On November 12, the AiFi Summit 2024 Devcon, co-hosted by GAIB, Codatta, and Kite AI (formerly ZettaBlock), successfully concluded at the Park Hyatt Hotel in Bangkok.
Article author: AiFi Summit 2024 Devcon
Source: AiFi Summit 2024 Devcon
This AiFi Summit had 1,300 registered participants, with over 500 attendees. 27 projects and investment institutions, including Paypal, BNB Chain, Base, NEAR Protocol, Story Protocol, 0G, Aethir, io.net, Exabits, Plume, Space and Time, Hyperbolic, Faction, Hashed, and Coinbase Ventures, delivered brilliant speeches.
Sarah, the Asia-Pacific head of BNB Chain, delivered the first keynote speech. She mainly introduced the construction of the entire BNB Chain ecosystem, various support policies for developers, and updated the audience on the current progress of BNB Chain in AI applications.
In the second keynote speech, Kony, CEO of the organizing party GAIB, shared his views on the potential opportunities in the current computing power market. He mentioned that AI is the most important era after the mobile internet, and computing power has captured a significant portion of value in the AI boom. Compared to other financial assets, investing in GPU computing power assets can yield returns unmatched by other targets, but the current problem in the GPU market is the inability to efficiently connect participants on both sides. On one side are operators, who have to pay huge financing costs when scaling GPU operations for external financing; on the other side are investors, who find it difficult to invest directly in computing power assets and usually can only choose to invest in semiconductor stocks like Nvidia. GAIB provides investors with more decentralized, transparent, and AI cash flow-based on-chain assets by tokenizing computing power assets and their returns and providing liquidity.
The first roundtable discussion at the AiFi Summit was themed: 'AiFi: Financialization of AI & Compute Assets.' Core members from GAIB, Exabits, io.net, Aethir, WitnessChain, and Plume discussed the current opportunities, challenges, and industry regulations surrounding AiFi.
Jonathan, CIO of Exabits, mentioned that currently, if users want to use GPUs, they can only rely on major cloud service providers like AWS or Azure, but these platforms tend to serve large enterprises, which limits the development of startups. We need more democratic and open GPU resources to support small and medium-sized enterprises. In the Web3 world, everyone can become an investor in GPUs to break AWS's computing power monopoly, which represents a huge industry opportunity.
Asa, the Asia-Pacific head of io.net, mentioned that 50% of GPUs in independent data centers outside the three major cloud vendors are still underutilized, as these data centers lack opportunities to reach users. However, GPUs need to ensure continuous operation and face maintenance issues, making it a significant challenge in the AiFi sector to build an incentive mechanism that guarantees the interests of investors and other participants.
Kartik, the ecosystem leader at Aethir, mentioned that the entire system includes demand-side computing power, operating-side computing power, and investors. Convincing them to jointly participate in a market that relies on on-chain mechanisms is challenging. The regulatory risks exist in that in some countries and regions, incentivizing data center services through tokens may cause certain problems, so compliance boundaries need to be defined in client agreements.
Ranvir, co-founder and CEO of WitnessChain, proposed that computing power, as a new asset, requires a new pricing mechanism. There is no unified formula to calculate the commodity price of computing power; different platforms and different GPUs have cost and performance differences, and different GPUs with varying performance will contribute differently to the same task, creating new opportunities for financial mechanism design.
Teddy, the CBO of Plume, also mentioned the need for caution regarding regulation when new assets emerge. A certain compliance framework already exists for AI-related assets, making asset trading legitimate and feasible, which is something Plume is helping ecological projects with.
In the upcoming keynote speech, Yi, the CEO of Codatta, explained how decentralized data trading can propel AI towards AGI and the position and mission of Codatta in this process. He mentioned that only vertical domain data can enhance the reasoning and planning capabilities of foundational models in specific fields, and only by collecting a large amount of diverse vertical domain data can AGI be achieved. Each piece of data we provide as data contributors can actually be applied to multiple different scenarios, and each scenario will have different companies to commercialize it, meaning that the vertical domain data we provide will generate income over time, which is why we view data as assets. Therefore, we need to make data asset trading easier and achieve relatively fair pricing in the market.
The second roundtable discussion focused on the Open Data Economy, where core members from projects such as Spheron, Theoriq, Space and Time, Hyperbolic, Base, and Nevermined discussed the current state of the AI data ecosystem, infrastructure support, and future ecosystem demands.
Ron, co-founder and CEO of Theoriq, mentioned that we are currently seeing many applications that go beyond simple dialogue robots, as well as governance robots on DAOs. These applications combine the cooperation of multiple agents, and beyond the crypto field, they are increasingly appearing in marketing, analysis, and other scenarios. Many people believe that the greatest use of data is in training models, but we see that data plays an increasingly significant role in decision-making processes. Different agents acquiring different data and cooperating can create maximum value.
Scott, co-founder and CTO of Space and Time, stated that Space and Time is currently building a rule engine for agent systems using smart contracts, allowing agents to use your funds in a trustless environment, achieving the ideal on-chain form of agents. Space and Time's products can allow users to query the historical behavior of agents and set strict execution policies for agents.
Don, CEO of Nevermined, believes that to win in the data market, two conditions must be met: one is to form a monopoly on data transactions, and the second is to impose restrictions on data assets to prevent data contributors from uploading meaningless assets. A feasible approach is to build analytical tools around data assets in relevant scenarios, maximizing data value extraction and profitability.
As one of the organizers, Chi, the CEO of Kite AI, announced during her keynote speech a brand upgrade and the launch of a new AI platform, Kite AI, during the summit. She discussed the current difficulties in the development of centralized AI and how KiteAI is expanding the boundaries of AI through its solutions. She pointed out that a lack of data distribution channels and mechanisms for confirming data ownership makes it difficult for a large model to train on a vast amount of personal and even corporate data. Over the past year, the proportion of datasets on the internet with open-source licenses has dropped from 95% to 75%, making it challenging for companies training models to obtain the best quality data for their models and to achieve breakthroughs in model performance. The industry needs decentralized AI solutions to acquire more valuable data.
In the third roundtable discussion, members from GM Network, Mind Network, 0G Labs, NEAR Protocol, and Chainbase discussed how Web3 companies can participate in AI competition, data privacy, application deployment, and other topics.
Max, a founding team member of GM Network, mentioned that users have been generating a vast amount of data, but this data has not been well utilized, causing it to lose value. We need to combine the collected data with AI to make intelligent devices smarter.
Leon, the Asia-Pacific head of Mind Network, mentioned that although there are no perfect data privacy protection measures in reality, different methods combined may explore feasible solutions. To protect user privacy, Mind Network is currently implementing encryption at three different levels: one is encrypting data in distributed storage, another is encrypting through fully homomorphic encryption during GPU computing processes, and the third is encryption at the application level.
Chris, an AI researcher at 0G Labs, mentioned that in traditional AI models, even with open-source models, it is difficult to know what data was used in training, and it is unknown how they will perform in new scenarios, making model results difficult to trust. 0G has excellent data storage infrastructure, allowing data to be loaded directly from the cloud into the training process, and in the future, it can achieve more secure and trustworthy models through personal verified data.
Chris, COO of Chainbase, mentioned that there are currently two narratives in the market: one is crypto for AI, and the other is AI for Crypto. Utilizing crypto to solve the problems of big companies controlling data, computing power, and models has been discussed many times. However, many AI for Crypto use cases have recently emerged, such as truth terminal and AI payments, with more projects starting to collaborate to support the AI ecosystem. Users are very concerned about whether data can earn money, and the key task for platforms is to ensure proper profit distribution between data contributors and consumers. Developers are not a vision-driven group; the most important thing is to help them save time and make money.
In the subsequent keynote speech, Bu Fan, Head of IPFi from Story Protocol, and Prakarsh, ecosystem leader of Spheron, shared their views on the assetization of decentralized AI and how their organizations are adapting to this trend.
Bu Fan mentioned that there are already many practical scenarios combining AI and Crypto in the market. The first is user-facing chatbots, where creators create AI characters and issue commercial licenses on-chain; the second is AI meme coins, where creators can legally connect with source IP assets on-chain and issue tokens; the third is providing model training data (like images), which can continuously generate income through on-chain royalties. However, these are still very early applications, and the models have not yet taken shape. Creators can continue to explore scenarios combining AI and Crypto. The Story protocol focuses on standardizing IP activities through tokens and disseminating IP in various forms. He believes that most AIs are also a form of IP, and if IP can be assetized, then AI can also be assetized. For example, images used to train AI models can be IP, and the AI model itself can also be IP. When the AI model generates new content, it can conduct on-chain IP distribution transactions to achieve assetization.
Prakarsh mentioned that in the AI era, computing power will become the underlying anchor asset for most agents and AI applications. Distributed computing will have many application scenarios, and they currently see promising scenarios including knowledge sharing between hospitals while protecting data privacy, and AI dialogue systems based on local computing power and model support, ultimately forming a personal AI system.
The fourth roundtable focused on how to connect the Crypto and AI worlds, where investors discussed the problems faced by centralized AI systems and how Crypto + AI could break the deadlock.
Hiroki, the research leader of Lemniscap, pointed out that there are two difficulties in building a decentralized AI network: one is that the scalability of a distributed computing network cannot be compared to centralized competitors, and the other is that the quality of data contributed by individuals is difficult to control.
Will, an investment partner at Faction, stated that currently you can have AI plan your entire vacation, but the plan cannot be realized because AI cannot help you make payments. Will believes that AI agents need to have crypto wallets, as these wallets will act as bank accounts, and the payment technology stack will have huge opportunities since all financial transactions must go through these agents.
Ryan, an investment partner at Coinbase Ventures, believes that currently most models can only access public data and cannot obtain sensitive private data such as financial and medical data. Crypto can promote model access to private data pools, thereby enhancing AI performance in specific fields. Agent systems are still unable to perform very complex tasks, as they do not actually understand the content of smart contracts and take action. We need large models capable of acquiring, understanding, and providing human-readable interpretations of smart contracts.
Dan, an investor at Hashed, pointed out that the incentive system for distributed AI is not very complete. In the entire AI value chain, only a few people have made significant positive contributions, but their contributions are not reflected in the incentives. The lack of good distribution mechanisms has led to unfair distribution. Additionally, community-owned models must be secure and controllable, and ownership of parameters should be returned to the community for research, rather than providing a black box like centralized companies. If the model involves scenarios like emotional companionship, it should be governed in an open environment.
Sylvia, director of Bullish Capital, mentioned that the design of the incentive model must fully consider what the underlying needs are. For example, if edge devices are needed, it is essential to consider how to find them among numerous decentralized computing devices. Therefore, without a clear understanding of the model architecture optimization problem, it is impossible to design a truly effective incentive model.
The above is a comprehensive review of AiFi Summit 2024 Devcon. Even in the face of challenges such as regulation and incentive mechanisms, the AiFi sector is also full of opportunities. With the overall market reaching new highs and the AI sector booming across the board, the industry is showing a positive trend, with talent continuously flowing in and more innovations emerging.
For more content, please follow
GAIB: https://x.com/gaib_ai
Codatta: https://x.com/codatta_io
KITE AI: https://x.com/GoKiteAI
