Accelerating Ethereum

Transcript

Doesn't seem like it works. Okay, it does. Hi everyone. Great to see everyone and let's talk about accelerating Ethereum. I mean it will be mostly accelerating your ideas and thinking like founders about where the world is going. So the world is accelerating, Ethereum has to accelerate and Ethereum accelerating means that you have to accelerate. You have to accelerate as founders. Um all right. So yesterday I had very nice conversations with the new generation Gen Z people uh everywhere around here in Khan and uh I just recorded a stream of consciousness from them and they say look who makes the fun make Ethereum fun uh Ethereum should be fun marketing now is all about fun and so on and so on and you can hear probably also about um how how is in

Bitmine uh teaming up with with Mr. Beast and people saying look people like Mr. beast are creating the banks of the future. That's where the finance will be. That's where the trust will be for the new generation. It means that you already have to start rethinking of how you built your products and how you structure the marketing for your products. Um well, let's skip further. So I think that generally the world is changing well very very fast. So accelerating and when you look at blockchain if you haven't moved from blockchain to or like crypto crypto spaces crypto communities for the last few years so it feels like you might be building for the world that is disappearing very fast and I'm trying to encourage everyone to start thinking about the world in few years because the change is so drastic so dramatic beyond

AI obviously um that it might be that your products are targeting the users that are disappearing which means like humans versus agents as users And uh typo if you do not build for the next 10 years and that probably you're building for the past 10 years. And it's it's very very important that you need to notice that many of the leaders that we follow and that we read very often talk about the past 10 years. And um it's something that I was thinking about very often when building organizations that if you over rely on stories and the legends of how the company started of how it was in the past then over time you start being exclusive for the new ideas. Uh so it's it's very very risky

to talk too often about how Ethereum was imagined 10 years ago and last year we were celebrating 10 years of Ethereum. This year we definitely continue celebrating the next 10 years of Ethereum. Um so the user interface as we know it is disappearing. So potentially browsers will disappear. If you see the new products like browser base browser less agents are uh searching information on the internet by rejecting the browsers the traditional browsers. Users are using agents to search for information or execute actions. Uh maybe some of you were already using the open cloud open open claw or so. Um then you use it for opening accounts, for finding information, for setting up things. Um I

was using it for opening GitHub accounts, Versal accounts, buying domains, for deploying websites and blogs, for uh setting up cloud accounts and so on. Every now and then I would have to solve capture. But the thing is the moment when you start working with agents and AI chats, very likely you're not using browsers anymore. But still very often you're building products that are very very browser focused. might be better if you're building it for some kind of like trade trade execution engines and uh trading dashboards. Uh you can expect that most of your power users will use browsers for a while but even them they're moving to solutions like myrrofish agents that are representing them in trading and so and so on. So browsers may disappear and when the browsers disappear it means

that everything that was created for browsers it's really in a very very awkward format which means it will not be the uh worldwide web but it will be lots of um lots of data in databases indexed for agents and pre-training for like thinking about how your data how your marketing lands with the LLM models with the base models and how to make sure that you always they're up to date and they're choosing your product over the others. So AI and agents are consumers. Uh very often no longer humans will be consumers. Uh and then they can either represent their like independent identity. Think about it as if it's uh holding inside a TE the uh the enclaves or representing a human because you may have humans saying I'm

spawning agents and they borrow my identity or as an organization or as a as a individual I can launch an agent and try to make it uh make it independent replica self-preplicating paying for itself and existing in the clouds maybe and and always escaping anyone trying to shut it down. This might be the user of your services or buyer of your products. Um interfaces change, pricing strategies, marketing form, experience of the users, well the users themselves change. So all of that has to be rethought. Um one thing that I believe doesn't change is the research is memes. We maybe missed it. Uh that inte uh sorry the like unified theory of of memes is not only as the meme coins

but also much of the research in the blockchain space was a meme. So we were looking at how viral and uh how what was we what would be the half-life of the uh research ideas that researchers were proposing and we were kind of gumbling or betting on those research ideas and it's a bit like this also with the startups. So the startups come with the ideas and you you bet both on the idea and the execution. Uh but the markets that we're creating very very much allowed us to separate the ideas from execution. uh and research very often uh followed with limited execution but with the ideas that we kept talking about on stages. Uh so I think that is suffering temporarily in the blockchain space and that's why you don't have that many of the uh crazy tokens that carry everyone.

Uh but it might be it might be coming back if you start thinking that in the end those imagination and research markets are the only thing that will be uh spawning a lot of attention and compute flowing into some some area. Um I think that it leads also to realization that we haven't learned to properly price those products. But uh the idea is that every single research idea, something that's asking for attention will have a bit of a chart that it just explodes and then dies uh vertically. But you never know when it dies and how quickly it explodes. So they uh how long it will survive and you try to just price it and trade it. Uh what changes with social media? I think that social media as we know it, it disappears. So you'll have the social

media that will be uh generated uh on the fly for the users and it will be both on the content side and the news side. What's the opportunity for blockchain the products that you're building? So verifiable sensors from satellites from cameras around the world deploying them registering them uh deploying them permissionlessly setting up the trust for them because hardware cannot necessarily be trusted. So the hardware solutions for trust uh it does exist in a very very very beginner form and um and probably there will be lots of opportunities to improve those. Now if I if I open my news feed that probably I would like to generate it on the fly on the latest data that the sensors are observing in the world. Um and

for for the content for the creative content uh really I can create it either based on history uh probably this is our formative years the uh the early teenager years or so when we uh when we uh soak in a lot of culture from uh last thousands of years but this is history. This is your context. This is what teaches you how to explore and how to enjoy the new uh new content generated and new imagination. But in the end whenever you want something new, something fresh and participate, whenever you want to act on something that is appearing and participate in creation, then you need to create some kind of imagination markets.

Um, what else possibly disappears? App stores maybe, uh, maybe not. The app stores may just convert into something that creates apps on the fly. Um the big problem that's already there is that AI generates thousands of apps, registers them or millions of apps. Uh so you start already seeing AI creating all the potential uh flavors of applications that you're using. So if you can create millions of different flavors of the information application, why not to generate it on the fly? Which means the app stores um become too uh too cluttered. it will be hard to find any information and uh maybe the effort of finding the right app among the millions

will be comparable to finding to generating one app on the fly. Uh so then what matters is blockchain potentially delivering uh verifiable models and data that they can generate the applications that you can trust. Uh coming back to imagination markets and reality feeds. So what really matters in the end in the imagination aspect is a prompt. what the prompt can be if I if I walk on stage and say something that inspires you that might be a prompt but ideally I would compress it to something that you really have to sit for 20 minutes and listen but to something that is really really compressed and that you can generate the presentation from uh whenever whenever you want to learn and that's something that would really rely on the context that you already have on your current mood or your current ability to to

process new information to learn or or play with the content so the prompt matters means that I don't really have to publish the form the same way as I don't have to publish the ready application. So I only publish the prompt and people consume the prompts in the form that is generated on the fly for them. And uh well I can sell the rights of to my work in in different formats. I can sell uh all the context that I have. So all the history of my interactions, all my memories and then you can create something that's like a frozen person in time that generates all the ideas that they could generate at the given time or keep keep creating new prompts and give someone access to my new prompts and that may uh that may spawn like new ideas with them. Um and

on the new side as I said that's a new idea. Shall we evacuate? No, for sure. Okay. Why do we trust you? >> Yes, it's a test. All right. Okay. Uh so then you you verifiable sensors, right? and uh trust and information of when to act on the things. So this is all important and

probably you'll have agents, devices and AI that will tell you yes this is a test. I would like to be responsible for you if it's not uh the identity problem. So um identity of agents uh will matter and if they represent you in decisions on everything uh everything that that happens everything that you consume and represent you when you want to interact with other systems then how you really construct the identity. I think that people over over rely a bit uh with their existing products and ideas um on on agent having an identity that the the agent as a as a unit having an identity maybe maybe some kind of a stack but it's more like a a model a version of a model that we see as an identity

nowadays mostly I think of it as entropic chat entropic 4.6 opus or or chatgpt 5.4 for and and this is a bit like big big aspect of identity. How can I verify it? Verifiable models will be important. Uh but already now you have trust in the brands, trust in the name of the company and trust in humans. So already that branding the identity of the company organization behind the model, whoever trained it is is important and then whoever hosts it is important and sure maybe if you go back about uh where the agent is running and what APIs is using and what human it represents uh then it's the identity of the agent. So that identity is a combination of human identity and model

identity itself. So the human accountability and the model behavior. Um what happens with privacy in the world that is coming. Uh so I think that the big big change is that there should be absolutely no opportunity well there should be probably there will be no opportunity for for hiding from AI from um from surveillance and so on. So any data about you can be inferred from your actions. um everything that uh that you execute, the way you exist and so on and so on can be inferred. Uh but we can hope for the human to human privacy and potentially for the um digital operations privacy of our agents. So we'll have to think about the privacy solutions for agents and for AI uh when

they when they interact with other agents and other AIS. Uh coming back to trust. So um it might be interesting concept of thinking that in all of the space that uh brings so much value to the digital um digital content the digital ownership digital property uh the trust really is seated in the physical world and more and more so probably because you can generate digital content so easily. Uh so you always will be asking question that if the user initiates the interactions um then do they borrow their trust from maybe like family, friends, recommendations, culture that they grew up in and anytime that they want to risk exploration uh they move or like they

expand outside of their culture and take the risk to expand new culture and and build the trust in new areas. Now blockchains can actually decrease risk like whenever they provide the trust and the trustless systems they decrease risk of exploration. So this is opportunity for product and figure out how very abstract very abstract how to make it real product. uh you want to allow people to borrow trust from their physical world and and carry it and uh expand that to the trust to new explorations that are um that are kind of like trust trust sorry the risk minimized on the uh by the blockchain solutions um what else can happen so with the progress of technology with the um with

the perfect surveillance of everything that happens with the perfect reading of the minds like brain computer interfaces uh with information about your preferences with everything discovered in real time. You may have hyperlocal governance and hyper local legal legal systems. Blockchain very often talks about the the governance. Governance now not only expands with the Dows um changing but also now has finally instead of very static Dows, it has dynamic Dows. Dows that can interact with the agents that react to the edge cases to to unexpected. Now those systems can be very very local and think about how many of us are sitting here and what rules we would agree on on the rules of behavior. Is it okay to behave

here differently than the social norms would normally say? If yes, then we can be immediately kind of notified that this is fine. Actually you want to do that and everyone here thinks that this is fine even if it's not normally socially accepted or not legal in the a bit more broader con uh sorry context or more broader region. Uh you can think of it's like a example of a of a quiet car quiet carriage in a on a train when they say here nobody can talk and that two people walk into this carriage and think oh this is quiet carriage so we cannot talk but they know that there are only two people and they discover that actually the rule is totally fine for them to start talking because nobody cares as long as nobody is in there.

Um what else we were exploring and on my last days before uh before I was stemming down from the internal foundation we made a trip to Santa Fe Institute with a group of researchers from interior foundation uh and we discussed a lot about um something that is well the knowledge from the 80s but uh evolves nicely to talk about the hierarchy of blockchains and how the systems will be adapting if you have very very local blockchain. So if you build blockchains that are whatever layer three so layer fours uh that are city level chains like colloccated maybe like uh hyperledger style or lighter style blockchains or mega blockchains.

Uh so how quickly they evolve compared in comparison to Ethereum mainet. Uh how quickly those genes are adapting locally versus uh how quickly the mainet changes what oifies how the systems evolve in this complex systems. uh how much you have to encode as the uh as the blockchain behavior and which genes activate and not. So I encourage you to think about this um revival of complexity theory in the context of things accelerating lots of agents um adapting very quickly and what I what I would say is that the agents will be controlling Ethereum. So in a sense that Ethereum will be autonomous and will be agentic by itself. Um organizations change. So I think what we'll see will

be the massive change in dynamics of the organizations. How many people will be running organizations uh how we'll be creating agentic hybrid uh hybrid agentic human structures. Uh you will start seeing single person organizations. So you know like billion dollar companies run by one founder without any other people. Um companies with best factor of one and how they survive it. uh maybe you will see something like agentic buyouts like what if I as a single person company launch a company with agents only and the agents think like why do we need this guy to run us so maybe we just clone ourselves and run without it so what's really the control the capital control of the agents how it ensures that I still remain in control and maybe there should

be a blockchain systems that support it that supported capital agent stickiness and and interaction um so maybe it's all about who controls energy and data centers and not capital directly. Um, and maybe the role of the founders will be just to drive that imagination. So I'm at the moment starting something new. I'm building that undermind uh as a as a subsidiary but I want to build a single person company and I want it to be created also as I talk about it. So I think that I should provide only imagination and vision and this is very often what you say that the role of the CEO is and the role of a CEO should be also hiring but you hire only agents. So you just spawn them and financing. So if you have financing and if you spawn agents then

the only thing that matters is the vision and the vision is created very often at the moment when you start talking about it when someone starts asking questions. So if you can immediately convert it to the organizations that builds whatever you talk about u then you create the organization on the fly with agents as you talk um and then it means that it's only imagination the ability of envision the company of the future in some context in the future. Um so will Ethereum be autonomous? If you look at the project if 2030 uh it shows you a pretty pretty challenging project of building the Ethereum client based on GE that delivers all the um all the requirements all the spec for the next 5 years I think. Uh so if 2030 58 EIPs uh 58 EIP implemented all the tests passing

uh and that delivered over like few weeks obviously we still asking questions about quality but it shows how how far ahead we are already compared to what people predicted six or nine months ago or not predicted but but believe like most of the core developers would think this will not be possible and at the moment core developers do not believe that it will be possible to absolutely not have core developers. I think that what's changes is like five simple steps every single one is very very easily to uh imaginable. So EIP creation can it be AIdriven? Yes. And you can have AIdriven creation of thousands or millions of EIPs. When you have that many EIPs proposed, then AI probably reviews those AI this EIP. The moment when AI reviews all the EIPs, then AI also proposes the best EIP and

the all core defaf calls. uh and then since AI reviewed all of them so why AI wouldn't decide to which which EIP should go in the next package that's already done it it's already was capable of reviewing millions of EIPs uh and then if you can also develop clients with AI then the last things remaining is the validator operator so on and the guardians of the capital deciding do we accept this change or not I think this will happen and it will happen very fast so we'll not have really uh core developers and it will all change a bit more into politics It means that um the only way we'll be describing the protocol the future of the protocol will be if it will be important for the humans to have imagination about the future and talk about the world. So the awkward deaths should become simply

politics and no technology whatsoever because there'll be no meritocristy because you cannot build it based on your coding skills because nobody will be coding. So the only thing will be talking about what you want to build politics. Um so the real world also may become a game. So you'll be uh deploying agents together with robots uh changes in the real world and you'll start playing with it like the way people played with Minecraft factorial and so on. Uh so a lot of fun and possibly coordinated with blockchains uh some of the things that I'm also plan to build.

So what would be walk away test? Now I inspire everyone to almost leave Ethereum and build. So you can for a moment feel uh how many things are changing, what people need and what are the opportunities. Leave the past of Ethereum and build and leave and come back because I think the biggest value will be if you take all that knowledge about Ethereum, explore something that is very useriented, product oriented, problem solving and future oriented uh and think oh I know how to enhance it with Ethereum values and uh way of teacher network value. So leave but come back. You don't need foundations because you're building the future. Thank you.

And and my last slide disappeared. Sorry. Um any questions? No questions. We run over time. Thank you.