Vitalik Buterin Conducts AI Experiment with Personal Health Data

Ethereum co-creator Vitalik Buterin revealed on X that he's conducting an experiment using AI to provide personalized diet and exercise recommendations based on his personal health and travel data.

The goal is to leverage frontier models without compromising his privacy, which requires a three-layer approach: avoiding personally identifiable information, preventing identity exposure through payment channels, and maintaining network and IP-level anonymity.

To achieve this, Buterin used a local model (Qwen 3.8 Flash Next) for orchestration and remote models via tool calls to benefit from higher-level thinking. He also employed zkAPI wrapped with Tor as a CLI tool to prevent data leaks.

The experiment worked, with the AI providing recommendations improved by information from frontier models. However, Buterin noted three main deficiencies: Tor's limitations in maintaining network-layer privacy, the skill file's request construction strategies being suboptimal, and the local model's speed.