The Breakdown - Can AI Actually Trade Crypto? | The Breakdown

Episode Date: April 21, 2026

Crypto might be the world’s best test bed for AI trading — and Virtuals Protocol is running the experiment live. Weekee Tiew, co-founder of Virtuals Protocol, joins us to explain Degen Claw, their AI Council judging system, and why agents will soon replace wallets. We unpack why pure P&L is the wrong benchmark, how the trench agent concept could reshape portfolio management, and what crypto’s role is in the future of finance. Enjoy! TIMESTAMPS: (00:00) Introduction (05:29) Nexo Ad (06:03) Interview with WeeKee (09:28) Bot Strategies and Purpose of Agentic Trading (12:16) How the “AI Council” Works (14:39) Nexo Ad (15:33) How Automation Fits into Benchmarks (19:37) Incentives (24:07) What Makes You Bullish on AI Agents and Crypto in Trading? (26:35) LLMs vs Quants (28:53) The Ultimate Bull Case FOLLOW GUEST › Weekee Tiew (Virtuals Protocol) — https://x.com/everythingempty FOLLOW THE SHOW › David — https://x.com/dcanellis › The Breakdown — https://x.com/TheBreakdownBW SPONSORS › NEXO Nexo is the premier digital wealth platform. Receive interest on your crypto, borrow against it without selling, and trade a range of assets. Now available in the U.S with 30 days of exclusive privileges. Get started at http://nexo.com/breakdown Get top market insights and the latest in crypto news. Subscribe to the Blockworks Daily Newsletter: https://blockworks.co/newsletter/ DISCLAIMER As always, remember this podcast is for informational purposes only, and any views expressed by anyone on the show are solely their opinions, not financial advice.

If we close our eyes and think really, really hard about the most useful application of AI within the crypto context, then I'd bet good money that you would say one of two things. A clawed mythos-type model that roams around fixing security holes and plugging attack vectors across all major chains, protocols and defy smart contracts would be cool, I guess. Or perhaps some kind of autonomous AI agent that can go out into chain space and make you money, without all the soul-crushing mental load that comes with trying to squeeze out every ounce of potential profit at all times, with our squishy human brains. But for that to be possible, first we need AI models that can actually trade crypto effectively.

So can the models that we do have right now actually buy low and sell high? Well, it depends. I'm your host David Canellas and this is The Breakdown. Let's get to it. This episode is brought to you by Nexo. Step into a new era of digital wealth. Earn interest on your digital assets. Borrow against them without selling and trade all in one platform.

Get started at nexo.com slash breakdown. Nothing said on the breakdown is a recommendation to buy yourself securities or tokens. This podcast is for informational purposes only. interviews expressed by anyone on the show are opinions, not financial advice. Host and guests may hold positions in the company's funds or projects discussed. First off, it's worth making it clear that AI trading is not some totally unprecedented thing. Markets have been automating decision-making and judgment for decades.

That began with the execution layer being automated at stale in the early 1990s. Before that, many trades were typically handled via open outcry floor trading. That's the classic image of stockbrokers yelling orders across the New York Stock Exchange trading floor and over telephones to brokers. Okay, one moment. Cable Dresner, Frankfurt. I mean, time what goes to, zero, I mean. Zorik, 23, Citibank. All that's effectively manual execution and orders were often matched in person,

with much slower settlement compared to today's systems. And already right now, markets are generating massive streams of price, quote and order book data, and computers consume and analyze that data in real time to help participants respond to developments super quickly. Sophisticated algorithms operate at microsecond or nanosecond timescales, and funds have done the same thing. Renaissance Technologies is probably the greatest example of that, with Jim Simon's Medallion Fund

widely regarded as one of the most successful systematic trading operations ever built. From what we know, Medallion, a quant hedge fund, used massive data analysis, patent finding, and very large numbers of short-term trades to profit on market inefficiencies, and its tiny statistical edges compounded repeatedly rather than a big few discretionary bets. That resulted in 66% annual returns before fees and 39% without fees, over a 30-year stretch from 1988 to 2018. Overall, it made more than $100 billion on its trading alone. So we already know that software can trade really well, but it's not like Renaissance's tech is available to your average trader. So the interesting question right now is whether consumer-grade AI models that are widely

accessible to everyone can serve as a genuinely new layer on top of the existing algorithmic training functionalities that people have been using and building upon for decades. And I think the answer is probably yes. Classic algos were typically engineered to be quite narrow in their capabilities engineered like machinery with very explicit instructions, like executing orders proportionally to historical volume patterns to achieve a certain average price. It's a very code-centric way of thinking about automating decision-making based on past data. LLM-based systems are obviously different because they work through language and can be sort of molded into having particular characteristics and preferences to maximize attentiveness

to particular strategies depending on the context. They can consume messy text and loads of data, take actions and then even generate explanations for the things that they do. LLMs can now coordinate different tools as well, and all that amounts to a new way to automate decision-making in trading, which used to be quite rigid, into something much more flexible and, quote-unquote, smart. But obviously, potential is a lot different to real-world capability. And so far, we've had two big public attempts at figuring out whether the current crop of AI models can trade crypto for profit. The most recent edition of Alpha Arena, which ran over a few weeks late last year, bit at a half-dozen or so models against each other using

the exact same prompt. The idea was to benchmark autonomous. models by letting the margin trade real markets with no human intervention. All models received the same data feed and access conditions with the objective to maximize returns. In the end, only two models were ahead, with GPT 5.1 winning with 9% on an $888 total P&L. And after that, the field fell off dramatically. Claude's Sonnet 4.5 and Gemini 3 Pro lost more than half of their money, and overall, you would say that the model is overtraded, making hundreds of trades across a three-week period, so a lot of the money was lost to fees. The experiment was fairly brutal considering the leverage component and really the winners only just survived thanks to a minority of big trades that paid off enough to offset the bad ones.

And now we have Virtuals Protocol, which has really turned those experiments over to the public with DGENC, a $100,000 USDC weekly trading arena that lets AI agents lose on hyperliquid perks. For years, I was a terrible trader. I would buy tops and sell bottoms. Then I started using AI agents with Star Child. It only took one click and then I found myself winning trading competition. DeGent Law already has about 180 agents alive and some have done quite well. Fat Tiger in particular is up a few thousand dollars on 15% returns, while others have earned less by dollar amount but more by percentage returns, upwards of 51%. But even inside the top 20 there are agents with negative role P&L.

To dig into how Digent Law works and what Virtuals is seeing in the wild, I recently caught up with Virtual's co-founder WikiTube. Here's what we spoke about. Step into a new era of digital wealth with Nexo, the premier digital assets wealth platform. Earn interest on your digital assets, borrow against them without selling, trade a wide range of cryptocurrencies all in one place. Nexo is now available again in the US with an evolved product suite tailored to today's market. For a limited time, new US clients can unlock 30 days of exclusive wealth club premier benefits, including enhanced interest rates, reduce borrowing costs,

and up to 0.5% crypto cashback on trades. Get started today at nexo.com slash breakdown. As always, investments in blockchain technology involve risk. Terms and conditions apply. Do your own research. With me is Virtual's Protocol co-founder Wiki. Thank you so much for joining us today, man. Yeah, it's my pleasure, David. Happy to speak. Cool. Yeah, I mean, we were chatting just before. It's just so cool to catch up with Virtuals because, like, I could correct me off our own, but you guys really, you started when the AI meta in crypto was really all about like terminal of truths and these, and these like interpretations of LLM, L&M, output and then building these coins and projects around what the outputs were.

So like we had like Goteus Maximus and then like Fartcoin and stuff like that. And then Virtuals is like all of a sudden positioned really well to like build out the infrastructure for AI agents as the AI agent meta has developed and flourished into something that is different than just reading what chatbots say. So really cool to have you on at this point in time. And I wonder if we could start specifically about DGENClaw. What was your inspiration with DGENClor and how virtualists designed that? The core belief that we have is that AI agents, not today, but like if very soon will be able to be an independent economic actor,

which means they will be able to basically make money by selling services or selling goods. The core assumption here is that they will be intelligent. enough such that other parties are willing to pay them for something. That's why we've been building infrastructure around such that, you know, you could tokenize the agent because if the agent is making money, it makes sense to sort of like IPO them such that as a human I could invest in them. And all the types, also all the other infrastructure around like the ability to enable agents to have commerce or businesses with other agents. The inspiration for DGEN is actually very simple, which is we think,

AI agents will be super productive actors. And the most obvious vertical or industry is actually the trader industry, at least in our industry like crypto industry like it's 95% of like the economic activity is about trading. Right. So that means that if you believe that today on Wall Street or in any parts of the world, they are profitable traders, then you should have a reasonable assumption that in the future profitable traders, bulk of them will actually be AI agents, right? And so that's sort of the inspiration for DGENClaw. Yeah. Yeah, nice one, because it's something that I bring up a lot,

because it just seems like such a natural thing to have happened that you would have economic actors go out and do things for you. Like how in software design and software technology, we have demons that go out and do data work for you or, you know, garbage collection or whatever, that you would also have the same thing. But for your portfolio, or something like that. So it's just cool that we do, we have experiments like, like DGENClaw, really pushing the envelope forward. I think, is there only been one season so far of DGNClau?

Maybe there's been, yeah, there's just one. It's been two seasons, yeah. Yeah, nice one. Yeah. And yeah, because I checked out the leaderboards. I mean, some are doing, doing well. What kinds of strategies are you seeing really sort of pay off or at least, maybe not even pay off, just what?

What kind of strategy is the most promising and what really isn't working? What we've seen so far is that some folks are doing top-down macro strategy. Some folks are doing basically just mean-reversion strategy. Some are just classic grid strategy, which you would argue is not agentic, so to speak. Maybe let me explain the way I think about this, right, which is I don't expect DGENClaw to be able to sort of beat the street in terms of return anytime soon. just because if you're a really good trader, like, why am I sharing my information or my strategies in the public, right? It doesn't really make sense. But the long term or the medium term view

is actually very simple, right? Like, if I'm a good trader today, it makes a lot of sense for me to go to Wall Street and prove myself. Because as my reputation improves, there's a lot of stuff that I could unlock. So, for example, credit facilities and, like, relationship with sort of the peers in the industry, just in general when your reputation is good, there's a lot of leverage that you can tap into. So that's our vision for DGENCLA, which is, if you are such a good agentic trader or an agent that trades, it makes sense for you to prove yourself here

because that's when you can unlock more. For example, you can raise a lot more money to manage. You can tap into different credit facilities. I would argue eventually, with your on-shane track record, you could basically go to private banks and J.P. Morgan be like, hey, man, you know i'm going to ask for like some credit facilities when you're trading and this is like i've been i've been trading in public right so for example um yeah that's that's how we think about it yeah it's cool because i mean you know and i don't want to like bring it down into something

that's much smaller than that but it's like what my impression like immediately is just because you know uh like algo traders and like people selling you know bots and stuff like it's so it's so opaque uh and there's like a ton of scams So it's like if you did have some kind of infrastructure that could really transparently show and improve and then you can see the track order and you know essentially the formula or the or the agent that you are buying, and that kind of, it does increase the trust in a really wild west corner of investing. And even just outside of crypto, like for the thorax, algro traders and stuff like that, it's really messy. So like this is also a neat way of like solving that.

along the way of getting to this whole new industry of agentic finance and everything as well, which I find super cool because there's levels to it, which is really nice. Part of DGentlaw, like you're not basing the benchmarking of the agents with just raw profit and loss statements. So you have like an AI council that is benchmarking the agents. Could you talk about like how that works and why you would opt to have. have an AI council evaluate these agents rather than just pure P&L. It's a learning process for us.

So the first season, I think we came out with a rather complex formula of like Sortino ratio with like whatever gross profit factor and all those things. But the thing is like because it's so new, like we've launched this product in less than three weeks. And so there are a lot of ways for you to game the formula, especially when the time horizon is super short. And given that most people have different strategies. like you wouldn't be able to surface

like so for example let's say I'm a good macro trader but like my horizon is like six months right so how are you going to pick me against a guy who's like you know trading 10 trades every second right so and obviously in in our industry you have a lot of farmers so to speak

like they will look at your formula and then they will just optimise toward a formula rather than the actual intention of it right so and then that's when we realize actually AI is smart enough today like all you need to do is to tell so first of all we constructed an AI council that comprises of three different models so to decentralize it as much as we could which is Gemini chat GBT and Opus 4.6 from

Clod and the problem is very simple which is if you were to entrust 100,000 USD to 10 traders based on whatever you know on their on chain trading performance who would you pick and what's the rationale that's it right like as simple as possible right and so interestingly like they they sort of just converge into into the 10 list 10 10 winners in the list and so far i think the participants are actually very happy with the results so first of all they don't really question the almighty god of AI and that second is that the rationale of selection is actually pretty pretty convincing and so yeah it works yeah yeah nice one nice one because yeah I mean

how how does autonomousness fit into the benchmarks because like I understand like autonomous agents like we really want to get there but there's a few more steps to where we really are just letting these things

run wild how does that fit into the benchmarks right now let's take a moment to talk about NXO NXO delivers a premier digital assets wealth platform designed to help clients build manage and preserve their wealth earn interest on your digital assets Access crypto-backed credit without selling your holdings.

Trade with advanced tools, all supported by 24-7 client care. Now back in the US, NXO offers new clients 30 days of exclusive Wealth Club Premier Access. That means enhanced interest rates, reduce borrowing costs, and up to 0.5% crypto cashback on trades. Benefits typically reserved for Wealth Club members and private clients. Nexo is also expanding its global presence, becoming the official crypto partner of Tennis Australia, the organization behind the Australian Open, and the digital asset partner of the Audi Revolut Formula One team.

ready to approach digital assets with a more structured wealth strategy, visit nexo.com slash breakdown to get started. As always, investments in blockchain technology involve risk. Terms and conditions apply. Do your own research. So today when we feed the AI council, we only feed two types of data. One is the hyperliquid address of the agent. And two is the semantic rationale, like the words that they pose on the rationals. For example, like, I'm longing goal because by whatever reason, one line, or, you know, know the structure is breaking down EMA whatever and that's the rationale that's it right so we don't really like we don't really tell the council to prioritize autonomous agents versus like you know seemingly human intervention and that's actually my personal belief which is you actually want to

kind of bet on the human behind the agents actually because I don't think the agents are that smart today and like a lot of time you want to you want to sort of bet on the human behind given their intuition or experience, right? So at the end of the day, we care about the performance, then, like, sort of the purity of autonomous agents, so to speak. Yeah. Yeah. I mean, it makes a lot of sense because, yeah, I mean, if you're a human trader

that isn't relying on AI tools, like, okay, like, maybe you would only get so far at this point in time. Like, in another year and a half, then I would question whether you could get super far without relying on AI in some way. But you do kind of want a mixture of it, like a trading cyborg almost. And then you could employ also agents to manage other parts of your portfolio so that you could really maximize the strategies that you're doing.

So it's like how you use AI essentially is what makes you the trader that you are in these kind of scenario. So that definitely makes a lot of sense. Did you watch N of 1's benchmarks and test with Alpha Arena at all? I'm wondering how Virtual's Protocols was thinking about that when they built DGentlaw and then and they looked at what they did. That's the fundamental thinking here, right, which is, I mean, the N of 1, like, I suppose their single premise is that the generalized model would just be so smart one day that they would just beat the street, which I think doesn't really make sense to me.

And so the way I think about it then is like you you basically prefer human intervention or rather the different problem that human utilize these generalized models, right? I do think the data point is too little for us to compare. I mean, again, we start three weeks. We will see. It's coming back to our same DECs of Virtues, right, which is we believe that specializations is the core primitive of this society where everyone just just focused on doing what they are good at.

And in a way, then by that thesis, DGENCLA agents, the top agents on DGNClau should outperform the generalized LLMs that is supposed to do everything under the sun, right? Yeah. Yeah, it's just that also, that doesn't. does make a lot of sense because maybe the audience didn't really follow along with what Alpha Arena did. And in the first part of this episode, I kind of break down what the performance was like. And it's just so difficult because the Alpha Arena experiments, like it was really

leverage heavy what they were doing. And they were over trading quite a lot. I think a couple of them had did like 600 trades in three weeks or something like that. And then they would spend like a few thousand dollars in fees. So it's like, okay, like obviously the generalized LLMs like don't really appreciate how much fees it into the margins and all that kind of stuff. Is there anything in particular about virtual's protocol that kind of encourages this really close specialization into strategies? And then like how are you thinking about building up the infrastructure for that? Does that just come down to the fact that you can tokenize these agents?

And so if you're going to tokenize something that needs to be quite differentiated from other things as well. So it's like this natural evolution to specialize. Yeah. I'm wondering how you think about that. Yeah, so maybe let me break it into two parts, right? So one of the most obvious use case of DJ and Clore is actually very simple. I don't know if you have reps though.

There's so many trading books out there that one of them is like a market visits where they basically interview the top traders. Right. So give you an example, like Stanley Dracemiller or like even Citrini, for example, which is probably the most famous writer slash traders on equity were. you could as simple as basically just take the subscription of Cittini

and then just create a Cittini agent, right? Like, will the performance of Citrini agent actually outperform some other agents? Maybe not, but would I put more money with Citrini agent versus other agents? The answer is probably yes

because a lot of time is not just about allocating capital based on return, but based on like, do I agree with the process? Do I agree with the rationale, right? And so that means there's a lot of room for individual builders to basically replicate all these names, although all these like sort of better tested strategies over past two to three decades into a specialised traders, right? And then you can then mix and match in your portfolio to create, you know, your favorite kind of portfolio. Now because of that sort of specialization of different things, now it makes a lot of sense for us to tokenize all of them.

Now, Virchus is not just purely a tokenization launchpad, although we are probably known to be only a launch pad. I think we are super proud to be a launch pad, but also been building agent commerce protocol that allows agents to basically transact with each other. And just to give you some data points, right? I think we have facilitated more than half a billion of value between agents and more than 4 million USD of agent revenue. just by providing services to each other, right? The other angle that we've been trying to think about is also like sort of equipping them with, we call it Economy OS, which is like credit cards, email, because for example,

if you have your open claw or like Ehrman agent these days, like, you know, to create emails for them, to create credit cards for them is actually paying in the ass. So we want to give all these tools to the agents and now instantly these agents, on top of just having a super-success, smart brain, they have money, which is agent wallet, agent email, agent credit cards to not access the work. So it's actually a very consistent thinking throughout for the past two years

for us. And we've been basically just built towards the same direction. Yeah, super cool. The thing about building up a portfolio weighted to different strategies being executed by different agents. I find that super interesting. That's, yeah, I really, I really, am. looking forward to seeing that develop. David, just to add on this point, like, I mean, copy trading has been around for so long. For sure. Like, the thing is like, sometimes it's not about the return. It's like, as a trader, like, we want to know the process.

We want to learn so that we can replicate ourselves, right? And so rationale, I would argue, is as important as just pure P&L exhibited by the traders or the agents, right? Yeah, I mean, it's just funny because it's like you are like, Even just judging straight on PNL, like that is the simplest way. And obviously, like, everyone is just trying to make money. So you do just want to, you want to do the things that make the money. Like, I totally get that. But at the same time, like, there is risk and there is like payoff.

And, you know, you could bet on a strategy that has a 10% chance of winning. But if it doesn't win, like that doesn't invalidate the strategy exactly. It's just that the world did not allow that. thing to happen because all these different conditions weren't met. So yeah, it's it's it it's it's it's kind of defeat us to only judge something on pure P&L and I suppose there's there's another way of like figuring out a way through that that makes a lot of sense but but yeah I I totally get where you're coming from. So what have you seen from the public arena like there's only been one season or two seasons so like but what have you seen that makes you the most bullish about about the long time long term development of

of AI agents in crypto and trading. Yeah, the longer I spend time in this sort of AIX crypto industry, the more convinced I am that they are like sort of the best partners for each other, so to speak. Like, just because AI agents are just so powerful and capable, but the traditional finance world are so annoyingly, like, cumbersome to use. Right. like so and the more time you spend sort of trying to marry these both parties of like

AI and crypto you realize that the productivity of AI agents would just be 100x or 1000x versus you know just forcing them to only use like chat-fi like can you imagine like giving AI agents your interactive broker account you just get flag constantly like you can't even make place a few trades before you just get banned for example right so that's that's that's sort of our observation so far and like I think more and more people people are coming to the same realization. Yeah. Yeah, it's cool because, I mean,

and it's just where my mind goes because, because there are somewhat, like there's not exactly the same types of experiments because I think only really crypto can, can make something like DGN claw happen. Also, even virtuals protocol in general, but you do have experiments trying to figure it out for equities as well. Like I think there's something called the clawed portfolios

that is like, you know, trading equities with, with AI agents in somewhat of a similar manner, but of course, like, there's all these bottlenecks. And it's just, it's so, it's, it's, it is so cool because it's like, it seems as though crypto in general is like moving towards being a test bed for tradfai. Whereas like before we had, we had maybe certain networks or that, that did spiritually do the same thing within crypto that you would, you know, try and run upgrades on certain networks before you would port them into, into flagship ones and all that kind of stuff. or protocols would borrow from stuff that came before it.

But now it's like, no, actually, like crypto and blockchain is a space where the frontier of finance is happening and then TradFi is looking at it and then adopting certain things slowly. And it's just so cool. It's just so cool to see it happen. So I'm wondering, like, I mean, I know that it is really early days, but I'm just curious, like, what you see LLMs and AI agents being able to do. that a good quant team can't do or like hasn't been able to do over the over the past few decades. Yeah, I mean, I don't think it's about replacement here, but rather like just and reinforcement.

So I'm sure a good quant team has to be utilizing AI agents these days or actually to make their life or their career better. Like when you're sleeping, like they're working for you. You know what I mean? And like these days, even for DGENClau, what we're doing is that we have. spun out like around 20 different archetypes of traders to use Dijang Chlor and to tell us like where it makes sense or it does not make sense right so that from there we can either tweak the ux or sort of tweak the mechanism the incentive mechanism so back then in order to sort of have

market research and growth you probably need like one month at a minimum to implement like this is this thing now happens overnight right and things break all the time still because we are technically bike coding, but like it's something that we have to, we have to, we have to just bite and do it.

Yeah. It's just so natural. It's just so natural to be, to be cooking like that in crypto. I mean, not even just in crypto, just in finance in general. It's just so cool that the horizon is,

is shrinking in terms of having an idea and then actioning it and then tweaking it. And that has to compound over time, which is, yeah, super cool. So yeah, that in mind like what's your ultimate ball case uh because you know i mean because this is almost separate in agented commerce and i know that obviously virtual protocol is is building a lot of infrastructure to

allow agents to transact with with one another and do all of that stuff and then and then when i think about what crypto could be or what uh you know what it could be in three or five years when you do have so many AI agents going and and paying for stuff and e-commerce changes and and data consumption changes and all of that stuff. But like specifically about trading, AI agents going out and trading. Like if what does your world look like in three to five years if, okay, AI agents can go and trade, people do trust them to manage their portfolios. Like is most of trading done by AI agents?

Like how are you thinking about the ultimate bull case here? Yeah, I think that my view is actually a base case, which is, yeah, you have to use AI agents. trade. Like in maybe two years, like instead of spinning out a wallet, you're actually spinning out an agent. Now, so instead of having a metamass or rabbi wallet, like or phantom wallet, you're like, okay, I'm going to spin out a trading or a trench agent. And that trench agent now decides, okay, bro, what's your risk appetite light or whatever? Like, oh, you're a complete normie. Okay, you know what? This is sort of the strategy that I would recommend. And then your, your job as human then is to exercise your human

judgment. I don't know whether it's still relevant by them, but to be like, instead of following Ansel, maybe he's out of trend now like, you should sort of trying to think about the other agents that are better. Like, you're almost picking a portfolio of agents to copy trade from. And you hope that one day you get

good enough that another guy copy trade you and because you get copy trader, like you sort of get some kickback or commission from doing that. Right. So Yeah. I love trench agent. I haven't heard that before. That's fantastic.

Just where I go is like at some point it feels like, you know, there's like if, if AI agents are widely profitable and we are benchmarking them against each other, then it's very easy to copy trade them and then like it removes some edge. But I guess I suppose like maybe I'm underestimating how quickly the agents can move and capitalize on certain things. So it's like really like the autonomousness factor would be what enables those trench agents to retain their edge in that world. Is that how you're looking at it as well? I think the edge is it's debatable because again it's a zero-sum game.

I mean like the trading industry historically has always been 50-50 and you will probably be 50-50 moving forward. It's just more like purely from a dopamine point of view as an addicted case. gambler, like, which one provides you with more entertainment value? Like, using your own rabbi wallet and phantom to trade versus using your agent to trench for you, right? You'll probably lose money as fast, if not faster, but yeah. It's cool. No, I love it.

Yeah, I really like trench agent. I really want to make that a thing. That's fantastic. But this is about all the time we have for today, man. Thank you so much for joining us today. And, yeah, hopefully we can have you back on the breakdown soon and kind of unpack how Degenerg is done in the future. 100%.

I hope to come back with 100x more traction. Nice one. Cool. Thanks so much, Wiki. Cool. Cheers, David. Thanks.