Chris, an AI researcher at 0G Labs, mentioned that in traditional AI models, even if they are open-source, it is difficult for us to know what data was used in the training, and we do not know how they will perform in new scenarios, which makes the model results difficult to trust. 0G has a good data storage infrastructure, and data can be directly loaded from the cloud to the training process, and in the future, it can realize the construction of safer and more reliable models through personal verification of the data.
Chris, the 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 large companies controlling data, computing power, and models has been discussed a lot. But recently, there have been many use cases of AI for Crypto, such as truth terminal and AI payment, and more and more projects are starting to cooperate and support the AI ecosystem. Users are very concerned about whether they can make money from the data, and the key task of the platform is to solve how to distribute the benefits between data contributors and consumers. Developers are not a group driven by vision, the most important thing is to save them time and make them money.
In the subsequent keynote speeches, Bu Fan, the Head of IPFi at Story Protocol, and Prakarsh, the Ecosystem Lead at Spheron, shared their views on the decentralization of AI asset-ization and how their organizations are adapting to this trend.
Bu Fan mentioned that there are already many landing scenarios in the market where AI is combined with Crypto. The first is user-facing chatbots, where creators create AI characters and issue commercial licenses on the chain; the second is AI meme coins, where creators can legally connect to the source IP assets on the chain and release tokens externally; the third is to provide model training data (such as pictures), which can be continuously monetized through royalties collected on the chain. But these are just very early applications, and the models have not yet taken shape, so creators can continue to explore scenarios where AI+Crypto can be combined. Story Protocol focuses on standardizing IP activities through token standards and disseminating IP in different forms. He believes that most AI is also a form of IP, and if IP can be asset-ized, then AI can also be asset-ized. For example, the pictures used to train an AI model can be IP, and the AI model itself can also be IP, and when the AI model generates new content, it can be distributed and transacted on the chain to realize asset-ization.
Prakarsh mentioned that in the AI era, computing power will become the underlying anchoring asset for most agents and most AI applications. Distributed computing power will have many application scenarios, and they currently see more promising scenarios including knowledge sharing between hospitals while protecting data privacy, and AI dialogue systems based on local computing power and models, ultimately forming a personal AI system.
The fourth roundtable focused on how to connect the Crypto and AI worlds, and investors discussed the problems faced by centralized AI systems and where Crypto+AI can break through.
Hiroki, the Head of Research at Lemniscap, pointed out that there are two difficulties in building a decentralized AI network: one is that the scalability of the distributed computing power network is difficult to compete with centralized competitors, and the other is that the quality of personal data contributions is difficult to control.
Will, an investment partner at Faction, said that currently you can let AI plan your entire vacation, but the plan cannot be implemented because AI cannot currently help you make payments. Will believes that AI Agents need to have crypto wallets, and crypto wallets will play the role of bank accounts, and the payment technology stack will have huge opportunities, because all financial transactions will have to flow through these Agents.
Ryan, an investment partner at Coinbase Ventures, believes that most models can only access public data, and cannot access sensitive private data such as finance and medical data. Crypto can drive models to access private data pools and improve AI performance in specific areas. Agent systems currently cannot complete very complex tasks, as they do not actually know how to understand the content of smart contracts and take action. We need large models that can acquire, understand, and provide human-readable interpretations of smart contracts.
Dan, an investor at Hashed, pointed out that the current distributed AI incentive system is not very perfect, and in the entire AI value chain, only a few people have made relatively large positive contributions, but their contributions are not reflected in the incentives. The lack of a good distribution mechanism has led to unfair distribution. In addition, the models owned by the community must be secure and controllable, and the ownership of the parameters should be returned to the community for research, rather than providing a black box like centralized companies. If the model involves scenarios such as emotional companionship, it should be governed in an open environment.
Sylvia, the Director at Bullish Capital, mentioned that the incentive model design process must fully consider what the underlying demand is. For example, if edge devices are needed, it must be considered how to find them among the many distributed computing devices. Therefore, before figuring out the model architecture optimization problem, it is impossible to design a truly effective incentive model.
The above is a complete review of the highlights of AiFi Summit 2024 Devcon. Even facing challenges in regulation, incentive mechanisms and other aspects, the AiFi track is also full of opportunities. With the new highs in the overall market and the all-round heat of the AI track, the industry is generally in a positive state, with a continuous influx of talents and more 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
Welcome to join the official community of BlockBeats:
Telegram subscription group: https://t.me/theblockbeats
Telegram discussion group: https://t.me/BlockBeats_App
Twitter official account: https://twitter.com/BlockBeatsAsia
