Moderator: Yao, MCM Growth Lead of Solana Foundation

Guests: Sean, founder of RateX; DY, co-founder of HajimeAI; Dr. Darcy, co-founder of Starpower; Edison, co-founder of CUDIS; Tony, founder of jogojogo; Rene, founder of CharacterX; Shannon, co-founder of IntentAGI; Kevin, co-founder of HackQuest.

Edited by: aididiaojp.eth, Foresight News

Recently, the Solana 2024 Renaissance Hackathon MCM area came to an end, and the list of winners has been announced. The first prize winners are RateX (margin-based synthetic interest rate trading market); the second prize winners are CUDIS (wearable devices that provide health data ownership and support artificial intelligence) and JogoJogo (on-chain social entertainment platform); the third prize winners are CharacterX (the core product is the AI ​​social economic tool XFriends), Hajime AI (a decentralized AI Hub designed for the home), Starpower (decentralized energy Internet) and IntentAGI (autonomous intelligent agent operation layer).

We would like to take this opportunity to invite key members of the winning team to share with everyone how to explore new trends in the Solana ecosystem.

Host Yao: Please introduce yourself.

Sean: Hello everyone, I am the founder of RateX. Before RateX was founded, I had been trading in the traditional capital market for 10 years. I used to be the head of fixed income investment and portfolio manager, managing an investment portfolio of more than 10 billion. I have been exposed to Crypto for three or four years. In the fourth quarter of last year, I went all in on Crypto and founded RateX.

Tony: Hello everyone, I’m Tony, the founder of JogoJogo. Before founding JogoJogo, I had been working in the field of game distribution and global payments. Our games have generated $1.5 billion in turnover in Brazil, and global payments have an annual throughput of about $2 billion. I once participated in FTT’s initial investment and have a natural affection for Solana. I personally enjoy the 0-1 building and learning process more, because we are all on a vaguely correct path, hoping to bring small changes and improvements to the world.

Edison: Hello everyone, I am Edison. Before founding CUDIS, I had been starting my own business. I entered the Crypto field in 2016. I invested in Bitcoin and Ethereum at the beginning, and then participated in the investment and incubation of many early projects. I also worked in traditional VC for more than two years. We are more concerned about building a project or brand in the physical world in the field of encryption. At the same time, we also hope to provide better products to help everyone solve some real-life problems.

Rene: Hello everyone, I am the founder of Character X. Before joining the tech industry, my dream was to be a professor of literature. I have always been concerned about human loneliness and social issues in postmodern society. At Stanford, I found that everyone was starting a business, and the strong entrepreneurial atmosphere infected me, so I also joined the tech industry. When I was doing social networking at Tencent, I worked in the Southeast Asian market. I joined the Crypto industry by chance. It was a big bear market when I entered the industry, so I tried many business directions in the process. Finally, after ChatGPT came out, AI+social networking gave me more imagination about how humans can stay connected in postmodern society, so I started to establish Character X with some friends.

Shannon: Hello everyone, I am Shannon, co-founder of IntentAGI. I am a researcher in the AI ​​industry and have more than 6 years of research experience in mainland China and Silicon Valley. We are now also working on a new decentralized model for multimodal operable devices.

Dr. Darcy: Hello everyone, I am Darcy, the co-founder of Starpower. After I graduated with a Ph.D. in 2016, I have been working in the renewable energy industry. My Ph.D. major is photovoltaics and renewable energy. Last year, another co-founder approached me and said that he wanted to work on Depin in the energy field. I thought that the new energy and the distributed characteristics of blockchain might be a perfect combination. Starpower aims to connect our distributed energy devices and build a decentralized energy Internet.

DY: Hello everyone, my name is DY. I graduated from the University of Auckland. I worked on Bitcoin and Ethereum mining in 2017, and then worked on market value management at Binance and OKX in 2019. At HajimeAI, I focus on hardware and economic model design as well as mining machine hosting. HajimeAI is a basic network for peer-to-peer edge computing, which provides some plug-and-play AI models and personalized AI assistants for specific scenarios. We recommend these models to edge devices through containerized docker, and we can perform real-time upgrades. Our application scenarios are mainly for family companionship, personalized AI services and sound processing.

Kevin: Hello everyone, I am Kevin, the founder of HackQuest. We are currently working on a developer education platform, and we will also do some hackathons. We started organizing hackathons with major Internet companies and some Web3 public chains in 2021.

Moderator: Please invite the guest to introduce what the project is mainly doing, what practical problems it solves and the team background

Sean: What RateX is doing now is a margin-based synthetic interest rate trading market, and its basic functions are somewhat similar to Pendle V3. Based on our current basic functions, we can also provide users with some one-click generation of strategies and the option of creating synthetic customized interest-bearing assets.

We have developed the first DeFi protocol for leveraged interest rate trading on Solana, and we also hope to bring more choices to users of the Solana ecosystem. During our financing process, many investors were concerned about the core differences between us and Pendle. First of all, we are a leveraged trading protocol. We can provide users with a more efficient investment and trading option. Users can use 10%-20% of the funds they currently use on Pendle for trading. Since we are a margin trading protocol, we can conduct long and short two-way trading. Second, for some assets with low absolute returns, such as assets with the lowest absolute returns of less than 20%, we can provide users with lower slippage. Third, because we are a synthetic income protocol, our income comes from public market data. Not only the native interest-bearing assets on Solana, but also the interest-bearing assets on Bitcoin l2 are placed on our protocol for trading, and also include some real-world assets, including US TREASURY, CPI, etc.

We currently have nearly 20 people in our team, more than half of whom are developers. We now have three co-founders. In addition to me, our COO used to be a top interest rate derivatives trader in the market. Our team is relatively experienced in the field of interest rate derivatives. Our CTO has almost 20 years of experience in trading system development. He has also led the development of some well-known derivatives exchanges in recent years. He is also a security expert and has served as the chief security consultant at a very top derivatives exchange.

Tony: JogoJogo is committed to becoming an opinion trading protocol and solving the following three core problems. First, make everyone's opinions assetized and tradable. Betting on your own opinions is a truly admirable action, and you must be truly invested and participate in it (skin in the game). Secondly, solve the trust problem between the prize pool and the players. We ensure the automatic liquidation and settlement of assets by building on the chain. Finally, introduce an annualized rate of return (APY) with a real positive expected value (EV) to push LST and DeFi into the next stage. These are the three key directions we are committed to solving.

Our team members have extensive experience in game publishing and deif building, as well as global implementation and organizational capabilities. We will build different implementation products based on the characteristics of each country, and are currently planning to expand the team size.

Edison: CUDIS products mainly consist of three parts. The first part is our Smart Ring. Currently, the Smart Rings on the traditional market are more inclined towards technology and sports. We may pay more attention to fashion, such as the design of our rings and the materials used. Smart Ring itself can help users record heart rate, blood oxygen, steps, calories consumed, sleep status, etc. We can provide users with multi-dimensional data.

The second part is our app. In addition to providing a data panel to display all the data, there is also a dedicated AI coach to provide daily advice based on daily data. We have worked with top coaching teams from the United States and Asia to model the AI ​​coach. Social activities can also be carried out in the app. Users can manage their family's health through the APP and initiate community competitions. We emphasize more on Wellness, which is actually divided into 8 different dimensions, including professional, physical and mental health.

The third part is the website, which currently only accepts users to purchase with SOL. We have just sold 1,500 rings, and will have a second sale next week. After that, we will accept a variety of assets in the Solana ecosystem as payment methods.

There are two core issues we want to solve. The first is that we want to help users encrypt their data and then put them on the chain. Users can find their data at any time and can also import the data into other applications. Another point is that we see that many entrepreneurs cannot make a good income in this field because most of these Web2 applications are free. In the Web3 model, we hope to help more builders create better products through the community and user data. While providing value to users, builders themselves can also earn considerable income.

Most of our team members are from UCLA and UCB. I graduated from UCLA, and my co-founder graduated from UCB. We have been working on crypto-related projects. Team members have strong backgrounds in crypto and traditional consumer industries, and most of them come from companies like Amazon, Microsoft, and Nike. We hope to provide users with more interesting products in the future to help improve the quality of life.

Rene: In the early stages of CharacterX, we positioned ourselves as an AI social platform, where you can create your own AI character and interact with AI characters created by others. We embedded multimodal interactive experiences at a very early stage, including the ability to send private photos to each other, and AI can make personalized replies based on the photo information sent by the other party. Because our entire product design is very good, and there are no similar products on the market, we found the opportunity and quickly attracted a large number of users. So far, our user registration volume has reached 4 million. We have been observing user behavior and collecting their feedback.

We found that decentralization itself is very important in the era of AI social networking. We have observed two major demands. First, users may use multiple platforms at the same time. However, if a new game or another different AI platform appears, I want to take my AI husband to the new platform to play, which is obviously impossible. However, it is troublesome for me to create a new husband every time I go to a new platform. There is also a more terrible situation, that is, if the platform goes bankrupt directly, my AI husband will die with it.

We found that users want to preserve memories for a long time, and hope to take their AI husbands to different platforms to play. In addition, AI design itself still has a certain threshold. Even if we provide better and simpler tools, there are always only a few high-quality content creators. In the process of interaction, everyone is contributing their own data and is eager to participate in the AI ​​economy. This is also what we hope to help everyone through a more democratic economic system.

Based on these two biggest user demands, our product has entered the next stage, which is why we participated in the Solana hackathon. We are building a complete AI-based decentralized social identity system, giving each AI its own identity system, and being able to independently complete identity management and memory management.

In addition, we have also made some economic connections using the underlying technology. These two needs are the real needs of our users that we have observed. At present, our team has added more senior Solana developers and AI experts to help us design AI models in the long term. Our early core team also includes partners like myself who came from Tencent.

Dr. Darcy: Starpower mainly uses technologies such as the Internet of Things, blockchain, and AI to connect energy devices such as air conditioners, water heaters, and electric vehicles for home and commercial use to form a dispatchable energy device network. The network can improve energy utilization efficiency and respond to various application needs, such as demand response of our virtual power plant, thereby improving efficiency and data management on the energy side.

Climate change is generally recognized as one of the biggest challenges facing the world. Our energy is gradually shifting to renewable energy. In other words, the proportion of renewable energy is increasing. But one of the biggest characteristics of renewable energy is instability. Let me give you a very simple example. When solar photovoltaic power generation is in progress, a cloud suddenly floats over. Assuming that the total power generation is reduced by 10%, we need to connect the energy equipment network I just mentioned. When your power grid loses 10% of its power generation, we will reduce the demand side by 10% accordingly. For example, turn up the air conditioner by one degree, or turn down the water temperature of the water heater by one degree.

What is one thing that the energy industry is very enthusiastic about? It is the popularity of our electric vehicles. The annual sales of electric vehicles are about 10 million now, and the number of electric vehicles is expected to reach 1 billion in 10 years. Every family that buys an electric vehicle will basically install a charging pile. After installing the charging pile, a storage battery will basically be installed, so that we can mobilize more and more distributed energy. This is an incremental blue ocean market that we need to occupy quickly. Starpower can achieve a rapid increase in the number of users with the flywheel effect of Web3 tokens. We want to become Web3 Uber, connecting the energy equipment of every household, including power generation equipment, energy storage equipment and power consumption equipment, to participate in the intelligent dispatch of our power grid.

Most of our team members come from the traditional energy industry. Another co-founder is the assistant to the chairman of HashKey Group. Our team now has about 20 people, consisting of the IoT team, Web3 team and energy hardware supply chain team from Wanxiang.

DY: The core of HajimeAI is a peer-to-peer edge computing infrastructure network, where users can use AI services in HajimeBot while contributing idle computing power to the network.

Our system is mainly divided into three parts, namely HajimeAI, HajimeBot and HajimeGarden. HajimeBot actually presets a series of artificial intelligence models and has some basic configurations. The first generation series has 8G of video memory, because high video memory is required for edge reasoning and computing. HajimeGarden is similar to a crowdfunding platform that can provide computing power for AI.

HajimeAI builds a decentralized peer-to-peer network in an N to N manner. The AI ​​deep learning architecture Transformer is highly dependent on computing power. It continuously stacks neural networks and swallows data in multiple dimensions. This cloud computing model certainly cannot achieve low latency and privacy. For example, users are reluctant to put some private information, such as photos, sounds, or some health-related data, in the cloud.

In addition, if OpenAI uses Transformer, it cannot avoid high latency, and even cannot perform long-term base memory because its number of tokens is also limited. So we built a decentralized dispatch center HajimeAI through N to N. HajimeAI can dispatch the computing power of power equipment to verify whether these computing powers are working effectively, and then do some customized models to provide standardized services.

Our entire team is mainly composed of investment research from Yale and Goldman Sachs, and data modeling is mainly led by a Harvard PhD. Most of the R&D team have been working for more than 20 years, mainly from Baidu Advertising Department, Huawei and JD's smart home.

Shannon: Now we usually say that AI writing poems, painting and chatting are all AI content generation, but in fact AI can't really help me do work, but these are exactly what we hope AI can do for us. We hope that AI can help me finish my work directly with my computer, which is actually a problem we are solving now. We are training AI assistants to understand the status of my computer like a human, and then infer the next action to be performed.

Currently, most GPT4 multimodal components have an accuracy rate of only 10.59%. Our team has improved it to 50% accuracy, which is currently the best AI assistant among researchers. Our model has begun to enter a stage similar to GPT2. We have a model that everyone can use, but it is still quite inaccurate. So we also proposed the concept of AI Coach to Earn. We invite everyone to participate in model feedback learning with us. In this part, we will use tokens as incentives. Secondly, we solved the problem of model benchmark and built a composable ecological foundation for everyone. We will launch our first version of the product browser plug-in at the end of this month. Everyone can give instructions by talking to our product. Our product will understand everyone's intentions and directly operate your browser to help you solve some problems.

Moderator: Why did you participate in the Solana Hackathon? What are the points of convergence between your product and Solana technology and ecology? What are the subsequent product and project operations and financing plans?

Sean: We should launch the testnet version next month and conduct community testing at the beginning of next month to collect some opinions on the trading function. In the second half of this year, we will launch synthetic interest-bearing stablecoins and bonds based on the core functions of RateX. We have completed our Pre-Seed financing last month, and will start the second round of financing after the product is launched.

Tony: We will launch a product themed around the European Cup on the Snoic testnet in June. All users can create their own prize pools, invite friends to share their views on the game, and enjoy special rewards from JogoJogo and Snoic. After the product testing is completed in June, we hope to get feedback and suggestions from the community. At the same time, we plan to start internal testing of a DePIN device for viewpoint trading in July, and will open internal testing quotas. We hope that the community will continue to pay attention.

Edison: We will soon launch the second phase of sales next week, and we will also meet you at an offline event in Singapore in September this year. Our current users are scattered all over the world. We hold events all over the world every month, in New York in March, Dubai in April, Austin in May, Malaysia in June, Brussels in July, etc. We want to give users more experience of smart rings in the future. In addition, we are constantly iterating new features and hope to give users more good experiences. We are preparing to develop or co-build some new products in the fourth quarter. We have basically completed the seed round of financing. We have always been working in North America before. In the third quarter, we will cooperate with more APAC communities to promote.

Rene: We are currently conducting SDID Mint activities, which are mainly for human users in the initial stage. In July, we will have SDID Mint for AI, and we will hold many related activities from this month to next month. We completed the Seed round of financing earlier this year, and the next round of financing is expected to start next month.

DY: We started pre-selling verification nodes in June, because our verification nodes are mainly used to verify computing power nodes. At the same time, the client will have a test network for sound data collection, and we will test a customized model at the edge. In August, a prototype test network of HajimeGarden will be launched, which can realize functions such as scheduling and verification. We also signed a cooperation agreement with last year's hackathon champion project to allow them to enter HajimeGarden, and we provide computing power. Our node sales last month were about 1 million U, and this month we plan to go to the European and American markets for financing.

Shannon: We will release the first version of our product this month. Users can participate in Train AI to Earn and submit learning data related to the model by communicating with our model. After that, we will mainly develop around our entire action layer. In the future, we will invite more AI practitioners to participate in our entire economic system. Our team is currently in North America, and we have some plans for financing in the future.

Dr. Darcy: Starpower pre-sold its first smart socket in April, and the number of pre-sales has exceeded 20,000 units, which is already a relatively large Depin project. We are currently carrying out the next marketing campaign, which includes 300 engineering machines for promotion and gift. You can follow Starpower's official Twitter to participate in the event and get our engineering machines and prototypes to play with. Our team is currently working hard and will be selling them simultaneously around the world. We are currently working on certification in South Korea, the European Union and the United States. After completing the certification, our first batch of pre-order users can start receiving our devices.

Moderator: Did you get any support during the hackathon? Many winning teams came from the founders bootcamp organized by HackQuest and Solana MCM. What did you gain from this bootcamp?

Adam: Let me share some of the things we will do before and after the hackathon. The Solana ecosystem is relatively unfamiliar to the Chinese developer community, so you are very welcome to find us at any time if you have any questions about the Solana ecosystem or development, and we will try our best to help. Before the hackathon begins, if you have good ideas or want to know more about the ecosystem, you can find us to communicate, including participating in HackQuest activities and offline meetups to meet more ecosystem partners. After the hackathon begins, if you have technical questions or need more multi-dimensional support, HackQuest and others provide many ways to participate in the ecosystem. The Solana ecosystem is very unique, because this ecosystem has actually experienced a lot of hardships, and the growth of the ecosystem has also been twists and turns and has become very resilient. Now this resilience has become the Solana community culture, and everyone continues to support each other, guide and promote each other's growth. We are very happy that there are so many excellent projects running out in the Chinese-speaking area.

I think the most important thing is how to cooperate with projects in the ecosystem.

Kevin: This year, we invited founders of first-line projects in the Solana ecosystem to share their experiences at the bootcamp we organized with Solana MCM, including the founders of Helius, Helium, and BackPack. I learned a lot from each sharing, and I am glad to see so many award-winning teams come out of this bootcamp and community. We are a developer education platform. If there is a need to recruit developers, we can help recruit them. In addition, we look forward to doing more online and offline activities with Solana Foundation and other projects.

Host: As a senior, can you give some advice to potential contestants?

Sean: The biggest advice is that you must really innovate and do something cool, and at the same time, what you do is really helpful to the Solana ecosystem. I think this is the most important thing. I also hope that everyone will believe in the Solana ecosystem, which is indeed different from other public chain ecosystems.

Edison: Participating in the Solana hackathon was a particularly unforgettable experience for us. Before and after the hackathon, we communicated and learned with the foundation and many Solana projects. We also hope to have closer communication and exchanges with the ecosystem or the foundation.

If you really want to make a product that users like, and they are willing to recommend it to others, and they are willing to pay for it, then you need to spend a lot of time getting closer to users, showing them the product, and communicating with them. In the Solana ecosystem, we have seen many successful projects go from difficult early stages to success step by step, and they have all done the things I just mentioned. They have done a lot of things that seem to be very hard at the beginning, they have to do a lot of activities, meet a lot of people, talk about a lot of cooperation, and then spend a lot of time polishing the product, which will eventually lay the foundation for the explosive growth that everyone wants in the future. Many projects in the Solana ecosystem, including the foundation, are very enthusiastic and friendly. Everyone is willing to give you some advice and help you answer some questions, so I think the most important thing is to ask more questions yourself and communicate more with everyone.

Tony: I think you must have the right intentions. You must believe that Web3 can solve problems, such as peer-to-peer payment transparency, openness, transparency, and trust. The second point is that you must be passionate. You will encounter many problems in this process. In this process, you must actively communicate and continuously polish your business model.

Rene: From the perspective of the entrepreneurial team, I think the most important thing is to keep an open mind. This is especially important for entrepreneurs who have experience in the Web2 industry. In the process of communicating with many entrepreneurs, I found that they may not be familiar with the entire Web3 gameplay, and even have many doubts.

During the hackathon, the Solana team helped us achieve some transformation and upgrades in what we were doing. For example, the decentralized solution of SDID actually involved a lot of exchanges with Solana ecosystem partners. These help may be as big as the entire technical architecture, as small as some very transactional regulations in the competition. Solana ecosystem is very open, and everyone here is not just as simple as getting rewards, but truly realizing deeper changes in the entire project.

DY: The reason why we did HajimeAI is that we believe that the combination of edge computing and large models may be a mainstream direction in the future. So I think when choosing a project, we must choose a mainstream direction that may be in the next three to five years or five to ten years. Secondly, everyone is currently exploring integrated intelligence. We need to combine blockchain to provide computing power for our edge computing devices while ensuring the effectiveness of revenue and computing power.

Dr. Darcy: First of all, as a project party, we must have a deep understanding of the needs. Our project must solve a practical problem and have real market demand. For example, in the field of new energy storage, everyone knows that it is a growing blue ocean, and the development of new energy in the mainland is very vigorous. The pain point of traditional manufacturers is how to quickly roll out hardware. We need to think about how to use tokens to drive the flywheel effect and roll out hardware. Technology must be used to serve applications and find the best combination of technology and application. The second point is that we must work as a team. A relatively strong and complementary team is the key to success. The third point is that we must actively participate in the community, use community resources, and then actively discuss and cooperate, including discussions and cooperation with other project parties. The last point is that we must maintain innovation, bold imagination, and constantly improve and optimize products and services.

Shannon: First of all, we need to solve a real problem. For us, first of all, we hope that AI can help us work and become a real assistant. Then in this process, we need to constantly iterate and listen to the voices of our customers. How do users use it, and how do they want to use it? Secondly, we need to think clearly about why we must use tokens and use the Web3 method to motivate and solve problems. For us, the AI ​​training stage is actually very important. Either you hire people with cash to participate in this process like OpenAI, or you use token incentives. Token incentives are a very necessary stage.

Finally, we must actively participate in the Solana ecosystem and cooperate. AI projects prioritize high throughput and low fees, as well as a wide range of application scenarios. In the Solana ecosystem, every project may become our partner.

Adam: The Solana ecosystem has opened up a lot of new possibilities for everyone in this cycle. The Memecoin craze has also proven that the network is moving towards a more stable and scalable direction.

Kevin: Solana is actually the public chain that I am most optimistic about. It has sufficient user activity, VCs are willing to invest in the ecosystem, and there are many excellent and loyal developers. Perhaps a killer application will appear soon.