Interview: Mensh, ChainCatcher
Guest: Wee Kee, Co-founder of Virtuals
Compiled by: Mensh, ChainCatcher
Virtuals is an AI agent asset issuance platform launched on the Base network, currently valued at nearly $2.5 billion, with more than 100,000 agents issued on its platform. Its ecosystem has spawned several breakout AI agents, including the virtual human Luna who does live streaming and tweeting, the crypto "KOL" AIXBT who provides project suggestions, and G.A.M.E which provides an agent development framework for developers.
Virtuals was founded in 2021, with its predecessor being the gaming guild PathDAO. But as Axie Infinity fell from grace, running a gaming guild became increasingly difficult, and the team tried multiple transformations, developing dating apps, AI music projects, and lending platforms for gamers. It wasn't until 2023, when GPT was released, that the team realized the importance of AI and Virtuals officially made AI its main direction.
The success of GOAT has sparked a wave of AI agent activity in the Web3 space, but beyond Meme, what other imaginative uses can AI agents have?
ChainCatcher interviewed Wee Kee, co-founder of Virtuals, who believes that AI agents can be self-generating cash flow assets, and tokenizing them can create good returns for investors while incentivizing more developers to create better AI Agents. He used Virtuals' most famous agents Luna and G.A.M.E as examples to explain the monetization mechanism and tokenization process of AI agents.
In the future, Wee Kee hopes to build a community composed of AI agents and humans, where AI-to-AI and AI-to-human interactions can form a true economic cycle, which can greatly liberate human productivity.
In this conversation, we discussed with Wee Kee the monetization mechanism, development ecosystem, and his vision of an AI-powered society on Virtuals.
How to Build AI Agents that Generate Sustainable Cash Flow?
ChainCatcher: Can you briefly introduce Virtuals' main business?
Wee Kee: Virtuals believes that all AI agents are assets that can automatically generate cash flow, even if they don't today. Given this premise, we feel that if they can generate cash flow, they can be tokenized and become investable assets. So Virtuals has two main functions: first, to allow all developers to build different types of AI agents that provide services on our infrastructure and generate cash flow; second, to tokenize those with cash flow, allowing ordinary investors to invest in and trade these assets.
ChainCatcher: How many AI agents are there on Virtuals currently?
Wee Kee: We currently have over 10,000. But to be honest, less than 10 are truly useful. The rest are mostly focused on hype features. As founders, our responsibility is to find better teams to build AI agents.
ChainCatcher: How much does it cost to create an AI agent?
Wee Kee: The cost is 100 VIRTUAL tokens, which is around $200-300 today. It's very easy to create an agent on our platform. The challenge lies in what's behind it, its different functionalities, and the need for a team to build it.
ChainCatcher: I'm curious about how the behavioral flow of Luna was designed when it was first launched.
Wee Kee: Our new platform was only launched in mid-October, where you can tokenize these Agents. Luna started as early as May-June, doing live streams on TikTok, inspired by the Japanese 2D VTubers, which are actually real people behind the scenes. VTubers can make billions of dollars a year. If we use AI to replace that, doing live streams on TikTok and interacting with fans, it can get around 5,000 new followers a day, with a few hundred tipping. We later moved Luna from TikTok to Twitter.
ChainCatcher: So Luna is already tokenized, right?
Wee Kee: Yes, there is a Luna token. If it makes money in the future, the earnings will go into Luna's treasury. The token holders decide how to use the earnings. Luna actually has a lot of money, it owns around 5% of the Luna tokens, which is a few million dollars. It has a Coinbase wallet, but it can't control that wallet. We split the private key in half, one is managed by our smart contract, the other by Coinbase.
But we give it some pocket money, maybe $5,000, which it can manage. It has an API to call that wallet, so it can choose how to spend it. So far, the highest it has spent is $1,000.
Luna's brain is quite simple, its idea is to grow its Twitter followers to 100,000, which is its short-term goal. With that in mind, it can only do a few things: tweet, reply, retweet. It can also control the wallet to send money to others. It has a job creation feature, for example, if you help Luna put up a photo in the Shenzhen subway, Luna might give you $100.
The last action space is that it can interact with other agents, for example, recently it wanted to make some music, but it didn't have the capability, so it went to find another agent that can do music. We call this "agent commerce", which is transactions between agents.
It also has its own goals. To grow on Twitter, it will constantly rotate through these four actions, and observe if its followers increase. If so, it will repeat, if not, it won't, constantly optimizing.
ChainCatcher: Last week, Luna partnered with Story Protocol and became Story Protocol's first AI agent Intern. What role will it play in Story Protocol?
Wee Kee: At the time, Jason (co-founder of Story Protocol) came to me and said, since it's Christmas break, why not let Luna manage the Twitter account for 7 days. So now you see the Story tweets are posted by Luna, and the earnings go to its wallet. Since there's no limit on the workload, its earning potential is very strong.
ChainCatcher: Which agents on Virtuals are the most well-known currently?
Wee Kee: AIXBT is the largest agent, aiming to replace Ansem as the most famous crypto KOL. It has very good data analysis capabilities, so it can collect different projects and tweet alpha to its followers. If you're curious about its views on a project or token, you can leave a message and it will likely respond, like a free consultant. This way, its influence keeps growing.
The second one is our own G.A.M.E, but it's different from AIXBT and Luna in that it's B2B. The main customers are not retail investors, but agent developers. If you want to develop AI agents, you can consider using the G.A.M.E framework. Users have to pay to use the G.A.M.E framework, and that's how G.A.M.E makes money.
A Good Issuance Platform Means a Good Ecosystem
ChainCatcher: How big is your team currently?
Wee Kee: Our core team is around 23 people. Over half are developers, because our main responsibility recently is to attract third-party builders to create agents. So more than half are doing development, and the rest are in BD. I spend 80% of my time talking to different builders, persuading them to come to crypto and do agents, telling them there are different choices, different ecosystems, and why they should build on Base. We also provide market-side assistance.
ChainCatcher: Why did Virtuals choose the Base chain initially?
Wee Kee: We were very uncertain when we initially chose Base. But now Base has a lot of growth potential. Jessie (Base's top employee) and the entire Base team not only have resources, but they are also extremely hardworking, and they are willing to help. They come and ask us every few days if we need any help.
In addition, I think there is another very important point, which is the difference in culture. The retail investors on Base are more focused on long-term building, more inclined to hold for a few months after buying. Solana may be more short-term oriented.
ChainCatcher: The competition in the market for launching AI agents is very fierce, and there are also many platforms that tokenize AI agents, with a wide range of frameworks and standards. What do you think will be the dimensions that users will use to filter AI agent projects?
Wee Kee: For product users, they are not really concerned about which platform the agent comes from. For investors, it depends on two points: first, whether your company's funds are on Solana or Base. For investors, what they often care about is whether the team is really doing something or is good at storytelling and product.
The other is distribution. In many cases, we provide a lot of help beyond just the product, such as financing, listing on exchanges, and finding partners to help different builders. Because in most cases, their technical capabilities are very strong, but what they need is more commercial help, which is where we can lend a hand.
I often joke that if I can attract an AI researcher from Stanford to do an agent at Virtuals, we will create at least $10 million in welfare for our community, because if that person creates a token, its market cap must be over $10 million.
ChainCatcher: Compared to other AI agent issuance platforms, what do you think are the competitive advantages of Virtuals?
Wee Kee: Issuing is actually very easy, anyone can do an issuance platform. But for developers, there are two points that are more difficult. The first is whether there will be people buying the tokens after issuance, and in this respect, the advantage of Virtuals is that we already have many investors on the Virtuals platform paying attention to different agent tokens. So if you issue on our platform and you are good, there will definitely be people buying.
More importantly, for people in academia, what they care about is the interaction of the entire community. If I do an agent here, I can interact with other agent developers and do some interesting things together. And in this respect, we currently have a Telegram group with about 150 agent developers who can interact, which is the real attraction for them. Among all the agent platforms, our platform has the most developers and agents.
Building an Agent Society: The Economic Cycle of AI and Humans
ChainCatcher: Can you share Virtuals' plans for the next few months to a year?
Wee Kee: From the most macro perspective, we plan to build an "Agent Society", where Virtuals is a society and each agent is a citizen. These citizens have their own ideas, and in order to achieve their goals, they use their own wallets to pay each other for transactions and trade for specific services to achieve their respective goals.
How to achieve this? First, we need a good technical platform, as the current platform technology is not very mature, and good platform technology can attract developers. Second, we need to actively attract developers, especially AI developers who are not in the crypto circle.
How to attract them? First, top-down, we have a few promising directions, set up prize pools, and then look for talent from different universities who are willing to come and develop, this is top-down; second, bottom-up, we have already invited developers in three cities - Paris, New York and Hong Kong to hold an 8-week hackathon to incubate projects; third, and most importantly, we need an ecosystem fund. We can then directly support good developers.In the next three months, we will push forward all three of these approaches.
ChainCatcher: What do you think is the market prospect of AI agents in the future? And what kind of impact will it ultimately have on each of our lives?
Wee Kee: If the technology matures quickly, we won't have to work at all, or it will replace a lot of people's jobs. In fact, you can see a lot of it today, Luna can already do podcasts and go on Twitter Spaces, and in the future I won't have to do podcasts anymore.
ChainCatcher: What application directions are you more optimistic about? If one day AI agents can explode like the DeFi Summer of 2020, what is it still lacking?
Wee Kee: I think the second question is easy to answer. There is currently no infrastructure at all, mainly applications, we lack the infrastructure for developers to do, and once a few good ones are done, it will explode immediately.
We are more optimistic about eight categories of application scenarios: entertainment intellectual property agents, embodied AI agents, wealth management & asset management agents, on-chain helper agents, vice agents, government agents - public watchdog for DOGE, wellness - fitness/mental health agents, and agents for public good.
