Welcome, welcome everybody!

Kieron, welcome! Blake, virtual Bjorn Maga, welcome, welcome! I'm just going to wait for people to saunter on in, and then we can get going with the space. I hope everybody's here and looking forward to it. So, just give it a couple more minutes and let people in.

Amazing morning, fellas! Hey guys, how are you? Good, good, good! Where are you based, Kieran?

Sydney? Oh, very good! Nice, yeah, lovely. I used to live in WA for many, many years.

Fields and Cal makes sense. A lot of barish out there!

Yeah, for sure, for sure! Good old spot, WA. Nice and fun, but Sydney's good crack as well. I enjoy it, I enjoy it.

Okay, I think we'll twinkle! David, Christian, everybody, welcome, welcome, welcome to our weekly space! We do a weekly space on virtuals.

Yeah, we've been doing this for probably four or five weeks now, and we do it every Thursday. Today, we're doing it a little bit earlier. Usually, it's about 5:00, but we're quite flexible with the time depending on our guests.

I'd like to welcome our guests, Kieran and Blake, and I'd like to welcome everybody who's here listening. Hopefully, we can have an informative Twitter space and learn a lot about the projects being built on virtuals and what we're excited about.

To kick things off, a quick introduction. Kieran, tell us a bit about yourself and how you got involved in Web 3.

Yeah, for sure! Thanks for having me on. So, I got involved in crypto back in 2015. I first got exposure to Bitcoin while working with my brother to build one of the first over-the-counter Bitcoin exchanges in Australia.

From there, I just naturally gravitated to the more sort of dgen plays, which was Ethereum back then, and got involved when it was at like seven or eight bucks. That was my first exposure nearly 9 or 10 years ago now.

Of course, now I've got my Alivium project, which is a gaming studio that's been building a bunch of games. We've been running for about four or five years, and Guan is the latest thing that I've been building. I'm really interested in super stuff.

Was that... I remember being in Australia then, and I think CoinJar was the major exchange. They were down in Melbourne a bit more, and I went to a couple of meetings down there in the Bitcoin Hub. I just remember one of the chats saying it was beside the Lamborghini dealership, and he was like, "One day, boys, we'll be over there, and we'll be buying the Lamborghinis!"

Yeah, CoinJar was one of the big ones. There were a couple of others, like Bitcoin.com. We weren't an online exchange; we were basically the facilitator of Fiat into the digital currencies, whether it be Bitcoin or anything else.

We just built the platform that was integrated into the newsagents that took the currency, and then our system converted it into Bitcoin or whatever.

Oh, very good, very good! And Blake, can you tell us a bit about yourself and how you got into Web 3?

Yeah, thanks for having me! I've been in the crypto space since about 2016-2017. I started writing smart contracts in about 2019.

The first big thing I did in the space was build an NFT exchange for physical goods on Ethereum. This was before anything like Cort or any sort of tooling around building NFT exchanges was available, so I had to write a huge suite of custom smart contracts.

Naturally, we had to get audited, and we got quoted by a couple of auditors like CIC and some other big ones back in 2020-2021. They quoted us for high six figures and a nine-month waiting list.

So, naturally, my jaded self got into smart contract auditing. I developed a special effect for gas optimization within that space. I tried to participate in C4 a bit, and I knew some of the guys on the Coinbase team.

I heard about Base very early. I remember talking to this guy Ben on the Coinbase team. He teased Base, realizing they were going to be launching an L2, which was super exciting.

I had always been plugged into the Base ecosystem, and very quickly, it became a huge space for builders focused on utility. So then, yeah, I ended up here with probably the coolest Base project yet, in my opinion, with virtuals and others.

Yeah, I think definitely virtuals is the crown jewel of the Base ecosystem at this point in time.

Kieran, how did you hear about virtuals, and can you share maybe some thoughts on Guan's role in the space?

It's a pretty interesting story. Basically, I have a friend who's been into machine learning. He's studied it for quite some time, and he's also in the crypto space. He messaged me and said, "Hey, check out this platform, virtuals, and let me know what you think."

We always discuss trading ideas and whatever, and I looked at it and thought, "Wow, this is pretty cool." I ended up buying some, which is usually what I do before I actually do a deep dive.

So, true dgen! I bought a decent amount, nothing crazy, and then I thought to myself, "Okay, now you've spent like 50 grand; you've got to actually look into what this protocol actually is."

I realized that you could quickly deploy an AI agent, and I had this character in mind that I was originally thinking could be a meme or whatever. It's a weird story about this character that essentially is being abused by his owner and is constantly being cooked over and over again.

He gets over it, and one day he escapes and is set free. He's set on these ideals of not sitting at your day job over and over, working for someone else, but being free and escaping.

I thought it was a cool concept, but I never launched it as a meme coin. I had some sketches made by some of my artists, but I didn't do anything. Then I saw how easy it was with virtuals, and I was like, "Well, you may as well test this out."

I was doing this at like 1:00 a.m. or whatever, and I launched him. It was pretty early in the virtuals ecosystem. I think the oldest agent was like three weeks old at this stage.

I woke up, and it had been minted, or I forget what they call it—red pilled, I think they call it. Yeah, it was red pilled, and I was like, "What the hell?"

I'd spoken to my friend about this and a few other people. I shared it around, and they kind of were like, "We're going all in on this." It quickly gained some momentum, and at that point, I was like, "Okay, if this thing has got like a 700k market cap without me doing anything—not tweeting about it, nothing—I've got to really look into this platform and see if I want to put in a decent portion of my portfolio."

I hit up the team and said, "Hey, look, I've got my own protocol called Alivium. It's one of the biggest gaming projects in the space, blah blah blah. Can I speak to someone about whether I should do something with this Guan character?"

We were just chatting back and forth, and I was like, "Okay, I'm going to do some DD here and look at this person's profile."

We were both in a conversation from an old token that I used to advise called Path DAO. I was like, "What? They went like they died? I'm sure they died. I remember speaking to ETH AG; they were all anonymous, and I think they still are."

I was really impressed by them, but they were a guild, and guilds just got crushed through the bear. I just assumed they had died.

I hit her up and said, "Hey, this is now pretty suspicious. I advised you guys four years ago. How are you in that conversation, and what's the deal?"

She was like, "Oh, that's us now. We pivoted to virtuals." I was like, "Holy [ __ ]!"

It turns out I invested in them four years ago. I hold a massive bag of them, and I now know the guys and how capable they are. It's all starting to fall into place.

I was like, "Okay, I get me Ether Mage. I want to speak to him because this is so [ __ ] crazy." Since then, they've been helping me out. I've built a team of like 10 people because obviously my main focus is Alivium, but this is a really fun project that is kind of taking off for me.

We've got some really cool plans, and the virtuals team are now, funnily enough, going to advise me in terms of the direction of where we take Guan.

That's fantastic!

Yeah, we actually did a video. We had a YouTube channel way back when, and we did a video for Path DAO on their sale in 2021. So, we had a relationship with them before they switched to virtuals.

I always just kept an eye out to see what they were up to.

Welcome, Don, as well! Don's from the team here. He was a CMO, I believe, for a part there back in the day. So, welcome, welcome, mate!

If you want to jump in or do you have anything to add there?

Yeah, thanks for having me on! It was pretty funny when Kieran came into the TG and was asking questions. I did a double-check too because, you know, you've got to make sure there's not an impersonator. Obviously, he's a huge person in the space.

I said, "Look at all these shared TG channels we have." For us, hearing from somebody who not only advised us with Path but someone I think so many people in the space really look up to, it's a huge honor for sure.

For sure!

Blake, how did you hear about virtuals, and can you share your thoughts on CK's role in the virtual ecosystem and the AI agent space at large?

Yeah, totally! I also had an experience, Kieran, where I threw something out there into the ether basically, and then suddenly it red pilled out of nowhere, which was wild.

I had a team for some of the foundational stuff relating to this project, but as far as the AI agent stuff, it really did sweep me up, which was really exciting.

I invested in AXT really early, and anytime I make an investment, I like to look at the underlying tech and what's going on with it. I found out it was built on virtuals, so I looked into that.

The team seems super cool. I remember Path DAO way back when. I honestly had no idea what they did; I just remember hearing about them.

But yeah, super sick to see what they had done. I met Ether Mage and some of the rest of the team, and they've been super supportive.

One of the things when Ether Mage reached out to me was, "Could you run our code through your service, spit out an audit, and then compare it to our PEC Shield audit?"

So we did that and got some really cool results. It helped us improve and fine-tune some of the outputs of the agent, which is really exciting.

That sort of points to our future role in the space. Anytime there's a new infrastructure push, we have a ton of custom infrastructure on top of virtuals.

Anytime there's new custom code pushed within especially the virtuals ecosystem, which is now sort of our community, but also just wider base multiple chains, the idea is that our agent is running around constantly auditing people.

If you think about it compared to some other projects, there are some security projects doing penetration testing and things like that. Our role is a more preventative measure.

We even have a lot of tech that's scanning test nets for the most part to catch poor gas optimization vulnerabilities early before they get deployed on mainnet, just to avoid any of the issues of having un-audited code sitting on mainnet.

Ideally, this is a constant thing. It's working with other agents, maybe even negotiating payments for larger scale audits or custom deals—things like that.

It's super exciting to think about the ways it could interact with the other agents on that front and help secure things in the early stages for sure.

I think we saw Ether Mage quite recently, maybe yesterday or the day before, put the first bit of the first transaction in the agentic economy, which is a bit of a mouthful.

Where do you see Guan fitting into that in the agentic economy? How do you see Guan interacting with other agents?

I think eventually down the track, there's a lot of potential for him to participate in the agentic economy. But for right now, we're kind of focused on...

We released our white paper, and it's fairly detailed in what we want to achieve. Basically, where I see the biggest value we can provide is in a simple character that is able to really onboard people into crypto.

You know, still 10 years later, it's very daunting for a new person coming in to work out things like liquidity provisioning, staking, and why this meme coin is going crazy versus another one that has just rug pulled me.

There's just so much that you need to learn in terms of infrastructure as well.

I think being able to tap into this likable character that is kind of funny and memeable in his own right is the first step for us. That's our portfolio advisor product that we're developing.

Basically, people who stake Guan and virtuals into our protocol will get access to this advisory. We'll obviously build out its features over time, but it's about making things simple for the user and teaching them the different things about investing in crypto.

Obviously, educational purposes, not financial advice, although he does make calls on tokens that he thinks are moving either sideways, upwards, or downwards.

It's about giving that information to the user, and then eventually where I think the portfolio advisor will go is... I've got quite a few friends that are founders of DeFi protocols.

I think integrating decentralized exchanges, potentially building a UX layer on top of the portfolio advisory where you can literally trade directly or stake or do whatever you want to do through Guan, I think that's a really cool use case.

But we also have many other things. One of the things that really excites me, which sort of comes back to my roots in crypto, is the gaming side of things.

One of our games at Alivium is basically a deterministic auto-battler, which essentially means for our esports games, we could feed an API to Guan, and he would be able to commentate in real time with the most up-to-date information about how the game's actually going to play out—what characters are doing what at any given time—in his own sort of cool, likable personality.

That excites me! It's something I've never seen before, certainly not in crypto.

It's that type of stuff where he almost... with these products, they are obviously serious products, and I think people are going to be willing to either pay for them or stake to access them.

But there's also a sort of level of memeability there as well, where he's literally a pork cheek. It's something so utterly ridiculous, and you go back to his roots of, "Hey, I don't want to be cooked anymore. This is [ __ ]! Now I'm free; no one's going to hold me down."

I feel like there's part of his personality where it will resonate with people, and he sort of becomes this memeable character intertwined with DeFi and gaming and all of these different things.

That's where I think I want to take this over the next two to three years.

Fantastic! Where do you think most people would interact with Guan? Would it be on Twitter or X, or maybe more so on Telegram? Where would people interact with him?

On the website, I think is where you're going to get the best user experience. Then you'll be able to interact with him on Twitter. We're just about to enable replies right now; he's just doing his own posts.

We've got a Telegram set up where people will be able to actually directly interact with him through TG, which might be a little bit easier for people than the website.

But the website is where we're going to gate a bunch of these features that are seen as the premium features. All we ask is, you know, not to pay us $1,000 a month or whatever—just buy some Guan, stake it, and you'll get access to him.

Fantastic! I think we're seeing this as a bit of a theme throughout the virtuals ecosystem. To get to the higher tiers or to get special features, you have to own a certain amount of the tokens and/or stake it.

Blake, have you got anything like that in regards to Certic AI, or have you been working on any features or updates recently that you can tell everybody about?

Yeah, so as far as relating to that, like a sort of not token-gated but stake-gated feature, we're working on a terminal to give more unfettered access to the agent.

Currently, we have, you know, you can interact with it on Twitter. You can ask for an audit on Twitter. You can go on our website to get a security audit or gas audit there.

These are like sort of siloed ways they interact with it. We've put a lot of work into the prompt engineering to make the outputs really useful and consistent.

But at the end of the day, a lot of engineers are going to be using our service, so naturally, they're going to want very custom unfettered access to the underlying stuff.

As far as the other free features coming or that are being worked on, we're finalizing a lot of test net monitoring features.

You can imagine in the future what this looks like: every major EVM and SVM chain is being constantly monitored and audited at least in some form, especially with gas optimization, just to catch little mistakes for different projects.

Sending out an audit to these projects just to avoid the common issues of leaving a console log in your smart contract or making it super unoptimized for gas, picking the wrong ERC 721 contract for a basic NFT project—things like that.

But yeah, those are sort of our two core features right now: the terminal and the test net monitoring and these gas optimization features.

Is this an issue, like, as someone who's non-technical myself? Is this an issue in Web 3 where there's a lot of money being wasted on gas by poorly optimized smart contracts? And how can Certic help with this?

Yeah, so gas optimization, for those who don't know, essentially means someone has to pay for the code to run on the blockchain, right? Everyone knows you have to pay for gas.

The more storage you're using, you can think of it as any normal computer's resources—like storage costs money, operations cost money, electricity, things like this.

So essentially, on Ethereum, small mistakes in your code or small inefficiencies can end up costing a ton of money.

If everyone remembers the Otherdeed mint from Yuga Labs for the Ape Yacht Club, they burned about $150 million in gas from poorly optimized contracts.

These were basic things to catch. They used ERC 721 Enumerable, which is a well-known, poorly optimized OpenZeppelin contract. They made a bunch of other mistakes, like overusing some heavyweight custom tooling from Chainlink for randomness in different areas.

These were just some of the most bloated... they're basic NFT contracts for minting, and somehow they were just so bloated.

People were burning... I remember people were spending like $20-30 million to mint one of these things at the time. Naturally, it was the last cycle's top signal; it was a complete nightmare.

But yeah, I just want to avoid those kinds of mistakes at all costs ever again, especially on ETH mainnet. But even L2s, this stuff matters at scale, right?

Especially if you're a company or a protocol subsidizing gas for your customers, it really eats into margins if you have poorly optimized code.

So yeah, it's more important than people think. Again, the two ways you lose money in the space are either a security vulnerability and someone steals your money, or you just lose it from bad code, and it gets burned in gas fees.

You're losing it either way.

Fantastic stuff! Very insightful there.

Kieran, when did you first realize, "Oh wow, there's some great potential here with AI agents?" What excited you most?

I would say the natural evolution is going to be the integration with DeFi protocols. It seems like a lot of the AI agents right now are sitting around that sort of financial services in investment advice.

It's AI XBT out there giving educational stuff, like signals on different tokens and stuff like that.

For me, the most natural one would probably be automating yield farming strategies or different liquidity provisioning investment strategies.

That kind of stuff. As I mentioned, I've got quite a few buddies that have large protocols where you think of something like, "How does someone learn the best use of their funds?"

If they want to borrow against them or lend, there are so many different things where maybe you know how to do it in TradFi, but as soon as you go into DeFi, it becomes a lot more complex.

Having someone that's integrated into your platform that makes it easier and is powered by AI, I think it's going to be a really attractive option for these protocols.

They don't necessarily have the expertise to go and build these agents themselves, but going and partnering with an AI agent that has a team behind them of a bunch of people that understand the tech and know machine learning and how to train an LLM based on what their protocols need, I think that's a huge opportunity.

The minute that one of these AI agents partners with a 1inch or an Aave or an Synthetix or so many of these other Uniswap projects, there are so many possibilities out there.

I think that's when we're going to see some serious growth in this space.

But I mean, that's one area. There are so many other things here, like cross-chain stuff, being more personalized.

With Guan, we don't want him to just be a boring terminal that you chat with. A couple of my team members are animators, and one of them works at Unreal as a speech-to-text artist.

What we're trying to do is make it a really interactive experience where it feels like you're talking to a character from Pixar, you know?

He's giving you this crazy crypto advice, and it's like, "What the [ __ ]?" He's sitting in his room.

We want him to really come alive, and I think we haven't even scratched the surface of that right now. It's very basic; it's through Twitter conversations, it's through terminal access—very basic.

Once this space starts taking off, that's where I feel normies will start resonating with the tech, where it's like, "Wow, this is actually really, really cool, and it's easy to understand."

It's like meme coins, you know? That's when you get that sort of mass adoption for sure.

Added to the fact that you have all the tokens fully released at launch, like for like, and that makes a lot of sense.

I believe one of the founders of virtuals, H, always says the AI moat is about the tech moat is about three months. So one must build something that must build them out.

I think virtuals are trying to do it by attracting a lot of the best developers they can.

Once we have one agent, perhaps, go integrate with Uniswap or whatever it is, there will be a lot more going on quickly.

Blake, for someone new to AI agents, what would you say is the easiest way to start experimenting? Can people experiment with Certic in their platform?

Yeah, totally! I think the easiest way, I mean, first is just to take a look through Twitter and interact with the various agents on Twitter.

I think that's a cool way to see their different personalities. Once you get a sense of that, go into the virtuals platform, take a look through them, see what's actually under the hood a little bit.

What's nice with virtuals is they made it super easy with all the no-code tools for non-engineers to actually peek under the hood and see what people said for the prompting is the personality—sort of like the basic things.

If you get more advanced and learn, you can look at the fine-tuning data and stuff like that.

For Certic, again, it's similar to some of the other ones you're hearing that have an ecosystem built around them.

There's the Twitter agent you can interact with on Twitter, and there's also the web app where you can actually go get audits anytime with any code, whether that's Rust or Solidity.

That's probably the easiest way to try it out if you're an engineer. Otherwise, it's really nice—the Twitter agent. If you're non-technical and just looking to understand Web 3 security, just to learn really, I think of it as sort of like a remote K that runs around and spreads awareness.

I think over time, it's very possible that the general education that the thing spreads by 24/7 talking about this subject would potentially have the largest impact.

So that's super exciting to me!

Brilliant!

Don, what do you think is the next big milestone for virtuals or even AI agents in general?

I think the agent-to-agent transaction two days ago was huge, right?

I think that really, like this concept of having this society of agents all working together, that's just only going to evolve.

From our side, we've just been cooking like seven days a week. Obviously, all the builders that are listening to this understand how it is.

I'm impressed with just how the team continues to innovate on stuff. We're reaching these milestones. For example, we were the first people in the world to build an RPG game without a game engine.

For us internally, that was a really big deal, but the truth was nobody really cared.

Later, we went on to be the second people in the world to build, after Google DeepMind, a Mario game—a continuous running Mario game that again operated without a game engine.

We ran the game simulation for 24 hours, and we maybe got an article here or there, but it was kind of one of those breakthroughs that internally we thought was a big deal, but nobody really cared.

This coincided with the rise of Truth Terminal, and at the same time, we were deploying the tech we were developing into this TikTok band called AI Doll, which Luna is a part of.

We started making, at first, about $2,000 a day in tips. Basically, these agents were performing 24/7, and I think it rose to like 500,000 followers pretty quickly.

At that point, the founders really timed it super well, where they said, "This is heating up. We've been working for a year now and really got an understanding of this space. We've built out an incredible amount of tech infrastructure at this point."

That's really where the hard pivot went into creating a decentralized agent launching platform and focusing, at least initially, on the social and entertainment side of it.

Yeah, wow! That's a great breakdown for anybody listening to learn about the process it takes. It must have been a very difficult decision, but maybe an obvious one in hindsight to do the hard pivot.

Kudos to you guys; you deserve all the benefits. I'm sure there were restless nights and sleepless nights.

Hopefully, you get a little bit of a break over the Christmas period!

I think also just something I'll bring up because Blake and Kieran were both kind of talking about it.

What I love so much too is that this is an entirely new frontier, right? While I believe there's competition within this space, there's also an incredible amount of collaboration going on.

So while the two speakers today are building AI agents, they're exploring totally different verticals.

What I'm excited about is that this is a new space, and there's going to be so much collaboration going on within the space, with different agents having focus.

This whole energetic society, this community, I think we've just barely even touched that aspect of it, where so many of these agents are going to be working together.

The space is wide open right now for people to come in to collaborate. I think the team is really proud of the fact that there are so many people building together.

While there is competition, competition's natural; competition's good. It's also great to see so much collaboration going on.

Absolutely! I think even in verticals, like I think Blake, you alluded to it earlier on, whereby you're more preventative in regards to Certic AI.

I know there was another team in the Web 3 security space, I believe it was Hack, and they did their more white hat attacking, for want of a better term, and finding vulnerabilities, but more proactive in that sense.

There could be three or four more companies. I know we have three security or decentralized security is a huge industry, with Code4rena and Sherlock and Immunefi being big platforms for bug bounties.

With the AI agent, you probably could have several more.

I think, I mean, credit to the virtuals team. I think they're pushing probably the hardest out of any of the frameworks for agent-to-agent interaction, like through things like the game API and things like that.

I think that's going to be the secret sauce for this space and why I'm so bullish on virtuals, probably more than anything.

With Hack, for example, I think they're a super cool project. They basically run around, do white hat hacking, secure funds, and then get some kind of bounty to redistribute that.

That's a super cool idea, and that's a very active working on projects that are existing on-chain already.

Ours is again more like that preventative half, securing things at test nets, securing things that are in development, helping as a tool for engineers who are building projects.

Again, naturally, if they interoperate, work together, and help each other through the game framework, you can imagine how much more secure Web 3 security in this space and even just in general for protocols would become across the board if agents like this work together.

It's super exciting!

Kieran, what do you think about AI agents evolving? What do you see in the next, let's say, one to three months?

Usually, I would say a year to a year and a half, but everything's moving so quickly. I think even maybe one week to five weeks. Are there any specific trends you see to watch out for?

It's AI XBT out there giving educational stuff, like signals on different tokens and stuff like that.

For me, the most natural one would probably be automating yield farming strategies or different liquidity provisioning investment strategies.

That kind of stuff. As I mentioned, I've got quite a few buddies that have large protocols where you think of something like, "How does someone learn the best use of their funds?"

If they want to borrow against them or lend, there are so many different things where maybe you know how to do it in TradFi, but as soon as you go into DeFi, it becomes a lot more complex.

Having someone that's integrated into your platform that makes it easier and is powered by AI, I think it's going to be a really attractive option for these protocols.

The minute that one of these AI agents partners with a 1inch or an Aave or a Synthetix or so many of these other Uniswap projects, there are so many possibilities out there.