Accelerating AI and Blockchain at the deAI Summit TOKEN2049
We have just wrapped up our inaugural deAI Summit at TOKEN2049. And it was a resounding success! More than 150 attendees showed up on a sunny Monday afternoon in downtown Singapore.
Zac Cheah, co-founder and CEO of Pundi X, kick-started the deAI Summit with his keynote speech. In his speech, Zac shared about the history of Pundi X, how it got started in the ICO days of 2017 with the vision of blockchain-based PoS system, then launched its first Layer 1, Function X, which is now rebranded to Pundi AIFX to focus on solving the issue of centralization in AI data tagging and annotation. Most importantly, Zac emphasized his belief that the age of AI is upon us and AI will be as pivotal and groundbreaking as the invention of the Internet itself.
We kickstarted the deAI Summit proper with Panel Discussion 1: What can blockchain bring to AI. The panel featured speakers from high profile projects such as 0G Labs, Circle, Akash Network, and of course, Pundi X. It was moderated by Teng, a former Delphi Digital analyst and cofounder of new research firm Chain of Thought.
Teng started the conversation by setting up the context of blockchain and AI, asking the important question of why AI needs blockchain at all. Zac was first to answer, sharing that he believes it’s much safer if data is stored on IPFS and onchain, rather than being centralized in a handful of large corporations. Tiffany Lai, VP of Ecosystem from 0G Labs, echoed the same concern as Zac. She also shared that decentralized AI provides benefits such as onchain provenance, incentivization via tokens, and transparency. She highlighted that it’s important to know what data is being used, or if there are any biases in the data, if data contributors are being properly compensated, and blockchain enables all of this.
Meanwhile, Zach Horn, Head of Marketing at Akash Network, emphasized that compute is the most valuable resource that we have, because fundamentally all Internet and digital services run on compute; and our demand for compute will only increase year over year. Given the importance of compute, it’s crucial that there is democratized and open access to this valuable resource, which is what Akash offers.
Blockchain and AI is a bilateral relationship, said Ian Chong, Circle’s APAC Developer Ecosystem Lead. He speculates that in the future, everyone will start using AI agents to help with tasks and manage different parts of our lives, and Ian believes that AI needs blockchain because blockchain provides the trust layer, enabling AI agents to trust each other at scale in order to work harmoniously together. As Ian works on Circle’s grant Program for developers, he also shared some interesting use cases he’s been seeing, such as the use of agents to monetize IPs and AI girlfriends.
Next up, we had Greg Osuri, founder of Akash Network, give his presentation on compute, entitled “Compute: Past, Present, Future.” It started with a history lesson on technology and computation, outlining how the Internet was created, followed by the creation of the Big Tech companies we know today and the centralization of cloud computing. What drove the rise of centralization? Convenience and scalability. But heavy centralization leads to higher costs and more difficult access to compute, which leads to, in Greg’s words, “digital feudalism.”
Enter decentralized cloud computing. To quote Greg, “decentralized compute represent a return of the Internet’s founding principles that we originally began with.” Compute is so important that it should be considered a public utility, like air and water. The centralization of compute is slowly being chipped away. Greg showed a market map of the dePIN sector and pointed out that all different layers of the cloud are being decentralized, from storage to compute networks to machine learning to mobility networks, with Akash Network leading the charge. Thus far, Akash has had 27k applications deployed on it, and is only going to up.
After Greg, we had Ian from Circle to share with us what Circle has to offer to developers. Circle’s main product is of course the stablecoin USDC. Ian shared how some projects such as Lemon, an Argentina-based crypto exchange and digital wallet, is using USDC to help Argentinian citizens to hold their wealth in USDC as their local fiat currency is suffering from severe inflation; and AirTM which is using USDC as a way for businesses and freelancers in the emerging markets to be able to do cross-border transactions and remittances with very low fees.
For developers, Circle has the Circle Platform, which comprises programmable wallets, smart contract platform, and cross-chain transfer protocol. Focusing on programmable wallets, Ian shared how this is relevant for AI developers. In the future, every AI agent will have its own onchain wallet. Given that there are so many blockchains out there, it’s important that the AI agent’s wallet is able to have cross-chain access across the many L1s. So what are programmable wallets? According to Ian, think of them like embedded in-app wallets-as-a-service. With this, AI agents will be able to own wallets at scale.
Our second panel discussion of the day is about data ownership in the AI era. The panel comprised Andrew Goldsky, Global Marketing Director at Gaia, and Benson Huo, the CEO of Nubila Network; and it was moderated by AI Network’s Rajeet Singh. As we know it today, data is incredibly centralized. Although users created the content/data with their own efforts, centralized platforms like Instagram and LinkedIn can use their users’ data to train their own AI model without compensating its users, while OpenAI and others scrape the Internet for data for their own models without compensating anyone as well.
Meanwhile, decentralized AI aims to empower users to take more charge and ownership over their data. Gaia’s Andrew shared that Gaia, which is building a distributed and incentivized GenAI agent network, gives users more ownership over the training data that they provide to AI platforms, with the goal of enabling users to monetize their own data. Nubila Network, meanwhile, is focusing on weather and ESG data. Nubila’s CEO (or Chief Raining Officer as he prefers to be called) Benson shared that environmental data today is controlled by a single authority, typically the country’s weather or meteorology station. Through its distributed network of all-in-one weather stations and Web3 native portable weather stations operated by its community, Nubila can not only collect extensive climate data but also analyze it in real-time, and can offer these services to businesses and governments. Device owners are then able to participate in the profits generated from the sales of these services.
At the end of the day, as Andrew very rightly put it, the only real way to track the use of user-generated and user-owned data and to enable proper incentivisation in AI networks is through a blockchain.
Our third and final panel discussion explores the venture opportunities in the world of AI-driven Web3 startups and what VCs are looking for in AI Web3 startups. The panel comprised Rio Ogawa, Head of Marketing from Impossible Finance, Nathanael Christian, CMO of Tokenomy, Raymond, founder of Mossfire Capital, Abby Huang, Head of Woo Innovation Hub at WOOX; and it’s moderated by Peko Wan, co-CEO of Pundi X.
Raymond kicked off the panel by sharing the criteria he looks out for when conducting due diligence on Web3 AI projects. He said that projects must leverage both smart contracts and some kind of AI technology, and apply it to a real use or to solve a real problem. Raymond also highlighted that he’s looking more at the application layer, rather than LLMs as they are too difficult. Abby shared that the intersection of AI and blockchain is not new, sharing examples like Ocean Protocol and SingularityNET. And like Raymond, Abby too said that her team is looking at applications that are at the intersection of AI and another vertical, such as gaming, social, and security. Community is also something to look out for, shared Rio, not just tokenholders and users, but developers as well. Most importantly, all the panelists agreed that there is a lot of hype in blockchain and AI now, but what’s most crucial is that projects must aim to solve real world problems and pain points, must have project market fit, and not just append a “.ai” domain to their project URL.
Conclusion
It was an intellectually stimulating and productive 3 hours. During the networking that followed, we felt excited by the conversations that our attendees were having. Two main takeaways we had are:
- Everyone is cautiously optimistic about blockchain and AI. It’s a genuine blue ocean, but at the same time, full of hype and hot air. So careful due diligence is required.
- Real world use cases and product market fit are the most important criteria for blockchain start-ups, AI or not.
- The application layer is getting more and more serious attention
Thank you everyone for attending deAI Summit. We hope to see everyone at the next TOKEN2049!
