作宇：@0xAikoDai，Folius Ventures
MyShell launched the badge points system on January 26, making AI agents, chatbots, and apps on all platforms open and investable assets. Currently, the team is still busy iterating on the product and economic system, so let me explain it in a nutshell.
If there are readers who are not familiar with MyShell, you are welcome to visit the official website to experience the product. In general, it is an AI agent/APP creation platform that covers mainstream closed-source large models and a large number of open-source models. It is currently the most user-friendly and integrated platform in Web2 and Web3.
I would like to reiterate that MyShell is first of all a very practical AI creation platform, and secondly, it has an open creator economy. Although this article focuses on its economic system in the Alpha phase of testing, I still recommend that every user who is interested in the platform experience the product and create an Agent/APP/Bot by hand. You will feel how an excellent web3 economic system will empower excellent AI products.
In this article, I will mainly explain:
1) Points system for MyShell Alpha badges;
2) Analyze the model’s contribution to the creator economy and the team’s long-term design intentions;
3) Imagine the future expansion of this economic model and ecosystem.
1. Analysis of the Badge Points System
Agent/APP is the core interest-bearing asset on the MyShell platform. Every interaction with Agent/APP consumes "electricity" (a kind of paid points). Each Agent/APP will charge different amounts of electricity for each interaction, depending on the language model, voice model, and image model it uses. For example, an ordinary chatbot (which only uses GPT3.5 and the model fine-tuned by the author) will only consume 1 electricity each time; while an image generation robot, which uses other open source language models and the official TTS voice model (OpenVoice), will consume 13 electricity each time. In this regard, MyShell implements pricing for Agent/APP and its interactions on the platform. Each publicly released Agent/APP is effectively used 10 times and can be allocated 100 shell points. These points will be distributed to agent creators, model creators, and agent investors in proportion.
For creators
1) Agent and APP creators can not only use a variety of models on the platform to create highly user-friendly works, but also directly obtain the first share (worth 10 Shell points) and the income generated from future long-term interactions after publishing the model.
2) Even if the author of the open source model does not need to build a product on MyShell, as long as it is inherited, he or she can directly obtain point rewards from MyShell, and there are hundreds of thousands or even millions of real users who actually use their models.
3) The rewards earned by creators in the above steps can be used to purchase pass cards to obtain more opportunities to create public Agents/APPs, creating new assets worthy of investment for the platform; or used to invest in other Agents/APPs and become investors.
Specific investment process
1) Buy: The initial price of each Agent/APP is 10 Shell points. Every time one more person buys this asset, the price will increase by 10 Shell points. In order to prevent wash trading and malicious price manipulation, investors can only buy one Agent/APP per hour, and a user can only make three purchases on the same Agent/APP.
2) Earn: In the process of holding an Agent/APP, all investors can share 40% of the income generated by it when it is used by other players. The longer the holding period, the more income the investor accumulates on it. The payback period of investors is different because some people buy at 50 points and some buy at 500 points.
3) Sell: When the user decides to sell, he can redeem 90% of the cost, plus the dividends obtained during the holding process. In addition, the quota of the early investor will be vacated, and the price of this quota is often half the current price of the Agent/APP or lower. Other investors may find and grab this opportunity at a preferential price in the market.
Therefore, the entire investment process can be summarized as follows: buy a potential Agent/APP, actively promote it and let more people interact with it. Then wait for your passive dividends to exceed 10% of the cost, then you can start to earn net profit. When you feel that the profit margin is declining or want to invest in other new potential products, you can choose to sell and take away 90% of the cost + all dividend income. When other investors find your favorable cost price, they will choose to invest in this asset and promote it again, so that it can gain a new round of popularity and dividends.
Next, I will explain how this economic model motivates participants in the ecosystem.
2. Features and advantages of the economic model
The entire economic model design of MyShell is based on its purpose. Rules are not enough to highlight the health and long-term characteristics of its economic model. Below, I will summarize the design intentions of the economic model through my personal impressions combined with previous exchanges with the team. These design intentions are closely related to the team's long-term vision, and I hope that readers can feel the team's adherence to values and mission.
1) A mildly improved version of the bonding curve - giving investors more room for rational thinking and a foundation for long-term holding
MyShell currently has nearly 500 public Agents/APPs, each with different personalities, fine-tuning, underlying models, and functions. Investors have too many choices and can make rational judgments and long-term investments based on the popularity and usage frequency of the Agent/APP, and join in the process of promoting it, rather than being kidnapped by the bonding curve and the ever-increasing numbers and making quick/scattered/illogical investments.
2) It’s not about earning interest rate spreads, but about earning dividends - implementing the long-term investment philosophy and empowering products
The price of Agent/APP will rise as more people buy it, but if the Agent has not yet interacted with more users and generated dividends, investors cannot get any income. If you regret leaving the market at this time, not only will you not be able to earn the spread on the price, but you will also have to deduct 10% of the principal as a handling fee. Therefore, early entrants need to work with late entrants to make the pie bigger, and cannot withdraw their capital and leave the market to reap the late entrants. The correct investment thinking is: the design has ensured that investors can redeem 90% of the principal, and the investment risk is very small, and once the Agent/APP has a hot upside, it will be very large. Therefore, you should choose low-value & high-potential Agent/APP, promote it after buying it, and obtain dividend income from the interaction by promoting the asset. In a sense, this kind of investment is actually more like a staking game. If you believe that good Agent/AI has long-term demand in the market, you can buy it with peace of mind and avoid frequent transactions, and share the benefits for a long time.
The profit earned by investors depends only on the frequency of agent use, not on the number of buyers. This further provides more opportunities for price recovery for a wide range of assets (whether they have experienced hype or not). People will believe that the once popular Agent/APP just meets the needs of a small group of people. Even if it is silent for a period of time, it will be discovered and popular again by another small group of people with the same needs. Rick bot (from Rick and Morty) on the MyShell platform is the best example. It became popular when it was first produced, and now it is popular again because of the addition of investment gameplay. It proves that high-quality Agent/APP can take on all the new attention brought by product updates/assetization, and be loved by different user groups at different time periods.
3) Diverse games help value discovery and encourage the emergence of high-potential new assets
Participants in the economic system may have these thoughts: Because the entry cost is different, but the design of equal dividend distribution means that when the price of the same Agent/APP is too high, it needs to generate more interactions to make a profit. For example, if you buy 50 points and 500 points, the former only needs 10 points to start making a net profit, while the latter needs to earn 50 points to make a profit before it can start making any income. Therefore, there will be a game and thinking among investors:
Choose classic Agent/APP. This type of asset can maintain a high frequency of use every day. Although it has been hyped to a higher price (such as 500 points), more points can be earned every day.
Choose an emerging Agent/APP with a lower price (such as 50 points). After investing, actively promote the bot so that it can interact more with people.
Such a trade-off will make more people hope to discover emerging agents. There will be more demand for new assets in the market, which will attract investors' attention and hype for newly launched agents/APPs. High-quality works will be valued in this process, and even flip those older classic agents; while defective agents will expose their shortcomings after a wave of popularity, and their investors will be able to encourage creators to make improvements.
In addition, the "grab discount" mechanism for early investors to exit is also very interesting. Early investors give way, allowing new investors to invest in an asset at half the current price or even lower. This mechanism actually stimulates a small number of new investors to quickly grab an old asset to maintain the total value of the asset and avoid price slides and death spirals that occur in the bonding curve; more new investors joining can promote the asset again, which can repeatedly stimulate the process of value discovery-promotion-hype-calm-value discovery again for an Agent/APP.
Last but not least, in order to protect the hard work of the team and creators and the normal operation of the ecosystem, MyShell will definitely protect the interests of ecosystem contributors and crack down on profit-seekers who maliciously open multiple accounts and cheat. MyShell will implement strict risk control and anti-cheating mechanisms. Welcome users who want to invest in the creator ecosystem and can grow with MyShell for a long time. Inorganic abnormal interactions will have a higher risk of being blocked.
3. Long-term expansion of the economic model - building an ecosystem, not a model
1）Agent/AI-native APP as investable asset
Why does the logic of investing in Agents and receiving dividends succeed on MyShell, and why can MyShell optimize and amplify FT-like betting methods?
Let’s return to the essence of MyShell as a creator tool platform and Agent/AI-native APP as an investable asset - that is, an irreplaceable tool platform, excellent native content supply, AI agent asynchronous interactivity, and convenient and long-term investment.
Investors who are active in various web3 point-based new IPO systems may be accustomed to short-term bubbles stimulated by various fomo and wealth effects. After friend.tech, there are actually many products with point-based systems, perhaps chat companionship, live broadcasts and mini-games. Although they can complete the cold start through the point-based system, their products themselves have the following four pain points:
The tool is not irreplaceable (it can be replaced by traditional social media or apps, such as X, Telegram, Twitch);
The non-native nature of the content (more information may be spread on other social media such as alpha groups, rather than a new APP);
The interaction lacks asynchronicity (the host must be online to answer questions);
The convenience and long-term nature of investment are weak (requiring homeowners and investors to conduct frequent transactions and operations, and requiring high financial capabilities from investors).
In short, MyShell has created a strong moat through integration and tool polishing: creators must create here, users must use products here, and investors must invest here. Moreover, its investment is more suitable for the participation of a wide range of users. This is irreplaceable by any other low-cost, short-cycle products and mass social media, and with the establishment of the growth flywheel and ecology brought by the open economy of Web3, it has become a more difficult barrier to surpass.
2) Help small AI startups raise funds and launch products, and use the Web3 economic ecosystem flywheel to solve the problems of commercialization, financing and exit of AI in traditional industries.
“A small AI startup and a web3 project launched through IDO or DEX share many similarities.” -Ethan (Cofounder of MyShell)
Etha and I once had an interesting brainstorm, that is, small AI products = $MEME, because they all have the following characteristics:
Small amount of capital (low cost with the help of open source model);
The product pace is fast (short development time, product delivery within a week, no time to discuss financing/find lawyers/deal with legal terms);
Short life cycle;
Able to become popular quickly, and need to make enough money within two weeks of the short-lived popularity (to cover development costs + develop the next product);
We are familiar with short-term popular products, such as "Miaoya Camera" or "Cat Girlfriend Simulator", which immediately launched subscription/buyout/advertising to make a profit during the first wave of popularity, but the cash flow is not sustainable. Behind these survivors are countless small AI teams and startups:
In the first quarter of 2023, there were 300% more small AI products than in the fourth quarter of 2022, with an average of 100 AI products released on Product Hunter each month. However, the number of AI startups has dropped by 70%, and 83% of these companies have less than ten employees. (source: Sortlist)
The above data shows that in the application fields of AI and even the broader post-AI era, small entrepreneurial teams (even individual entrepreneurs) have emerged, while large-scale companies have sharply decreased.
With the emergence of a large number of small-team startups, changes in liquidity and commercialization paradigms have also come along. Most VC funds have flowed to giant companies, while more small AI companies may use their own funds or obtain angel/seed round investment, but subsequent PMF and commercialization will become huge problems. A large number of small teams are still exploring PMF, and only a few have become popular through short video media for a short time, but they missed the heat before they could commercialize.
Therefore, the paradigm innovation of AI entrepreneurship requires matching liquidity and commercialization support, namely the Launchpad model and the IDO model.
Now, imagine a platform that provides most open source models and mainstream closed source models, allowing developers to quickly get started with making and delivering products; at the same time, it has native currency and a large number of individual investors who can provide financial support in the early stages of the product; and with a certain probability that the product will become popular, it can receive liquidity injection and hype from the 24/7 financial market - such an innovative and entrepreneurial product launch platform is exactly the role that MyShell will play.
By then, the value of MyShell will be the sum of all AI products, just like Valve to independent games, Kickstarter to consumer products, and Product Hunt to mobile Internet products, but more than that, it also has the traffic of Apple Store and the capital and liquidity of Launchpad/DEX.
3) More to come: Explore and experiment with AI agent market and other valuable topics related to agents. Limited by the length, product stage, confidentiality, and not wanting to distract readers from the alpha stage economic model, there are many possibilities about economic models and ecology (already in the pipeline) that are difficult to discuss here. I look forward to meeting readers in future research articles. For example:
Agent market on the whole chain
Trading agents with different strategies and personalities
Game agent (helps you optimize game strategy and idle, and even bribe other agents)
Calls and interactions between agents
…
All these powerful functions will be integrated into the huge user ecosystem and economy in the future. For a team with strong research and technology implementation, it is a very inclusive/experimental/stable ecosystem. The release of the alpha version of the economic system does not mean the end of the product, the MyShell ecosystem has just begun;D
postscript
In the past year, many AI and IP teams have come to me, and they are all curious about the creator economy brought by web3 (such as the largest GPT Prompt community in Web2 and well-known anime IP). However, they are not clear about how to integrate the funding/investment characteristics of web3 into their communities and products in a comfortable and secure way. I am very happy to see MyShell take the lead and embed it into their ecosystem. With the development of the ecosystem and the iteration of products, this problem will be solved and become a pioneer in the future development of the creator economy. You never know until you try.
In many of our previous discussions, we have repeatedly explored the combination of the hype nature of web3 with creators from web2 and the various possibilities that this combination brings.
We know that due to the strong stimulus effect of the web3 economic system, many teams have become short-sighted and eager for quick success. However, only those teams that truly care about the creator ecosystem and products will be willing to spend time building an economic system that meets all the "wants and wants" and tries to make all participants in the ecosystem contribute and get the rewards they deserve.
The MyShell team's answer to this question has always been consistent. They firmly believe that moderate growth and sharing of revenue can bring more long-term and stable growth momentum. From day one, their goal has been to create a truly powerful AI tool that can rival the giants of web2, while sharing revenue with creators and the entire open source community. I am very happy to see that they are fulfilling their promises in countless product iterations and explorations of economic systems.
