Chainfeeds Introduction:
Wintermute announced that it will use Chaos Labs' newly launched Edge Proofs Oracles as the oracle for its U.S. presidential election prediction market OutcomeMarket. Chaos Labs founder Omar took Edge Proofs Oracles as an example to analyze the mechanism, advantages and limitations of the AI/LLM oracle model it represents.
Source:
https://x.com/omeragoldberg/status/1836067673278734443
Article author:
omer
Viewpoint:
omer: Edge Proofs Oracles ensure verifiable data origin, integrity, and authenticity, enabling blockchain applications to trust the external data they rely on. This feature is critical for prediction market oracles, which are a specialized subset of proof oracles designed to bring off-chain data onto the chain in a secure and trusted manner, ensuring accurate verification of real-world results such as elections. The prediction market oracle mechanism is as follows: 1) Declare reputable sources: Specify trusted sources such as the Associated Press, CNN, or Fox in advance. These sources can be used as inputs to solve prediction market events (such as determining the winner of the US election); 2) Prove the source, integrity, and authenticity of the data: The oracle ensures that the data remains in its original form, has not been tampered with, and is exactly the same as published by the declared source, thereby maintaining the integrity of the input; 3) Determine the outcome of the event: Advanced AI or LLM models process text and generate insights that answer specific questions, such as "Who won the US election?" without adding subjective interpretations; 4) Reach consensus: To ensure the reliability of the results, the oracle network reaches consensus among multiple nodes. This step ensures that no single entity can unilaterally decide the outcome of an event, thereby ensuring transparency and decentralization. Relying on verifiable machine learning models for oracle solutions offers advantages over human-driven voting systems, especially in high-stakes environments where trust, accuracy, and efficiency are critical. When trained with fair data, machine learning models provide a more structured approach that minimizes external influences, thereby improving objectivity and fairness. Machine learning models are fully verifiable. Each prediction can be traced back to its inputs, allowing third parties to clearly and confidently verify the results. Machine learning models are highly scalable and can process large amounts of data in real time. In addition, ML-based oracles can be automated. Machine learning models are also resistant to manipulation that human-driven voting systems are subject to. While LLMs hold great promise, they are not a panacea. We must acknowledge their limitations. A recent example is the Google Gemini model, which was found to have rewritten parts of American history in certain cases, highlighting the fact that LLMs can sometimes produce flawed or inaccurate answers. These shortcomings stem from biases in the training data and the inherent variability of language models. Nonetheless, when it comes to prediction market oracles that need to extract answers directly from clear and authoritative sources (such as determining the winner of an election based on articles from reputable news outlets such as CNN or the Associated Press), we believe LLM remains highly effective. 【Original text in English】
Content Sources 
