Ethereum and Robots: New Asset Class

Transcript

Yes, I am founder at Netamines, but also now I'm founder at Complete Robot. It's it's something fresh. It's a fully subsidiary of Netamine. So, if you think that I'm leaving Netamine, I'm I'm not. So I'm just creating a slightly slightly more concise structure underneath that mind to focus on a very new style of organization. So I've been talking a lot for the last um two years I think about agentic AI about robots about how Ethereum will contribute to coordination of agents of robots how be solving the questions of identity and so on and uh and I thought also I've heard a lot that we should support builders be with builders and think about the real world

use cases of Ethereum. process of how to how to prove it better than to just jump in and understand all those challenges that otherwise you're skipping because I would keep people telling people that it's easy you can just do it now when I'm doing it you know it's it's slightly harder sometimes but it's super exciting um so I want to really focus on the ideas of where uh where the financial markets and Ethereum might be going towards understanding where the opportunities appear and what can be the asset classes the formats on Ethereum um and where it is. So, uh AI agents, compute, robotics, so the embodied AI or physical intelligence or recall it uh

lots of talking about automated payments uh through the AI agents, agents to agents, but also machine to machine payments. uh agentic commerce you see probably some of the best best results and uh most promising uh simple cases where it's immediately implementable is the agentic commerce uh but the compute metering and payment for compute uh recently so I've been working with the uh well I've been building my my startup at the HFZ residency recently at uh at San Francisco and there there is also the blockr run AI team and they do kind of this like um onchain compute metering and agents using payments um the machine to machine microp payments fleet billing

and actually the direction of fleet billing and self-deployable businesses is where I'm where I'm building now so thinking of the economy that you imagine kind of without humans not that there would be no space for humans but if you take it to extreme and imagine how the economy would operate fully autonomously with agents with robots uh what kind of changes to the structure to the mechanism would be required and what needs to be built now. So that's that's where we can all be going at least as an experiment and then it's fantastic to ask questions about uh where Ethereum fits all that. Um thinking about all the recent things that uh that I'm observing on Twitter. I'm not sure you know how it is when you when you fall into a rabbit

rabbit hole and you hear only about that one thing and then you start assuming that everyone sees all of this as well. But you may have seen it. So this is a figure I robot competing with a human in the in the package sorting and they run them over 100 hours and that the human one they say maybe it's the last time the human wins. Uh and I like that comparison that the robot can run 6,000 hours per year. The human can run 2,000 hours per year. Then I started realizing that some of us are robots because we are probably closer to that first number. Um but it's it's a lot of questions about what really happens with the labor with where the adoption happens because probably all of us think the robots are produced manufactured in

China. China is leading um in the US uh there is lots of movement towards building robots in the US. But now when you think where you can apply robots is everywhere where the cost of the robot uh is cost of a robot work over hour amortized over time over hour uh is lower than the than what you're paying humans which means that actually when you think about the US labor market it's very expensive so it might be that you'll have much more opportunities to start introducing automated work and robots into the uh into the labor force and then it opens a lot a lot of questions about what comes with it like does change the insurance market, the regulations, the opportunities for financing. Uh can you start operating

robotic fleets also as asset classes? Uh it can be not only robot taxis but also the paid per hour uh robotic work. Uh and you see some of those diagrams. So they they suggest that look over time the cost per hour of the robot amortized like the production cost will be will be lower and lower uh and probably the human labor cost will be going up. This is interesting assumption because maybe the challenge of the economy of the future will be that the the price of the human work will be dropping or as we start specializing we'll be specializing in the areas where robots absolutely cannot help us uh and the cost of that work hour will be going up. So some something there it meets and it means that there is some equilibrium level uh and it's it's look to it's good to look

at the comparative advantages of different countries and our skills as humans. Um this is robotic density. Actually when you look at uh China the density is not that high but when you look at the absolute numbers the number of deployed industrial robots in China are definitely is definitely above everything else. So uh when you look at Korea, Singapore, Germany, Japan those are leading in how many robots are deployed deployed compared to the uh manual labor in manufacturing.

Um these are the orders uh of the robots in industrial space in North America. So as you see actually there's lots of potential because this is not growing but at the same time the technology advancement in the last uh last year or two years is bringing a lot of excitement to the robotics community uh builders community manufacturers saying something moved. This is the most exciting time ever uh for many people that have been in the space for 15 years or more. It's not coming from me. I'm super fresh to that space. Uh but that's what I hear when I'm catching up when I'm listening to podcasts and so on. Uh so these are absolute numbers. So we see here as expected China is uh absolutely

in the leadership. uh but I think when talking to the uh to people in the US in the west that kind of sense of competition of trying to to show that we can that we can rebuild the manufacturing of the west uh is definitely very exciting and this is my trip from last last week I went to Detroit and uh I had a pleasure to meet with with the team behind BCON manufacturing uh and also visited new lab in Detroit they also build the uh new lab in Brooklyn here and and in New Orleans. So these are spaces for people to to do the like incubation accelerations for startups to join and like in this co-working we work style also be able to use the machines

and do manufacturing and testing drones and machines and robots and so on. There's also that re-industrialist movement in Detroit. Look, I've never been Detroit in Detroit before. Somehow when I followed the social media, my imagination was I'll go to Detroit and it will be just a collapsing city where nothing happens. But the energy in Detroit was like oh wow wow great this is the space which will be attracting people meeting in person. uh there's maybe even some idea that the blockchain community that laughed so much to travel and to meet in person now and facing the idea that AI is just dominating everything that is digital there is almost no reason to meet in person because you can transfer all the knowledge without talking to people but when you want to produce manufacture and

do things physically you have to meet you have to go to spaces where there are manufacturer uh manufacturing capacities and Detroit seems to be really full of energy and I encourage you to go to those spaces and New York as well definitely a revival of industrialization and around in Boston and Detroit uh San Francisco South San Francisco here you have the uh maybe you've heard Westm is starting to produce the uh engines the actuators for robots for drones and humanoids in San Francisco and South San Francisco. So there is definitely definitely that uh energy of wow we're building we'll show it we'll do something impossible we'll do it from from zero uh here in America from zero there in Europe um and and

this is very uplifting like the fact that also in Europe you have lots of uh humanoid robots or the cobots starting uh startups great thing to do uh now think about blockchain to robot and robots um I mentioned here two projects Because when we talk about uh blockchain, we mention very often the uh notion of identity. Uh so identity systems for humans in the context of when humans interacting with agents or robots. Uh I mentioned two projects world and boop. Uh boop is a project that was uh this was kind of incubated at Nethermind. Uh the verification of robot actions. I feel that every time when I think where is the gap now for the models for training

robots and for operating robots you you hear that it's very fresh space to analyze the robot failures and robot coexistence in the in the spaces with humans there's lots of mistrust now between manufacturers from different countries so if you think would us allow millions of Chinese produced robots on the streets in the US and vice versa Um suddenly those questions of verification, verification of actions, uh verifiable training, verifiable data collection, uh might be bringing opportunity. Uh maybe not in the day-to-day simple robots use cases, uh but for sure in some more uh security related operations. Uh you will have it.

There will be opportunities for the blockchain skills to be applied. uh the yield tokenization is something that I feel maybe refers closest to the uh to the topic of the talk the the new asset class. So in the space where you have the robotic fleets like to robot taxis uh drones for delivery for logistics uh drones for defense systems and so on and so on. you may start thinking of them as as packages to be deployed in different geographical locations and and treat them as businesses that generate yield.

Um there is maybe one one thought for everyone to think of that we very often say there's opportunity for microtransactions with agents with aentic AI. We also say that blockchain also solves identity and there is a bit of a friction between the two. If we solve the two properly then when you have identity which very strong then you add reputation to identity and then the question is do you still need micro transactions to the same scale? No. Because identity or reputation allows you to do to provide solutions the same way as for people the uh subscriptions the upfront payments or like the subsidizing for a while and then subscription to follower or or payments for per month credit and so on

and so on. uh when you want to introduce microtransactions this is all about paying before the action not trusting at all and introducing least friction so not even verifying identity but with microtransactions then in this models you have problems with KYC with AML solutions and so on because those agentic or robotic actions that are executed at scale uh suddenly might be money moneyaundering or or simply executing something that is very very much not desired by the society Okay. Uh so I think there is friction even as we see oh blockchain solves exactly these two topics. Um it's not that those two topics coexist. Uh they well they do coexist but they don't synergize super

well. They more like approach the same problem from two different directions. Um you're generating robots. I mentioned robot taxis and when you in the in the past years I was saying that very often whatever we in the blockchain space announce as a big vision for all of us um like people particularly when I observe Tesla they almost like strip down all the decentralization aspects and deliver the same type of product. So they say, "Oh, we want to promise the centralized social media and uh and like Elon Musk would say, well, let's make it centralized but open source and unbiased." Uh whether you believe in it or not, but the vision is described the same way, just not decentralized. Then

when I say robot taxis can be decentralized and autonomous like decentralized physical infrastructure networks deepen um Tesla would say well let's do taxis that can generate yield for you but they'll be centralized and so on and so on and so on as the same would be with with Starlings or with the energy and batteries. You'll see that almost every project that is delivered by those companies is also offered as the fantastic blockchain vision with decentralization and we tend to definitely suffer from executing uh not at the same level most of the cases but also we have the challenges.

So you can think of how to do it right in blockchain robot axis delivery fleets uh mining autonomous maybe dark factories cleaning robots agricultural legacy uh independent decentralized farming uh autonomous farming something that I've been looking into uh recently as well like what are the opportunities it kind of exists like in a centralized way uh but in the open-source uh fully collaborative way I think there's great great opportunity so We we buy land, we like to play those things like in NFT space and so on to buy something and experiment together in a fully crazy open-source way. And I think that that opportunity to play with the physical

world like buying buying land and playing with robots and deploying them and financing that through blockchain and experimenting together on the uh design and recipes um will be something that'll be coming to blockchain space for sure. uh yeah and maybe a summary what we've seen so far in blockchain. So the deep solutions where you practically deploy uh hardware into the physical space and you participate in the network so kind of investor or participant in the network is also operator and the token rewards are distributed in the networks uh helium fabots and so on. uh the token price. As we've seen very often with this projects, there is a a very very strong mismatch between the like

misalignment of uh misalignment of incentives and uh keeping the the founders and contributors long term. And there is also a feeling of uh of unfairness in the in the tokenized solutions nowadays. very often because of regulations not allowing for the tokenized solutions to carry the same fairness that you would see uh most of the cases for the for the traditional finance. I I definitely feel that the the blockers that pushed many of the blockchain solutions into the very tricky uh designs are not allowing those solutions to really flourish the way they could.

uh infra where you introduce like lending supporting with stable coins examples of uh USD AI which uh which is a depth that is backed by the GPU fleets and if you think that GPU is kind of a machine then you can also think about the fleets of the robots executing labor um or or any industrial robots robot axis and those can be backing the the debt issued in stable coins. Now the real assets tokenized this is exactly this. So the asset class where over time you can simply look at the the physical space at the globe and think where I would like to deploy a new farm

where I would like to deploy a set of robot taxis. Ideally I would select a factory of this robot taxis and uh say please produce for me a fleet of 100 taxis and deploy them on the streets and they'll start generating value for me. Then you have no intermediaries at all. You have straight to market production, straight on demand production. Wherever capital flows somewhere, it says please produce, I deploy and then I start uh collecting yield. So, so that's how I see the fully automated ideally permissionless decentralized space. I think this is very much it's wishful as the permissionless deployment in the physical world most

likely will not be allowed like so you you can expect very heavy regulation in the space and now it's this multi-year battle for sure of what can be allowed to be tokenized to be processed for blockchain. Um maybe some few more comments around this. So when you have robots, they do a lot of processing. They use a lot of tokens uh for the uh for their models of analyzing the reality, analyzing their mistakes, their behaviors, planning and so on. So the token use keeps growing for the agentic AI, but it also will keep growing for the physical AI. And uh this has been mentioned by many that um token use becomes this asset class that can be tokenized as well. Well, tokens

can be tokenized. So uh this is almost like pure on energy level. So maybe that's coming back to idea of bitcoin choosing to mine either bitcoin directly or the intelligence through compute. So it's either compute or money. Uh and everything starts connecting on the most basic level. Uh and then there's question of how it connects with capital through staking. uh so I feel there is there is something that can be very well described through research papers on the economy of choosing where to allocate and what's the role of capital versus compute once again the mention of this one um a few like not so recent developments but I see after coming back and looking at

X42 in the last weeks I see that more and more companies are joining it um it has it seems to have some traction obviously it's it's not the the level of hype at the time when it was introduced in Engentic AI but I think it has uh enough of the adoption to to show that at some point it starts being much much more widely and naturally uh accepted the the levels of like transactions per second that has to be processed by the systems that you'll be using X42 definitely demand some scaling solutions so there will be layer twos or so um and once once again I'm referring to the company with which I'm at the uh at the residency the blockr run so they

created this Franklin agent uh that does the payments and manages wallet so uh it's great to see that vision first like experimenting trying dying out a bit trying again you've seen probably massive success of the Hermes agent that came from the blockchain space from the cryptocurrency world then decided to be very very AI but you can always think that that there is some DNA of the team thinking of how to how to use it blockchain way, how to respect the privacy and so on and so on. That's why last time when I was speaking I was saying try to leave blockchain for a while and come back carry that culture and DNA uh to the space of agents and robots.

uh and I think in the future for us there's a complete solution of agentic allocation capital autonomous detection of sites where to deploy things when to construct things deploy operate them through agents and deploy deploy deploy recursively so that's what I'm focusing on now thank you so much for listening