Machine Learning For Web3: Realizing The Potential and The Challenge of Censorship Resistance

With the launch of Large Language Models (LLM) like OpenAI ChatGPT and GitHub’s CoPilot, the world is having a front row seat at revolutionary technologies that will define the 21st century. These models have already conquered web2, it is time that we bring the revolution over to Web3! In this talk, we will give a brief overview of LLMs, and explore the real-world applications of language models to smart-contract phishing. We will also discuss the unique challenges and opportunities that Web3 presents for ML, such as Web3 production ready data indexing services that are optimized for batch read versus write, the lack of standardization in smart contracts and ABI decoding, and the lack of developer tooling to support the emerging languages of Web3 to name a few. In the second half of the talk, we will delve into the technical details of the technologies we need to invent in order to build a censorship-resistant ML stack. Finally we propose an open-innovation approach to solving these challenges in a WAGMI-esque style. Attendees will leave this talk with a better understanding of the potential and challenges of ML in Web3, as well as a roadmap for building a censorship-resistant ML stack. By understanding the potential of ML in Web3, its real-world applications, and the challenges of censorship resistance, attendees will be better equipped to navigate the fast-evolving field of Web3 and develop solutions that are more robust, secure, and fair.