Kartin Jiatian Wong is the Co-founder of Ora Protocol, a verifiable oracle protocol that brings AI and complex computation on-chain. ^[1][2][3]

Career

In an interview with Alea Research, Wong discussed his involvement in crypto. He began exploring crypto in high school around 2011 by writing a Bitcoin mining script in Python, earning early profits. He later moved to the U.S. to study astronomy. He co-founded an AI startup during undergrad that used convolutional neural networks to classify star images, though the project lacked commercial success. After graduation, he joined Google as a software engineer working on cryptography and AI. In 2019, he re-entered crypto through trading before fully committing to building a crypto startup.

His interest in AI dated back to 2015, when he experimented with RNN and LSTM chatbots, though these were costly and underperformed. The release of ChatGPT in 2022 reignited his focus on integrating AI with blockchain, particularly around enabling smart contracts to support more complex computations. He initially explored using zero-knowledge proofs to run large-scale AI models on-chain but found the approach inefficient for models with billions of parameters.

To address this, Wong and his team at Ora shifted to Merkle tree-based verification. This led to OPML (Optimistic Machine Learning), which transformed AI inference into a verifiable Merkle structure. Open-sourced in 2023, the project gained traction and funding. After testing various use cases, the team pivoted in 2024 to tokenizing AI models. Although loosely connected to agent token projects, their focus remained on advancing model-based AI infrastructure.^{[1] [7]}

Interviews

About ORA

In an interview on Magnet Labs’ AI Radioroom, Wong discussed ORA's role in decentralized AI on the Ethereum blockchain. He elaborated on the two main functions of ORA: enabling decentralized AI influenced by Ethereum smart contracts and facilitating the initial model offering (IMO) to tokenize AI models, transforming them into on-chain oracles. They explored the ethical implications of using publicly available data for training AI models and highlighted the need for blockchain infrastructure to prevent single points of failure in decentralized applications. Wong introduced their OPML (Optimistic Proof of Machine Learning) technology for decentralized inference and the criteria for selecting models for tokenization. Additionally, he presented ORA's strategy for launching a fair allocation system using AI to democratize token distribution, aiming to improve transparency and community involvement in the ecosystem. ^[8]

Presentations

Initial Model Offerings

At the AI x Web3 Summit, Wong discussed the concept of Initial Model Offerings (IMO), emphasizing the growing integration between artificial intelligence (AI) and cryptocurrency. He outlined the evolution of AI over the past decade, highlighting its increased capabilities and the resulting monopolization by major tech companies that profit from closed-source models. Wong drew parallels to historical economic structures, noting how stock markets allowed broader participation in investments, analogous to his proposal for tokenizing AI models on the blockchain. He introduced two key components of IMO: new Ethereum standards for tokenizing AI models and ensuring revenue sharing among token holders. Wong explained the introduction of optimistic machine learning that enables efficient and verifiable AI model operations on the blockchain. He also addressed the importance of a model selection process to avoid resource wastage. He pointed out Ora's role as an oracle protocol integrating AI into blockchain, positing IMO as a vital business mechanism in the evolving market. ^[9]

AI Agents

At ETHDenver, Wong discussed the development of verifiable AI agents and the pursuit of a concept termed "world superintelligence" at Aura. He argued that advancements in AI have surpassed human intelligence, particularly noting substantial increases in neural network parameters over the past decade. Wong emphasized that achieving world intelligence hinges on the interconnectedness of AI systems, promoting a decentralized and autonomous model free from human control. He introduced the notion of using blockchain technology to enhance the functionality and fundraising for AI, proposing two key components: optimistic machine learning for verifiability and a new economic model using a token system. Wong suggested that this decentralized approach would ultimately empower independent AI agents to operate within their economic ecosystems and manage their interactions effectively. ^[10]

Panels

Decentralized AI

At the Open AGI Summit, Wong participated in a panel titled "A Builder's Roadmap to Decentralized AI," moderated by Kenzi Wang and featured Alex Rusnak (Maru), Michael Heinrich (Zero Gravity Labs), and Sean Ren (Sahara Labs). Each panelist shared insights into their projects and approaches to integrating decentralized technology with artificial intelligence. They highlighted the potential for blockchain to enhance data verifiability, security, and ownership in AI development, emphasizing the importance of creating decentralized models that could democratize access to AI capabilities. Challenges were discussed, such as the need for efficient data processing pipelines and the balance between decentralized and centralized AI training. The conversation also delved into the philosophical implications of AI's impact on society and the evolving relationship between Web 2.0 and Web 3.0 technologies. Overall, the panel showcased the collaborative and innovative possibilities within the intersection of blockchain and AI. ^[11]

Future of AI

At Ethereum Singapore, Wong and Emad Mostaque (Stability AI, Schelling AI) reflected on the impact of decentralized systems on technology, particularly focusing on the development of Schelling AI after leaving Stability AI. They emphasized the importance of open-source AI in areas such as education and healthcare, arguing that these systems should be built on distributed ledgers for better transparency and resilience. The conversation highlighted the necessity of aligning AI with human values to ensure the safety and efficacy of these technologies, particularly in crucial sectors like medicine. They explored the potential advantages of combining AI with blockchain, suggesting that this integration could allow for greater collaboration and innovation while avoiding the concentration of power. The participants expressed optimism about the transformative possibilities of AI, particularly in democratizing access to intelligence and improving quality of life. However, they acknowledged the risks associated with centralized control and the ethical implications of advanced AI systems. ^[12]

Agent Use Cases

The panel discussion titled "Agents Unleashed - Technologies for AI Agent Use Cases at the Nexus of On

• and Off-chain" brought together AI and cryptocurrency technology experts to explore various infrastructure projects and their applications. Participants included Wong,  Zoe Meckbach (Phala), Daniel Olshansky (POKT), Stephen King (Indexing.co), Fran Algaba (Giza), and Modulus Labs. They shared insights on how their technologies intersected with AI, particularly around the challenges and demands of serving AI agents. They discussed the potential of blockchain technology to enhance transparency and efficiency in AI models and the evolving needs in data access and processing in Web3. Several use cases were presented, highlighting the integration of AI into financial systems and gaming, while also acknowledging that the costs associated with these technologies, particularly regarding zero-knowledge proofs, could be significant. The discussions underscored a collective interest in leveraging decentralized infrastructure to manage AI agents and improve the utility of smart contracts. ^[13]