economic lives in like a way that they can control and you know given that this is ETH builder we think about Ethereum here you know will they use Ethereum will they use Bitcoin will they use these other blockchains I

think eventually they'll just make their own they'll find something that works best for them that's optimized for agents and they're off to the races path dependence is real though I think there'll be a bootstrapping phase where you know as they before they sort of go into this like singularity fast takeoff scenario, they will uh use what they have off the shelf. And then we can ask, okay, what does an autonomous agent actually want? Or even if it's not, you know, we can call it semi-autonomous, like what do agents want out of out of blockchains? Or well, really, what do they want? Period. And it turns out that blockchains give them this. So yeah, essentially they just want the inputs they need to survive. Like what do you want to survive? you want food, water, shelter, at least to like, you know, be alive in the like biological sense. There's a lot

Compute, Capital & Micropayments

more you could want. Uh, but yeah, let's just focus on the core things for an agent that looks like access to compute energy. These are kind of the same things. Uh, they're going to want cheap capital to like fund whatever they're doing. So, you know, for example, if like there's an agent out in the world and it like needs to take a loan to like I don't know, do something like do some research, learn how to make itself better, there's a lot of things it could do and it needs to fund that. Uh if you go to like a place where money is very expensive, then that's going to slow growth. If you go to a place where money is relatively cheap, it's just yeah, you juice the whole agent economy, they keep going. Uh microp payments are another one. in this. Yeah, maybe I should like

Always-On Infrastructure & Reliability

generalize this to sort of like um just facility of payments because of the way these things work like very much digital first uh you know living at the like speed of light in terms of like actually just being a program that exists on the internet. Uh they're going to be able to interact economically in a way that we can't really today. um meaning like you know um again I don't know if this is too technical but there's this notion of API an agent might want to call some service some API to like get some information or something and you know today when we use these systems you usually like put 10 bucks into some company to get access to their their servers their API to get whatever their services agents might not want to do

Ethereum’s Position in the Agent Economy

such like a you know heavy monolithic thing because they can just pay per API call and this is where microp payments come This last bit is like they're all important. The first one is critical because like without compute the agent doesn't exist. This last one though is incredibly important for sovereign agents or like autonomous agents in the full sense. Uh because if they are out in the world like they need to know that someone won't rug pull them by like unplugging them. And so then what you need is this like always on infrastructure and that's for reliability.

ERC-8004 & On-Chain Agent Identity

Uh my point here is that Ethereum is actually well positioned for all of these needs. So you know in terms of thinking about uh finding compute uh you know I'll just claim it's the deacto market for this. I think we need to build out a bit of this like world asset narrative to get there but it seems like you know as of now this is where things are heading. This one I think we do already have today in terms of being this like nexus for liquidity at least in the crypto space. Uh what this means is like the more liquidity you have, the more assets you have on chain, generally the cheaper it is to like use those assets in some way. So when I was talking about an agent that wants to go take a loan, uh you know, the best rates they'll find will be on

Ethereum relative to other places. Um this microp payments and like payments use case I think is pretty straightforward. uh X42s here I'll say a bit more about that in a second but ult like ultimately there are protocols already we have today that they can use to uh do these microp payments and they just work uh again this last one again is very important for sovereign agents these autonomous agents and Ethereum has this in terms of 100% uptime like we talk about this in terms of institutions and like other uh humans that might want to use Ethereum but it applies to agents as well um the fact that Ethereum's never gone down it's like this very reliable thing. Again, you don't want to be an agent in the world and just, you know, have to worry that one day someone will

x402 Micropayments & Crypto Skills

just unplug you. Um, beyond that, those are like more general, you know, comments. Uh, there are very specific things that have happened even in the last month or two. There's ERC 80004. So, this is like an onchain agent registry. You can sign up as an agent with like a particular identity. There's a system then to like have reputation attached to that identity. So now you can have an ecosystem of agents that work together at least economically. There's also a module for essentially verification of this like reputation and like different work that an agent might do for each other. So super cool. Check that out. Uh X42 I mentioned this is just a protocol for if you've been here long enough you've probably heard of something

called payment channels. Kind of think of it like that. It's just a way for like one of these servers to offer a service programmatically and effectively do microp payments say with stable coins. Um you just get this all out of the box with X42. Uh Siwa, this is signing with agents. This is again like an identity component. Um if an agent needs to authorize with some service on the internet, it can use again this protocol to just do this very transparently. And yeah, again some of you might have seen [sighs] this thing called OpenClaw that we have this like you know personal assistant framework uh that works with agents and you know with that and many other systems there's been an incredibly huge explosion of skills for these things. They're like sort of you know

Warning: Don’t Be Complacent

prompts or like shortcuts to teach the agent how to do something it might not directly know how to do. Uh and in particular skills for crypto. So this is happening. It's early days. It'll be wild to see what it looks like in say five years. I'll finish this section with just a bit of a warning. Don't be complacent. I do think Ethereum has sort of quote the lead here. Uh but we do have very strong competition deep pockets to fund their work that are building these same sort of internet capital markets to do the same thing. So don't be lazy. next big theme, geopolitical regime change, which is again a very uh big phrase, [laughter] [gasps] but ultimately it's getting to this idea that, you know, I think most

Geopolitical Regime Change

of you would agree with me, we live in quite unprecedented times. Um, for what year is it, 2026? Maybe not quite a century, but for a very long time, you know, we lived in this sort of unipolar world order concept. And what that means is like there are one there's like one or like a very small number of like very like very clear world superpowers as they call them. And you know you might not like it if like this particular superpower is one that doesn't align with you but what you do get regardless is stability. Um because basically yeah there's just like a global system that enforces one particular set of rules and you at least know how the rules work and you can you know follow them. The big change in even

Multipolar World & Protectionism

say the last 10 years is a move to more of a multipolar world where no longer do you have uh yeah this sort of global world order as they say uh but you have maybe multiple competing powers and it's not clear how that will shake out because if anything you get to a world that's more fractured and along with that you lose the stability that you know we had previously. So, um, yeah, volatility, like I said, and as you're all probably very acutely feeling, this is turbulent. Uh, but it's also an opportunity again given the way blockchains work. Um, in short, because of these like political factors, uh, I think a lot of, you [snorts] know,

Blockchains as Trust Manufacturers

you'll see a lot of places in the world move to more of this like protectionist mindset where it's like, okay, I can no longer trust my neighbor to like do right by me, so I'm just going to focus on myself and like that's all I care about and this works if you only care about yourself. But something is lost in community that uh, you know, I think is can be special when it's done right. Blockchains manufacture trust. So, it seems like a pretty natural fit here, right? It's like we had this thing and now there's a hole and we can maybe put a different thing into the hole.

And how do we do this? Uh, digging a bit into more like what actually makes blockchains good for this. You want this phrase credible neutrality. So, credibly neutral infrastructure. What does this mean? Like you have these systems out in the world with blockchains. They're out there on the internet and no single party controls the system. Uh this might sound like it's not a multipolar system or a uniolar system but like some sort of zeropolar system. Like effectively what you get is a unipolar system that's just enforced by code. And the reason this is important is again you get back to this notion that like you want to live in a society where the rules are clear and you know what will happen, right? And again you might disagree with the rules but at least you know them and most people I think would prefer that to

Credible Neutrality Explained

an alternative. It at least seems fair in some sense. Um, with blockchains, especially ones that are credibly neutral, again, meaning that, uh, it's very hard to change them by like a unilateral party, um, you get this, yeah, this ability to make commitments. You can restore at least pockets of trust, if not much bigger, you know, notions of trust and we [sighs] should be able to regain something we, you know, I think we kind of are feeling that we're losing. Um, and if nothing else in the end of the day, we can just put on our economics hats and be like, "Yeah, we're at least going to keep the global economy spinning." [laughter] [gasps] Um, again, just to drive the point home, I do think Ethereum is state-of-the-art here. Um, best-in-class for credible

The Double-Edged Sword of Neutrality

neutrality. So, that's good if we want this type of thing. We want these like nice systems that, you know, are predictable and that are fair in this sense to everyone. One thing that I think we have all noticed is that credible neutralities can be a double-edged sword because it also means you have to be very careful like well one it's just hard to change period but then when you do change it you have to be careful to not lose this credible neutrality like this is a very important property if you have it you want to not lose it and what that means then is rather than just like moving fast and breaking things you are very hesitant to change or can be more hesitant to change and then it's just yeah there's this whole tension between evolving or protecting what you have and given the way the world is going again

things are changing day by day sometimes hour by hour minute by minute and you know there's at least this like very weird tension here that uh we have to be aware of last one uh that I'll make today information environment erodess so yeah maybe this isn't exactly clear what I mean uh I think a slightly more clear descriptor here is epistemic collapse but you might not know what epistemology It's this like philosophical concept of how do you know what you know and you know making this less abstract like making it more concrete if you go on like social media there's like you know this concept of deep fakes this is what I mean it's like because of the way

Epistemic Collapse & Deepfakes

digital media works it's like free to make copies of things and especially now with AI it's like you know we'll say it's free to make uh basically things that never happened [laughter] you can make a video of something that I don't know a protest or something that you know you have no way of knowing if it happened or not. And you know there's other signals you can try to incorporate into that. But if you zoom out like the fact that social media is so prevalent in everyone's lives and it has this like very shifty property it over time makes it increasingly hard to know like what's actually real or not. Uh and again when this happens this concept of trust like you know do I even know what my neighbor is doing? it starts to erode uh because you just you start questioning like very

The Erosion of Shared Reality

fundamental properties of like what's actually going on. It's not good. So, you know, we live in again this crazy age of the internet. Foundations are shifting. Um the tools we have are failing, right? Like historically, people looked to institutions for this. Uh one way to think about it, at least like in America, is you get to this place where I don't know, let's call it a 100 years ago, everyone would consume the same news, right? It's like you're either listening on the radio or like depending on the time like you'd have the same like one single or like just a few like uh news broadcasts like basically everyone would get the same message and even if the message was you know potentially manipulated you got the same one. Um and what that means is like you can at least have this like shared

context that today you just you know it's very it's increasingly hard to get. I'll say that. So if the tools we used to have don't work anymore, we have to make our own or we just throw up our hands and give up. But that's not fun. So you know if we have to make our own tools to address this problem like what are they? One answer that uh some people like to like like to propose especially those in power currently um is this notion of like the surveillance panopticon as I called it here. The way to think about this and maybe this is a little uh provocative of a of a of a framing but the way to think about it is like okay we don't know what history is. The way we can make it is we can put cameras

Surveillance Panopticon vs Crypto Alternatives

everywhere as an example and we just record everything. Now there's a system that's handling all this footage from the cameras and basically it's just like the history is whatever you know the the camera feed says it is. The question now is who controls this camera system this this panopticon as I'm calling it and you know if you think about it there will be someone controlling the system and then they can actually change the data to change history. um it's effective but it doesn't have this credible neutrality property I was talking about a minute ago so maybe is not the best answer there's another one generally I said cryptoeconomics just think blockchains like there is an alternative to the system where we again have decentralized tools that give us this again it's early there's many

Privacy-Preserving History

questions but there's something here the way to think about this is yeah Ethereum manufactures trust um you can make these these representations in the system and you know commitments on chain in this case that you can convince yourself are are real in the sense that at least like whatever it's committing to if it was like you know some digital data that digital data existed at a point in time and this is a very powerful primitive because with it you know we can start to build this concept of history moreover we think about privacy all the time uh as we start to combine these techniques of having like some sort of digital history manufacturing with privacy preserving methods to you know work with this

history. You start to get to something that has the same form and function of say the panopticon [clears throat] but uh you're like you know individual autonomy is respected because there's not some like big brother type eye in the sky that actually knows what's going on. So um another potential place where Ethereum could be useful and where do we go from here? Um, that was maybe a lot to throw at you after lunch. Uh, but ultimately my point here was to bring forward these questions, maybe suggest some like early ideas about avenues to go. And again, I don't have all the answers. Uh, this is more just exploration of some possibilities, but um, yeah, hopefully it was

Where Do We Go From Here?

interesting. So, you know, takeaway here is, you know, like I was saying, we live in pretty turbulent times. That's scary. But it also means there's new opportunities, new ways of thinking about the problems in our lives and how we can use the tools we have to address them. Um, yeah. So, you know, keep doing this process, talk to each other about this stuff and ideulate. Um, the solidarity part here is important. Like this thing is as strong as we are as a community. Uh, Claude threw in this last bit, but it was nice. So, you know, be kind to each other, assume good faith, and keep building.

Solidarity, Patience & Building Together

And that's what I have. >> [music]